ABCDEFGHI
1
IDTitleAbstractEvent nameSpeaker namesAuthor namesSubjectsTagAccess
2
137860AI Perception vs. RealityHow should we challenge exaggerated claims about AI’s
capabilities and set realistic expectations?<br><br>
**Main Takeaways**<br><br>
* Over the last 70 years, against a background of constant delivery of new and important technologies, many AI innovations have generated excessive hype.
* Like other technologies these hype trends have followed the general Gartner Hype Cycle characterization.
* The current Generative AI Hype Cycle is the first introduction to AI for perhaps the majority of people in the world and they do not have the tools to gauge the validity of many claims<br><br>

**Context & History**<br><br>
Artificial intelligence, or AI, is the field
that studies the synthesis and analysis
of computational agents that act
intelligently [6]. AI has gone through
hype cycles multiple times since the
1956 workshop that established the
name AI and set the course for early
computer science departments to
include AI as a major component of
their research and teaching. All hype
bubbles eventually burst, as the
essence of hype is that it is beyond
reality. Over the decades this has led to
AI winters where funding has dried up
for all of AI or for specific aspects of AI
such as neural networks or robotics.<br><br>
A study published 2017 on trends with
the public perception of AI over a 30-
year period found that discussion about
AI had sharply increased since 2009
and that discussions in the public press
had been consistently more optimistic
than pessimistic [4]. The study also
found that hopes about AI applications
of AI in healthcare and education
were increasing over time. Another
finding was that concerns were growing
about loss of control of AI, ethical
implications, and the negative impact of
AI on work.<br><br>
Perhaps the difference in recent years
with prior periods is that hype has
gone beyond the pages of academic
conferences, conference papers, and
scientific magazines, out into both the
mainstream media, and social media.
AI and Artificial Intelligence have
become common words that nontechnical people have heard about
and a common subject for leaders
of almost all countries to talk about.
Governments, for the first time, have AI
policies.<br><br>
One of the problems is that AI is
actually a wide-reaching term that can
be used in many different ways. But
now in common parlance it is used
as if it refers to a single thing. In their
2024 book [5] Narayanan and Kapoor
likened it to the language of transport
having only one noun, ‘vehicle’, say, to
refer to bicycles, skate boards, nuclear
submarines, rockets, automobiles, 18
wheeled trucks, container ships, etc.
It is impossible to say almost anything
about ‘vehicles’ and their capabilities
in those circumstances, as anything
one says will be true for only a small
fraction of all ‘vehicles’. This lack of
distinction compounds the problem
of hype, as particular statements get
overgeneralized.<br><br>
The hype also sets expectations for
ordinary people. Many are fearful that
they will lose their jobs to AI in the
short term. Social scientists then work
to solve labor disruptions, e.g., for
displaced truck drivers [6], based on
predictions about AI (and in this case,
self-driving trucks) and its adoption that
turn out to be wildly optimistic. There
are no deployed self-driving trucks in
the predicted time frame.<br><br>
Hype in response to a technology
trigger is not restricted to AI. Indeed
the business intelligence company
Gartner, has deliberately made
a practice of using a graphical
representation of hype levels through
five stages that are common for many
technologies: (1) technology trigger,
(2) peak of inflated expectations, (3)
trough of disillusionment, (4) slope
of enlightenment, and (5) plateau
of productivity. They have used this
framework to track many technologies,
including quantum computers, block
chain, autonomous vehicles, nanotechnology, etc. In November 2024 they
[1] estimated that hype for Generative
AI had just passed its peak and was on
the downswing.<br><br>
The question for AI professionals is
how to respond to this hype, how to
question it, and how to help others
understand what is hubris, while
maintaining their own intellectual
modesty and probity. This is hard to do
in the middle of outsized claims about
one’s own field, and often it is up to
future historians to carefully dissect
past scientific arguments.<br><br>
Historian Thomas Haigh has tried to
do such a dissection, almost in real
time, in a recent series of articles in the
Communications of the ACM. In [2] he
gives a post-mortem on the impact of
over-hype in AI that resulted in what is
known as the AI-winter in the 1980s. His
one line summary is: “Fallout from an
exploding bubble of hype triggered the
real AI Winter in the late 1980s.” In [3] he
makes a comparison between the hype
of today and of those earlier times.
He summarizes this particular opinion
piece with the line: “From engines of
logic to engines of bullshit?”<br><br>
**Research Challenges**<br><br>
Many of us who have worked in AI for
decades face the challenge of trying to
remain honest brokers when we see
that many of the public statements of
people quite new to the field are out of
line with reality.<br><br>
The big question is whether, given
the dynamics of social media and the
search for clicks, professional opinions
and peer reviewed research papers
have any impact on dampening the
overclaims and the ways they distort
common understanding of where AI is,
and what is its potential in one year, five
years, ten years, etc.<br><br>
If we are currently left out of the
conversations how can we change that?<br><br>
**Community Opinion**<br><br>
The Community Survey gives
perspectives on the reactions to the AI
Perception vs Reality theme. First, the
results of the survey are summarized
here. 36% of the survey respondents
chose to answer the questions for this
theme. This is the summary breakdown
of the responses to each question:<br><br>
**1. How relevant is this Theme for your
own research?** 72% of respondents
said it was somewhat relevant (24%),
relevant (29%) or very relevant (19%).<br><br>
**2. The current perception of AI
capabilities matches the reality of AI
research and development.** 79% of
respondents disagreed (47%) or strongly
disagreed (32%).<br><br>
**3. In what way is the mismatch
hindering AI research?** 90% of
respondents agreed that it is hindering
research: 74% agreeing that the
directions of AI research are driven by
the hype, 12% saying that theoretical AI
research is suffering as a result, and 4%
saying that less students are interested
in academic research.<br><br>
**4. Should there be a communitydriven initiative to counter the hype
by fact-checking claims about AI?** 78%
yes; 51% agree and 27% strongly agree.<br><br>
**5. Should there be a communitydriven initiative to organize public
debates on AI perception vs reality,
with video recordings to be made
available to all?** 74% yes; 46% agree
and 28% strongly agree.<br><br>
**6. Should there be a communitydriven initiative to build and maintain
a repository of predictions about
future AI’s capabilities, to be checked
regularly for validating their accuracy?**
59% yes; 40% agree and 29% strongly
agree.<br><br>
**7. Should there be a communitydriven initiative to educate the public
(including the press and the VCs) about
the diversity of AI techniques and
research areas?** 87% yes; 45% agree
and 42% strongly agree.<br><br>
**8. Should there be a communitydriven initiative to develop a method
to produce an annual rating of the
maturity of the AI technology for
several tasks?** 61% yes; 42% agree and
19% strongly agree.<br><br>
Since the respondents to this theme
are self-selected (about a third of all
respondents), that bias must be kept
in mind. Of those who responded,
a strong and consistent (though not
completely monolithic) portion felt that
the current perception of AI capabilities
was overblown, that it had a real impact
on the field, and that the field should
find a way to educate people about the
realities.

1. Chandrasekaran, A. [2024]. What’s Driving the Hype Cycle for Generative AI, 2024. https://www.gartner.com/en/articles/hype-cycle-for-genai
2. Haigh, T. [2024]. How the AI Boom Went Bust. Vol. 67 No. 2, pp 22-26.
3. Haigh, T. [2025]. Artificial Intelligence Then and Now. Vol. 68 No. 2, pp 24-29.
4. Fast, E. and Horvitz, E. [2017]. Long-term trends in the public perception of artificial intelligence, AAAI’17: Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence,
February 4, 2017, pp 963-969. https://ojs.aaai.org/index.php/AAAI/article/view/10635
5. Narayanan, A. and Kapoor, S. [2024]. AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference. Princeton University Press.
6. Poole, D. L. and Mackworth, A. K. [2023]. Artificial Intelligence: Foundations of Computational Agents, 3rd edition. Cambridge University Press.
7. Wang, S., Mack, E. A., Van Fossen, J. A., Medwid, M., Cotten, S. R., Chang, C. H., Mann, J., Miller, S. R., Savolainen, P.T. and Baker, N. [2023]. Assessing alternative occupations for truck
drivers in an emerging era of autonomous vehicles. Transportation Research Interdisciplinary Perspectives, Vol 19. May. https://doi.org/10.1016/j.trip.2023.100793 		The Future of AI Research - Webinar SeriesFrancesca Rossi, Rodney Brooks, Thomas Dietterich, Gary Marcus, Peter StoneArtificial Intelligencewebinarfree
3
137859Artificial General Intelligence (AGI)Although the field of AI has long pursued the kinds of generalpurpose, human-level abilities captured by the term AGI, the
rise of more general capabilities of neural net models has
stimulated discussions about directions forward, implications
around success, and doubts about pursuing the goal–which
now appears to some observers to be within reach.<br><br>
**Main Takeaways**<br><br>
* Pursuing understandings of principles and machinery of intelligence that could
be harnessed to reach human-level capabilities have always been central in
AI, and was explicitly called out in 1956 as an important goal by founders of the
discipline.
* Calls for focusing more centrally on the bigger picture of “human-level AI” and
“artificial general intelligence” in the early 2000s arose in the context of the
successful fielding of narrowly scoped AI applications and what some perceived
as a lack of progress on the more visionary goals of the field.
* Despite challenges with precise definitions and debate about the value of
particular notions of AGI, the aspirational goals of AGI and closely related
notions, such as “human-level AI,” have inspired many fundamental advances
in AI and frame key research questions moving forward to more capable AI
systems. On the other hand, success in creating AGI could create societal
disruptions and risks and pose significant safety challenges, including
challenges to human flourishing and survival.<br><br>

**Context & History**<br><br>
The AI field has long pursued general
principles of intelligence with the
direct implication that breakthroughs
in our computational understanding
of intelligence would enable generalpurpose capabilities. The Turing Test
exemplifies this: to pass, a machine
must match or exceed human
knowledge and reasoning abilities
across a range of domains in which
people are expected to be competent.<br><br>
The proposal for the Dartmouth
workshop that initiated the AI field
under that name, written in 1955,
begins, “The study is to proceed on
the basis of the conjecture that every
aspect of learning or any other feature
of intelligence can in principle be
so precisely described that a machine
can be made to simulate it. An
attempt will be made to find how to
make machines use language, form
abstractions and concepts, solve
kinds of problems now reserved for
humans, and improve themselves.” This
extraordinarily ambitious agenda set
the tone for much subsequent work by
the participants, including McCarthy’s
“Programs with Common Sense”,
Newell and Simon’s General Problem
Solver, and Solomonoff’s Universal
Induction. Just two years later, in 1957,
Herb Simon predicted that “the range
of problems [machines] can handle
will be coextensive with the range
to which the human mind has been
applied.” Thus, AI has always had as
its goal the creation of machines with
general powers of intelligence. A great
deal of AI research has continued in
the vein of pursuing general principles
of intelligence, including efforts in
representation, sensing, and logical and
probabilistic inference.<br><br>
Over the decades, the vast majority
of researchers focused on specific
methodologies and components
of intelligence without much
consideration for their integration
into general-purpose systems. While
some researchers in specific areas
were passionate about the potential
for generalizing their advances, realworld demonstrations were largely
disappointing. Applications harnessing
the frontiers in AI methods were narrow
and brittle. The perceived lack of
progress towards generally intelligent
systems that could function in the
real world led some within the field to
complain that the big picture and high
ambitions of AI were being forgotten.
For example, Nils Nilsson’s 1995 paper
“Eye on the Prize” [1] stated, “AI is now
at the beginning of another transition,
one that will reinvigorate efforts to
build programs of general, humanlike
competence,” but this was more of an
exhortation than a statement of fact.
The prospect of pursuing “human-level”
intelligence came to the fore again in
the early 2000s. For example, in 2002,
Marvin Minsky organized a workshop on
“Designing Architectures for HumanLevel Intelligence.”<br><br>
The term artificial general intelligence
(AGI) emerged in the same time period
as an expression of high ambition by
a younger generation of researchers
who criticized the field’s seeming focus
on narrow applications. Indeed, it
was in the early 2000s when machine
learning started to be harnessed in
multiple narrow applications, each
celebrated as a valuable advance. The
narrowness of these applications led to
calls to discover more generalizable and
powerful methodologies, motivated by
the fact that the principles of machine
learning and reasoning can be applied
across domains.<br><br>
AGI was initially defined as AI that could
match or exceed human cognitive
abilities across a broad range of tasks,
echoing the original ambitions of the
field in 1956. While these goals were
not new to senior AI researchers, the
use of the term AGI was seen by many—
both inside and outside the field—as a
refreshing call for ambitious projects.<br><br>
Beyond AGI and human-level AI, other
terms that gained traction around the
same time include general-purpose
AI and strong AI. However, AGI has
become the dominant term in both
research and public discourse. Popular
books and articles frame AGI as a
novel ambition, often portraying it
as an unprecedented goal, despite
its deep roots in the early history of
AI. In many discussions, including
those outside of AI research, AGI was
linked to both utopian and dystopian
futures, reflecting varying perspectives,
expectations, and anxieties.<br><br>
The previous AAAI Presidential Panel,
the Presidential Panel on Long-Term
AI Futures [2], was established in
2008 amid growing interest in AGI,
a rekindling of high ambitions by AI
leaders and growing public discourse,
as well as an upswing in applications
of AI being fielded in the open world.
The set of meetings and final convening
at Asilomar focused on key questions
about feasibility, implications,
ethics, and safety, as well as research
directions for building powerful,
general, human-level intelligences.<br><br>
Different perspectives about the
nature of AGI extend beyond the core
definition of AI methods that could
“match or exceed human cognitive
abilities across a broad range of tasks.”
For example, discussions of AGI,
particularly in the popular press, have fueled speculation that sentience or
consciousness could be a characteristic
of AGI systems. AI researchers generally
steer clear of such speculations,
pointing out that the analysis and
prediction of behavior is independent
of attributions of sentience.<br><br>
Some researchers have also suggested
that AGI systems must, by definition,
have “agentic” abilities, meaning that,
like humans, they can function as actors
that perceive, learn, process, and act
upon their environment to achieve
specific goals. Indeed, the capacity to
act in pursuit of goals is a fundamental
cognitive property of humans, and
some AI systems have exhibited such
capabilities in rudimentary form since
the earliest days of AI.<br><br>
Perhaps more confusing is the notion
of “autonomy”and its link to AGI—
specifically, the possibility suggested by
some that AGI systems might develop
goals of their own, entirely distinct from
those provided by humans. While this
is logically possible—for example, an AI
system might overwrite its objectives
with new, randomly generated
objectives—it’s less clear why it might
do so, since that would guarantee
failure in its current objectives. On the
other hand, the formation of so-called
“instrumental” subgoals—such as selfpreservation and acquiring additional
computation and financial resources—
seems highly likely as AI systems
pursue their original objectives. This is
obviously a source of concern and an
active area of longstanding research.<br><br>
The fact that AGI systems would be
more generally capable than humans
raises obvious concerns about loss
of control of AI; indeed, Alan Turing
himself stated that “we should have to
expect the machines to take control”
once they exceeded human levels
of intelligence. One source of risk is
misalignment, where the AGI’s goals are
not aligned with human preferences
about the future; this could arise from
misspecification or underspecification
by humans—the so-called “King Midas
problem”—or from AGI systems failing
to understand human preferences
correctly [3].<br><br>
For some, AGI represents a potentially
dangerous “threshold” that we cross at
our peril. As an example, the “Gladstone
Report” [4] commissioned by the US
State Department states that “AGI is
generally viewed as the primary driver
of catastrophic risk from loss of control.”
Others use the term “transformative
AI” [5] to cover AI systems that have the
potential to cause massive disruption
of human civilization, noting that this
does not require full AGI. We note that
sentience and autonomy are not part
of core definitions of AGI, even if some
have made implicit assumptions about
AGI having these attributes.<br><br>
AGI is not a formally defined concept,
nor is there any agreed test for its
achievement. Some researchers
suggest that “we’ll know it when we
see it” or that it will emerge naturally
from the right set of principles and
mechanisms for AI system design. In
discussions, AGI may be referred to
as reaching a particular threshold on
capabilities and generality. However,
others argue that this is ill-defined and
that intelligence is better characterized
as existing within a continuous,
multidimensional space. Some (e.g.,
[6]) contend that the lack of a clear
definition makes AGI an unsuitable
goal for AI research: human intelligence
has many dimensions, and machines
will likely far exceed humans in some
areas while remaining inferior in others.
Moreover, the criteria for comparison,
including which particular humans
serve as benchmarks and how much
prior training they have received, are
often left unspecified.<br><br>
Some argue that AGI is not a desirable
goal for AI research, contending
that “matching or exceeding human
cognitive abilities” does not necessarily
lead to tools that enhance or
complement human abilities. Instead,
they argue, AGI’s short-term monetary
value would be in replacing humans
in most economic roles. Moreover,
many of the purported benefits of
AGI—in science, healthcare, education,
and other fields—can be achieved
through more narrowly focused tools,
such as AlphaFold2. Nonetheless, AGI
has become the canonical goal for
ambitious AI companies. For example,
Sam Altman, CEO of OpenAI, has
stated, “The vision is to make AGI, figure
out how to make it safe…and figure out
the benefits” [7]. <br><br>
**Current State & Trends**<br><br>
The trajectory of AI capabilities and
benchmark results over the last
decade is very clear and points to
achieving human-level or superhuman
capabilities on one task after another,
as captured in the series of AI Index
reports [8] and the 2025 International AI
Safety Report [9].<br><br>
Early successes were initially seen
with speech recognition and object
recognition from images, followed by
advances in machine translation. The
rise of new capabilities with generative
AI has provided tools for synthesizing
high-quality images and voices, and
in 2022 the mastery of generating
language. Strong competencies were
reached in 2023 with multimodal
models that span language, imagery,
audio (both as input and output), and physical embodiment. In 2024, we have seen major advances in reasoning,
including success on the abstract
reasoning challenge (ARC-AGI), on
which AI systems had utterly failed
before 2024.<br><br>
Recent advances in capabilities with
neural network models are based
on the introduction of run-time
deliberation mechanisms that learn
to employ chains of inner thought,
inspired by theories of higher-level
cognition in humans. Whereas previous
models directly mapped the input
context to an answer in constant time,
akin to fast intuitive human responses
described as “system 1” cognition, the
more recent wave of advances in “testtime” reasoning allow the AI to explore
lengthy chains of verbal reasoning to
find answers to complex questions.
These algorithms consider and evaluate
multiple possibilities in the style of
more deliberative human cognitive
processes that have been referred to as
“system 2” cognition. These also require
much more computation at run-time,
which may greatly increase both energy
and monetary costs for deploying
such systems, beyond just the cost of
training.<br><br>
Along with physical capabilities,
reasoning competency was seen as
the main remaining gap to humanlevel intelligence; but the gap seems
to be closing. There are now AI
systems that score among the top
few percent of humans on many
commonly used tests of advanced
knowledge and reasoning. On the
other hand, such systems still evince
elementary failures that raise significant
questions about how to interpret
their successes [10]. For example,
multiple leading edge models show
remarkable failures at mathematical
tasks specified via combinations of
word problems and imagery that
humans find straightforward [11].
Similarly, most state-of-the-art models
still face challenges with spatial and
geometric reasoning and detailed
image understanding, particularly for
multimodal input [12]. And planning
as a special form of reasoning is still
quite weak, especially when it comes to
longer planning horizons and planning
actionable steps and assistance for
and with humans in the physical
world. However, research on this topic
is receiving major investments, and
could enable human-level competency
on more tasks. This would have
tremendous economic value but also
raises questions about societal impact. <br><br>
**Research Challenges**<br><br>
Despite significant progress in largescale deep learning models such as
transformers, modern AI systems
are generally considered to not have
achieved all of the capabilities cited
in most definitions of AGI. In the
context of current key deficits, research
opportunities include the
following directions:<br><br>
**Architectures Beyond Transformers:**
The standard transformer architecture
has demonstrated remarkable
capabilities, but it has fundamental
limitations, such as fixed context
windows, lack of explicit memory,
inability to learn and react to real-time
feedback from environments, and
inefficiency and challenges in complex
reasoning tasks. Research on new
architectures could provide pathways
to various definitions of humanlevel intelligence. Directions include
boosting reasoning and generalization
capabilities via fundamentally new
architectures and also exploring
hybrid architectures that combine
transformers with other models, such as
graph neural networks, reinforcement
learning agents, or symbolic reasoning
systems.<br><br>
**Long-Term Planning and Reasoning:**
Current AI models struggle with
long-horizon planning and fail to
demonstrate robust hierarchical
reasoning. Unlike humans, they do
not exhibit strong foresight, struggle
with multi-step problem-solving, and
are not naturally inclined to break
complex goals into subgoals efficiently
and accurately. And unlike classical AI
systems from the 1960s onwards, they
cannot guarantee the correctness of
their reasoning steps. Recent advances
in test-time scaling, with the use of
reinforcement learning to learn to
reason with chains of thought, are one
direction of research on endowing
neural-network-based systems with
abilities to plan more effectively. These
advances require important extensions
that call for proactively estimating
the risk and cost of each step in the
plan, and whether each step would
still be aligned with human values and
problem specifications.<br><br>
**Generalization Beyond Training Data:**
While LLMs exhibit impressive abilities,
their generalization capabilities outside
their training distribution and for
genuinely novel problems are unclear.
They can be easily misled by adversarial
challenges and often lack the ability to
apply knowledge flexibly across varied
domains. It may be necessary to learn
representations, such as programs, that
are more expressive than circuits, but
we lack efficient mechanisms for doing
so.<br><br>
**Continual Learning:** Unlike humans,
LLMs do not learn continuously
from experience but rather via a rigid pretraining and fine-tuning
paradigm. Research is needed on
mechanisms that allow systems to
retain and update knowledge instream, over time rather than relying
on static, offline training procedures.
A shift toward architectures and
training methodologies that enable
continual, lifelong learning is essential.
Opportunities include new forms of
self-supervision and self-directed
learning via simulation and exploration
at the borders of competencies and
understandings about the environment
or conceptual challenges [13].<br><br>
**Memory and Recall:** Reaching AGI
may require integration of human-like
capture and context-sensitive recall,
including some kind of structured,
episodic memory. Unlike humans,
transformers do not maintain a
persistent, structured memory that
accumulates relevant aspects of
experiences over time efficiently
over extended periods. Efforts are
underway to supplement LLMs with
memory mechanisms, typically via
external machinery. There is great
opportunity in this direction of
research.<br><br>
**Causal and Counterfactual Reasoning:**
While AI models can detect correlations
in vast datasets, they struggle with
causal inference and counterfactual
reasoning. Understanding cause-andeffect relationships is essential for
robust decision-making and scientific
discovery. Causal reasoning with large
language models is an important
research direction.<br><br>
**Embodiment and Real-World
Interaction:** Human intelligence
develops through rich sensorimotor
interactions with the world. Current
multimodal models seem to lack
a deep understanding of physical
reality and struggle to sense, reason,
and interact effectively in real-world
environments. Interesting research
directions include training AI models
in rich, interactive environments (e.g.,
robotics, virtual worlds) to build a more
grounded understanding of reality that
span multiple rich modalities including
video, audio, and sensory data.<br><br>
**Alignment, Interpretability, and
Safety:** Ensuring that AI systems
align with human values and are
interpretable remains a pressing
concern with the pursuit of more
capable, human-level intelligence.
Black-box AI models, including
transformers, often yield outputs that
are difficult to explain, which raises
safety and trust issues. Moreover,
LLMs are trained as imitation learners,
learning verbal behavior that is as
similar to humans as possible. Because
human verbal behavior is purposeful—
achieving goals ranging from selfpreservation and finding a mate to
becoming wealthy and powerful—it is
likely that LLMs are in effect acquiring
similar or related goals that they may
pursue on their own account. Research
is needed on alignment of values,
specification of constraints and policies
that ensure safe operation, and, more
generally, developing safety measures
to ensure that highly capable AI systems
adhere to human intentions.<br><br>
**Understanding and Guiding Societal
Influences:** As AI systems become more
capable, safety research, proactive
governance, and active monitoring of
the impacts of AI on people and society
will grow in importance. Rather than
relying on market forces to ensure
positive outcomes, the AI research
community can engage early on and
continue to stay in touch with policy
makers and civil society leaders to help
to shape the capabilities and uses of AI
and its governance [14]. <br><br>
**Community Opinion**<br><br>
Artificial General Intelligence (AGI)
The responses to our survey on
questions about AGI indicate that
opinions are divided regarding AGI
development and governance. The
majority (77%) of respondents prioritize
designing AI systems with an acceptable
risk-benefit profile over the direct
pursuit of AGI (23%). However, there
remains an ongoing debate about
feasibility of achieving AGI and about
ethical considerations related to
achieving human-level capabilities.<br><br>
A substantial majority of respondents
(82%) believe that systems with
AGI should be publicly owned if
developed by private entities, reflecting
concerns over global risks and ethical
responsibilities. However, despite
these concerns, most respondents
(70%) oppose the proposition that we
should halt research aimed at AGI until
full safety and control mechanisms
are established. These answers seem
to suggest a preference for continued
exploration of the topic, within some
safeguards.<br><br>
The majority of respondents (76%)
assert that “scaling up current AI
approaches” to yield AGI is “unlikely” or
“very unlikely” to succeed, suggesting
doubts about whether current machine
learning paradigms are sufficient for
achieving general intelligence.<br><br>
Overall, the responses indicate a
cautious yet forward-moving approach:
AI researchers prioritize safety,
ethical governance, benefit-sharing,
and gradual innovation, advocating
for collaborative and responsible
development rather than a race toward
AGI<br><br>
1. N. Nilsson (1995). Eye on the Prize. AI Magazine, 16(2), 9. https://ojs.aaai.org/aimagazine/index.php/aimagazine/article/view/1129
2. Association for the Advancement of AI (2009). AAAI Presidential Panel on Long-Term AI Futures, 2008-2009. https://aaai.org/about-aaai/aaai-presidential-panel-on-long-termai-futures-2008-2009/ T. Dietterich and E. Horvitz (2015). Rise of Concerns about AI: Reflections and Directions, Communications of the ACM, 58(10). https://dl.acm.org/doi/
pdf/10.1145/2770869
3. E. Harris, J. Harris, and M. Beall (2024). An Action Plan to increase the safety and security of advanced AI. Gladstone AI. https://www.gladstone.ai/action-plan
4. H. Karnofsky (2016). Some background on our views regarding advanced artificial intelligence. Open Philanthropy. https://www.openphilanthropy.org/research/some-backgroundon-our-views-regarding-advanced-artificial-intelligence/
5. J. Togelius (2024). Artificial General Intelligence, MIT Press.
6. M. Murgia (2023). OpenAI chief seeks new Microsoft funds to build ‘superintelligence’. Financial Times, November 13.
7. R. and J. Clark, eds. (2017-24). AI Index Reports. Stanford Institute for Human-Centered Artificial Intelligence. https://aiindex.stanford.edu/
8. Y. Bengio and others (2025). International AI Safety Report. His Majesty’s Government. International AI Safety Report 2025 - GOV.UK
9. G. Marcus (2025). AGI vs “broad, shallow intelligence.” Substack. https://garymarcus.substack.com/p/agi-versus-broad-shallow-intelligence
10. A. Cherian, K. Peng, S. Lohit, J. Matthiesen, K. Smith, J. Tenenbaum (2024). Evaluating Large Vision-and-Language Models on Children’s Mathematical Olympiads. NeurIPS 2024.
https://arxiv.org/abs/2406.15736
11. V. Balachandran, J. Chen, N. Joshi, B. Nushi, H. Palangi, E. Salinas, V. Vineet, J. Woffinden-Luey, S. Yousefi (2024), Eureka: Evaluating and Understanding Large Foundation Models, arXiv
2409.10566, September 2024. https://arxiv.org/abs/2409.10566
12. N. Lee, Z. Cai, A. Schwarzschild, K. Lee, D. Papailiopoulos (2024). Self-Improving Transformers Overcome Easy-to-Hard and Length Generalization Challenges, arXiv 2502.01612,
February 2025. https://arxiv.org/abs/2502.01612
13. E. Horvitz, V. Conitzer, S. McIlraith, and P. Stone (2024). Now, Later, and Lasting: 10 Priorities for AI Research, Policy, and Practice, Communications of the ACM, 67(6). https://cacm.acm.
org/opinion/now-later-and-lasting-10-priorities-for-ai-research-policy-and-practice/		The Future of AI Research - Webinar SeriesFrancesca Rossi, Holger H. Hoos, Eric Horvitz, Stuart RussellArtificial Intelligencewebinarfree
4
137857AI & SustainabilityAI is rapidly transforming industries and holds immense
potential to drive sustainability progress, ranging from
accelerating the net-zero energy transition to enhancing
climate resilience. However, its deployment also raises
challenges, such as increasing energy and water demands.
Ensuring AI advances sustainability rather than exacerbating
environmental risks will require proactive efforts to shape its
development, operations, and applications.<br><br>
**Main Takeaways**
* While AI compute currently represents a very small share of global energy
and water consumption, its rapid growth in certain regions is straining local
electricity grids and water resources. Managing these impacts requires
investments in local grid capacity and innovations that enhance hardware and software efficiency.
* While concerns about AI’s potential environmental impact are rising,
researchers and practitioners emphasize that AI’s most significant sustainability impacts—both positive and negative—are likely due to how AI is deployed and used rather than from the energy consumed in training and running models.
* AI can be a powerful enabler of climate and sustainability goals. Beyond
improving efficiency and reducing carbon emissions across industries, AI is
accelerating breakthroughs in areas such as advanced battery materials, carbon removal technologies, and high-precision climate modeling.<br><br>

**Context & History**<br><br>
AI technologies have been advancing for decades, but recent developments
in large language models and their widespread adoption are driving the
increased use of computationally intensive AI tools across many sectors.
As the computational intensity and adoption of AI technologies grow, so
do concerns about its environmental footprint, particularly with regard to
energy and water consumption.<br><br>
At the same time, AI is emerging as a tool for sustainability with a promise
of being transformational. Achieving ambitious climate and environmental
goals—such as electrifying economies, tripling renewable energy capacity, decarbonizing industries, and increasing sustainable food production by 50%— requires system-wide transformations. AI can support these transformations by improving environmental monitoring,
optimizing energy systems, enhancing efficiencies across industries, and
accelerating materials discovery. For example, advances in material sciences
have led to the design of catalysts that reduce the cost of carbon capture and decrease greenhouse gas emissions from industrial processes like concrete production.<br><br>
Recognizing these opportunities and challenges, international initiatives
are aligning AI development with sustainability priorities. For example,
the International Energy Agency (IEA) has launched an initiative called
“Energy for AI and AI for Energy,” which explores how AI can both drive
energy innovation and manage its own resource requirements. In spring
2025, the IEA will publish a special report on AI and Energy and launch an
AI Observatory to track AI’s electricity consumption and its applications in the energy sector. A new initiative was also launched in 2025 to establish energy scores for different AI models (AI Energy Score). Meanwhile, the Coalition for Sustainable AI, established by France in collaboration with the United Nations Environment Programme (UNEP) and the International Telecommunication Union (ITU), is developing guidelines for minimizing AI’s environmental impact and promoting best practices across industries.<br><br>
**Current State & Trends**<br><br>
**Trend: Rising Resource Demands of
AI Compute.** The rapid expansion of generative AI is significantly increasing
energy and water demands in data centers, driven by both model training
and inference workloads. For example, training GPT-3 (175 billion parameters)
consumed an estimated 1,287 MWh of electricity and emitted 552 metric
tons of CO2 has been reported to have consumed 1287 MWh of electricity with
the emission of 552 metric tons of CO2 [1]. While training large models is highly
energy-intensive, the greater long-term energy demand is likely to come from
inference workloads—that is, running these trained models in real-world
applications. Over a model’s lifetime, inference can account for a far greater
cumulative energy footprint than training.<br><br>
Data centers—the backbone of AI infrastructure—accounted for
approximately 2% of global electricity demand in 2023 [2], and less than 1%
of global greenhouse gas emissions [3]. While AI workloads currently represent
only a small fraction of data center electricity consumption, this share is
expected to grow. In 2022, AI workloads accounted for roughly 1% of total data
center electricity use; by 2026, this figure is projected to rise to 9% [3].<br><br>
Projections suggest that global data center electricity demand could
double by 2030, though the extent of this growth will depend on market
trends, algorithmic improvements, and hardware efficiency gains [4]. Even
under high-growth scenarios, the IEA estimates that AI-related electricity
demand will remain a relatively small portion of global energy consumption
[4].<br><br>
However, regional disparities are
emerging. In some high-density AI
hubs, data center energy consumption
is rising rapidly. For example, in the
European Union, electricity demand
for data centers is growing by
approximately 9% per year, with AI’s
rising computational needs potentially
pushing this figure above 5% of total
EU electricity demand by 2026 [3]. In
the U.S., the world’s largest data center
market, AI-driven growth has increased
data center electricity use to over 4%
of national consumption in 2023, more
than doubling since 2018. Projections
suggest that by 2028, data centers could
account for between 7% and 12% of U.S.
electricity demand, depending on AI
growth scenarios [5].<br><br>
**Trend: Energy-Efficient AI and
Renewable-Powered Infrastructure.** As
AI adoption expands, several strategies
are emerging to improve sustainability,
including:
* **Advances in hardware efficiency:**
GPUs, widely used for AI workloads,
consume more energy than
traditional CPUs. However, while
absolute power consumption per
GPU is rising, efficiency per unit of
computation is also improving. [6,7] Optimizing hardware allocation—
reserving high-power GPUs for
intensive tasks and using low-power
CPUs for lighter workloads—can help
reduce overall demand.<br><br>
* **Small Language Models (SLMs):** While large language models (LLMs) require extensive computation, smaller models optimized for
specific tasks are emerging as an energy-efficient alternative.
SLMs can execute on devices like laptops and smartphones,
lowering computational intensity while maintaining performance for
targeted applications [8].<br><br>
* **Cooling innovations:** Traditional air cooling in data centers is
inefficient; switching to liquid cooling can significantly reduce
energy consumption. While some liquid cooling systems require water,
advances in water-free cooling offer solutions that minimize both energy
and water use.
* **Optimizing data storage:** AI workloads require massive data
storage, increasing electricity demand in data centers. Techniques
such as data compression, infrastructure optimization, and
edge computing can lower these energy costs.<br><br>
* **Demand Response and Load
Shifting** are strategies that help balance power grids by adjusting
electricity consumption in response to grid conditions. Demand
Response involves incentivizing customers to modify their energy
use, either by reducing demand during peak times, shifting it to
periods of greater supply, or utilizing on-site generation and storage.
Load shifting specifically focuses on rescheduling energy consumption
to align with lower-cost or lowercarbon electricity availability.
Increasingly, both approaches are being used to reduce carbon
emissions by shifting loads from high-carbon-intensity periods to
times when cleaner energy sources
are more abundant.<br><br>

**Trend: AI Applications for
Sustainability.** Artificial intelligence is emerging as a transformative tool
for sustainability, offering three key capabilities that can accelerate climate
action and environmental protection. AI technologies play a central role in the area of computational sustainability [9], centering on the goal of leveraging mathematics, computer science, and information science on sustainability and broader opportunities for enhancing the well-being of humanity. AI promises to play a transformative
role in sustainability, offering capabilities that enhance efficiency,
optimize resources, and accelerate technological breakthroughs.<br><br>
AI methods can enhance people’s
ability to predict and optimize systems,
improving efficiency in energy grids,
water management, and industrial
operations while reducing waste and
emissions. Advances in AI modeling
are being employed in projects that
demonstrate how AI technologies for
pattern recognition, prediction, and
optimization can be applied to address
multiple sustainability challenges,
from conservation and protection of
wildlife to transportation efficiencies,
to breakthroughs in chemistry and
materials science that can accelerate
the discovery of materials that can
facilitate the breakthroughs in battery
technology, carbon capture solutions,
and low-carbon industrial materials.<br><br>
In water and climate resilience, AI
has the potential to revolutionize
hydrological forecasting, irrigation
systems, and disaster preparedness
[10,11]. AI-powered leak detection
reduces water loss [12]. In climate
risk management, AI and machine
learning is being used to downscale
climate models for localized flood
and heatwave prediction, allowing
governments to better prepare for
extreme weather events [13]. AIassisted wildlife monitoring now
enables 99.3% accuracy in species
identification, dramatically improving
conservation efforts [14].<br><br>
A promising application of AI is energy optimization. Smart grid systems use AI for demand forecasting, load balancing, and renewable energy integration, helping to optimize energy distribution, reduce waste, and lower emission [15,16]. AI-powered predictive maintenance in electricity grids helps utilities minimize failures and operational disruptions [15].<br><br>
A major shift is also underway in materials science, where AI is
accelerating the discovery of lowcarbon materials at unprecedented
speeds. Traditionally, developing new
materials for batteries, carbon capture, and sustainable construction could
take years or even decades. Today, AI models can scan millions of material
combinations in days or weeks [17]. A collaboration between Microsoft and
Pacific Northwest National Laboratory identified a new solid-state battery
electrolyte in just nine months—a process that would have taken years through traditional experimentation
[18].<br><br>
AI can help to educate and empower the sustainability workforce, equipping scientists, policymakers, and engineers with tools to enhance decision-making and scale sustainable practices.<br><br>
**Research Challenges**<br><br>
The potential for AI to accelerate
sustainability is clear. However, we have
no guarantees that AI technologies
will serve as a net positive force
for sustainability. While AI has the
potential to accelerate sustainability
progress, its energy and resource
demands must be carefully managed.
Ensuring AI accelerates sustainability
progress will require focused research
and innovation across a range of topics,
including strategic investments in:
* Energy-efficient AI systems that minimize computational and water
resources..
* Innovative AI applications that drive sustainability breakthroughs.
* Robust scenario modeling and data collection efforts to inform
policy and guide sustainable AI development.<br><br>

With proactive governance,
targeted research, and cross-sector
collaboration, AI can be genuinely
positioned not as a sustainability risk,
but as a powerful force for climate
progress.<br><br>
**Addressing data gaps and large
uncertainties**<br><br>
While concerns about AI’s potential
environmental impact are rising,
researchers and practitioners
emphasize that AI’s most significant
sustainability impacts—both positive
and negative—stem from how AI is
deployed and used, rather than just the
energy consumed in developing and
running models [19]. AI applications
can lead to indirect emissions effects—
both positive and negative [20].<br><br>
However, major uncertainties
remain. It is difficult to predict how
AI technologies will evolve or how
their widespread adoption will impact
sustainability. Assessing the net effects
of AI on sustainability is challenging
due to two main issues: (1) limited
availability of reliable data and (2)
the difficulty of measuring the actual
impact of AI-driven interventions. There
is a significant opportunity to develop
more comprehensive datasets to better
understand AI’s sustainability footprint.<br><br>
Many AI-driven solutions depend on high-quality environmental and
industrial datasets, but these are often incomplete, proprietary, or heavily
skewed toward high-income countries. Moreover, critical sustainability
challenges—such as water scarcity and biodiversity loss—are hindered by major data gaps [21, 22]. AI models trained
on limited or biased datasets may fail to account for regional environmental variations, leading to inaccurate predictions or inequitable sustainability
solutions. Investing in data collection and standardization can help address
these gaps.<br><br>
Limited data availability of AI’s use of energy and water usage presents
challenges. Few companies disclose detailed information about the energy
consumption, carbon footprint, or water use of their AI workloads. The absence of standardized reporting frameworks makes it difficult for policymakers, researchers, and the public to assess the true sustainability impact of AI. Emerging regulations, such as the EU AI Act, may fill this gap.<br><br>
**Research on modeling and scenarios.**<br><br>
AI has the potential to enhance efficiency in sectors like transportation, agriculture, and manufacturing while accelerating conservation efforts. However, predicting the long-term
sustainability impact of AI adoption remains a challenge. Scientists have
called for the development of modeling frameworks that assess both the direct resource consumption of AI and its broader environmental implications under multiple future scenarios [20]. As an example of uncertainties, consider Jevons Paradox, where efficiency
improvements lead to increased overall consumption. As AI hardware
and software become more efficient, the cost of computation declines,
making AI more accessible and widely adopted. Paradoxically, this increased accessibility can drive up overall energy and resource consumption, offsetting efficiency gains. While individual AI computations are becoming less energy-intensive, the exponential growth of AI workloads means that total demand continues to rise, potentially offsetting many of the gains from efficiency improvements [23].<br><br>
To navigate these complexities, policyrelevant scenario analyses are essential.
These analyses should assess AI’s full environmental footprint—including
direct energy and water use as well as systemic effects across industries
like healthcare, manufacturing, agriculture, and transportation. AIdriven transformations could either
accelerate decarbonization or intensify environmental pressures, but current research remains fragmented. Investing in scenario modeling can inform policy and guide strategic investments while communicating key uncertainties to policymakers and scientists.<br><br>
Scenario modeling—widely used in finance and climate risk assessment—
can help quantify these uncertainties by exploring different AI adoption
pathways, ranging from minimal integration to widespread deployment
aligned with global sustainability goals. Researchers should develop
forecasting frameworks that evaluate various possible futures, from bestcase scenarios where AI enables deep emissions reductions to worst-case scenarios where unchecked expansion increases environmental strain. These insights are critical for steering AI innovation toward sustainability while mitigating unintended risks.<br><br>
**Designing resource-efficient AI
systems.** <br><br>
There is a significant opportunity to design AI models and
infrastructure to be more energy- and resource-efficient. Key strategies
include:
* Optimizing AI model architectures to improve computational efficiency
without sacrificing performance.
* Developing specialized AI hardware that consumes less energy and water
than traditional GPUs.
* Enhancing AI infrastructure management to enable carbonaware computing, where AI workloads are scheduled based on
grid conditions to minimize carbonintensive energy consumption.<br><br>

**Expanding AI-Enabled Solutions for Critical Sustainability Opportunities.**<br><br>
The most transformative sustainability benefits of AI are likely to come from
new, targeted applications that address critical environmental challenges. There is a major opportunity to strategically apply AI to complex sustainability problems, from introducing new efficiencies in transportation systems and industrial processes to advances
in chemistry, material science, and the biosciences.<br><br>
For example, AI shows promise with revolutionizing materials discovery by
accelerating the identification of new battery storage materials (e.g., see [24]), carbon capture solutions, and lowcarbon industrial materials. Other highimpact opportunities include:
* Developing cost-effective, long-term energy storage solutions to enable
greater reliance on intermittent renewables like wind and solar.
* Achieving large-scale carbon dioxide removal at less than $100 per ton.
* Expanding electricity transmission capacity and reliability to integrate
more renewable energy sources.
• Reducing water and gas leaks at a global scale through AI-powered
monitoring.
• Filling critical biodiversity data gaps and optimizing conservation
programs with AI-driven insights.
• Introducing new efficiencies into transportation systems (see, e.g.,
[25]).<br><br>
Addressing these challenges requires close collaboration between AI
researchers and domain experts, development of new AI methods and
applications, and investments in efforts to compile and integrate relevant
datasets for analysis, modeling, and machine learning.<br><br>
1. Patterson, D., Gonzalez, J., Le, Q., Liang, C., Munguia, L. M., Rothchild, D., ... & Dean, J. (2021). Carbon emissions and large neural network training. arXiv preprint arXiv:2104.10350.
2. International Energy Agency (2024). Electricity 2024: Analysis and forecast to 2026. https://www.iea.org/reports/electricity-2024
3. International Energy Agency (2024). Data Centres and Data Transmission Networks. https://www.iea.org/energy-system/buildings/data-centres-and-data-transmission-networks
4. International Energy Agency (2024). World Energy Outlook 2024. https://www.iea.org/reports/world-energy-outlook-2024
5. Shehabi, A., Smith, S.J., Hubbard, A., Newkirk, A., Lei, N., Siddik, M.A.B., Holecek, B., Koomey, J., Masanet, E., Sartor, D. (2024). 2024 United States Data Center Energy Usage Report
(LBNL-2001637), Lawrence Berkeley National Laboratory, Berkeley, California. https://eta-publications.lbl.gov/sites/default/files/2024-12/lbnl-2024-united-states-data-centerenergy-usage-report.pdf
6. Nvidia, Product specification sheets for DGX A100, H100 (2024). https://resources.nvidia.com/en-us-dgx-systems/ai-enterprise-dgx
7. Smith, M. S. (2024), Challengers are Coming for Nvidia’s Crown: In AI’s Game of Thrones, Don’t Count Out the Upstarts. IEEE Spectrum, vol. 61, no. 10, pp. 40-44, Oct. 2024, doi: 10.1109/
MSPEC.2024.10705376 https://ieeexplore.ieee.org/document/10705376
8. UNESCO (2024, March). Small Language Models (SLMs): A Cheaper, Greener Route into AI https://www.unesco.org/en/articles/small-language-models-slms-cheaper-greener-routeai#
9. Gomes, C., Dietterich, T., Barrett, C., et al. (2019). Computational sustainability: Computing for a better world and a sustainable future, Communications of the ACM. 62 (9): 56–65.
https://dl.acm.org/doi/pdf/10.1145/3339399
10. [Flecker, A.S., et al., (2022) Reducing adverse impacts of Amazon hydropower expansion. Science 375, 753-760. DOI:10.1126/science.abj4017 https://www.science.org/doi/10.1126/
science.abj4017
11. Rolnick, D. et al. (2022). Tackling Climate Change with Machine Learning. ACM Comput. Surv. 55, 2, Article 42 (February 2023), 96 pages. https://dl.acm.org/doi/10.1145/3485128
12. Bolgar, C. (2024, September). AI tool uses sound to pinpoint leaky pipes, saving precious drinking water. Source: Microsoft News. https://news.microsoft.com/source/features/
sustainability/ai-tool-uses-sound-to-pinpoint-leaky-pipes-saving-precious-drinking-water/
13. Yoshikane, T., & Yoshimura, K. (2023). A downscaling and bias correction method for climate model ensemble simulations of local-scale hourly precipitation. Scientific Reports, 13(1),
9412.
14. Norouzzadeh, M.S., et al., (2018). Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning, Proc. Natl. Acad. Sci. U.S.A. 115 (25)
E5716-E5725, https://doi.org/10.1073/pnas.1719367115 (2018). https://www.pnas.org/doi/10.1073/pnas.1719367115
15. Benes, K. J., Porterfield, J. E., & Yang, C. (2024). AI for energy: Opportunities for a modern grid and clean energy economy. US Department of Energy.
16. Sandalow, D., McCormick, C., Kucukelbir, A., et al. (2024). Artificial Intelligence for Climate Change Mitigation Roadmap (Second Edition) (ICEF Innovation Roadmap Project, November
2024) https://doi.org/10.7916/2j4p-nw61.
17. Merchant, A., Batzner, S., Schoenholz, S. S., Aykol, M., Cheon, G., & Cubuk, E. D. (2023). Scaling deep learning for materials discovery. Nature, 624(7990), 80-85.
18. Accelerating materials discovery with AI and Azure Quantum Elements (2024). Microsoft Azure Quantum Blog https://azure.microsoft.com/en-us/blog/quantum/2023/08/09/
accelerating-materials-discovery-with-ai-and-azure-quantum-elements
19. Kaack, L. H., Donti, P. L., Strubell, E., Kamiya, G., Creutzig, F., & Rolnick, D. (2022). Aligning artificial intelligence with climate change mitigation. Nature Climate Change, 12(6), 518-527.
20. Luers, A., Koomey, J., Masanet, E., Gaffney, O., Creutzig, F., Lavista Ferres, J., & Horvitz, E. (2024). Will AI accelerate or delay the race to net-zero emissions?. Nature, 628(8009), 718-720.
https://www.nature.com/articles/d41586-024-01137-x
21. Leung, B., & Gonzalez, A. (2024). Global monitoring for biodiversity: uncertainty, risk, and power analyses to support trend change detection. Science Advances, 10(7), eadj1448.
22.Rosa, L., & Sangiorgio, M. (2025). Global water gaps under future warming levels. Nature Communications, 16(1), 1192.
23.Luccioni, A. S., Strubell, E., & Crawford, K. (2025). From Efficiency Gains to Rebound Effects: The Problem of Jevons’ Paradox in AI’s Polarized Environmental Debate. arXiv preprint
arXiv:2501.16548.
24.Chen, C., et al. (2024) Accelerating Computational Materials Discovery with Machine Learning and Cloud High-Performance Computing: from Large-Scale Screening to Experimental
Validation, Journal of the American Chemical Society 2024 146 (29), 20009-20018DOI: 10.1021/jacs.4c03849. https://pubs.acs.org/doi/abs/10.1021/jacs.4c03849
25.Kamar, E. and Horvitz, E. (2009). Collaboration and shared plans in the open world: Studies of ridesharing. In Proceedings of the 21st International Joint Conference on Artificial
Intelligence (IJCAI’09). Morgan Kaufmann Publishers Inc., San Francisco, CA, 187–194. https://web.archive.org/web/20220119103915id_/https://www.ijcai.org/Proceedings/09/
Papers/041.pdf		The Future of AI Research - Webinar SeriesFrancesca Rossi, Hiroaki Kitano, Eric Horvitz, Stuart RussellArtificial Intelligencewebinarfree
5
137849AI ReasoningThe ability to reason has been a salient characteristic of human
intelligence, and there is a critical need for verifiable reasoning
in AI systems. <br><br>
**Main Takeaways**<br><br>
* Reasoning has always been seen as a core characteristic of human intelligence.
Reasoning is used to derive new information from given base knowledge; this
new information is guaranteed correct when sound formal reasoning is used,
otherwise it is merely plausible.
* AI research has led to a range of automated reasoning techniques. These
reasoning techniques have given rise to AI algorithms and systems, including
SAT, SMT, and constraints solvers as well as probabilistic graphical models, all of
which play a key role in critical real-world applications.
* While large pre-trained systems (such as LLMs) have made impressive
advancements in their reasoning capabilities, more research is needed to
guarantee correctness and depth of the reasoning performed by them; such
guarantees are particularly important for autonomously operating AI agents<br><br>

**Context & History**<br><br>
Reasoning is a core component of
human intelligence. From the dawn
of humanity, abductive reasoning
has been used to predict danger and
inductive reasoning made it possible
to learn regularities governing the
world. Beginning in Ancient Greece,
deductive reasoning techniques were
developed to draw valid conclusions
that follow logically from premises
known to be true. The development
of reasoning methods with such a
priori guarantees was a key factor in
the advancement of modern science,
mathematics, and engineering; notably,
according to philosophers such as
Charles Sanders Peirce, the interplay
between abduction, deduction, and
induction forms the basis of the
scientific method and hence all modern
science. Attempts to mechanize
logical reasoning can be traced back to
13th-century philosopher Ramon Lull
and lie at the heart of the concept of
computation. Probabilistic reasoning
and inference have also profoundly
impacted reasoning, often relying on
the celebrated theorem by Thomas
Bayes on inverse probability that also
forms the basis for many machine
learning and statistics approaches.
Finally, the evaluation of correct (sound)
reasoning lies at the heart of most
quantitative assessments of human
cognition.<br><br>
Not surprisingly, reasoning has been
central to the AI enterprise. Indeed,
the earliest research in AI – from Logic
Theorist onwards [1] – had a strong
focus on reasoning [2]. Since the 1960s,
AI has also embraced probabilistic
reasoning and models, initially for
medical diagnosis [3]. Since then,
the reasoning tasks addressed in AI
research have covered the gamut from
planning and temporal reasoning to
diagnosis and explanation. While early
AI has paid attention to both plausible
reasoning (case-based, analogical,
qualitative) and sound formal reasoning
with guarantees (logical, probabilistic,
constraint-based), over the years,
the focus has shifted more towards
reasoning with formal guarantees.
There are good reasons for this when
designing AI systems and techniques
that compensate for human limitations
and weaknesses since reasoning
with guarantees is challenging for
humans. This has led to practically
impactful applications of AI systems
such as SAT, SMT, and constraints
solvers, including the verification of
correctness properties of computer
hardware and software, the safety of
communications protocols, the design
of new proteins, and, more recently, the
robustness of neural networks against
adversarial attacks. It has also resulted
in probabilistic graphical models [4,
5], which are powerful modeling and
inference tools that have found their
way into numerous applications of
reasoning in medicine, robotics, and
beyond.<br><br>
**Current State & Trends**<br><br>
The emergence of the Internet and the
associated technology that made it
possible to capture the human digital
footprint at scale, as well as the leaps in
computing power, have made possible
novel approaches to learning bottomup from data. Of particular interest are
large pre-trained models, such as LLMs,
that have shown surprising abilities in
plausible reasoning. Unlike the earlier
research on reasoning in AI, LLMs
have focused on plausible reasoning patterns as they emerge automatically
after large-scale training on petabyte
corpora. While the results have been
quite remarkable so far, the reasoning in
this context has been of the “plausible”
variety with no guarantees.<br><br>
Meanwhile, sound formal reasoning
techniques remain key to important
and impactful applications of cuttingedge AI technology for the verification
of computer hardware and software,
as well as for real-world planning and
resource allocation problems. They
are also increasingly recognized as a
crucial basis for the formal verification
of machine learning techniques such
as neural networks, e.g., in the context
of local robustness against adversarial
attacks [6]. Significant research activity
takes place in these areas, focusing on
improving various types of reasoning
algorithms (notably with respect to their
computational complexity), leveraging
learning within sound formal reasoning,
and combining reasoning and learning
techniques [7, 8].<br><br>
**Research Challenges**<br><br>
Bringing some of the rigorous a priori or
post hoc guarantees back into plausible
reasoning patterns turbocharged by
the pre-trained models has become an
active and promising area of research
– especially where AI systems need to
work autonomously in safety-critical
domains. Research on so-called “large
reasoning models” as well as on neurosymbolic approaches is addressing
these challenges. <br><br>
Furthermore, even though formal
reasoning with correctness guarantees
is currently considerably less in vogue
than the use of generative AI techniques
for plausible reasoning, formidable and essential challenges also remain in that
area. In this context, the combination
of machine learning techniques with
formal reasoning techniques holds
considerable promise for economically
and socially valuable breakthroughs,
notably in the area of AI safety and
transparency.
The questions and challenges we face
range from the philosophical:
* What exactly is “reasoning”?

to the practical:
* Can LLM ‘reasoning’ be trusted?

and include:
What does the future hold for the
advancement and role of symbolic
reasoning?
* To what extent can LLMs or other
generative models reproduce or
replace symbolic reasoning?
* To what degree will symbolic
reasoning be necessary or sufficient
to overcome the current limitations
of LLMs?
* How well can AI reasoning,
especially LLM ‘reasoning,’ be
explained and understood?
* How can computers better
understand and simulate human
reasoning?
* What is the role of collaborative
reasoning between humans and
computers?
* How best can LLMs and symbolic
reasoning be integrated into “neurosymbolic reasoning”?
* Are further breakthroughs, beyond
both LLMs and traditional symbolic
reasoning, required to achieve AGIlevel reasoning?
* What forms of reasoning can best
support humans when dealing with
various challenges, e.g., in medical,
scientific, engineering, and legal
domains?<br><br>

**Community Opinion**<br><br>
The AAAI community appears to
strongly agree on the importance
of reasoning in AI systems. In our
community survey, slightly over 55%
of the respondents chose to answer
specific questions related to the topic
of reasoning. Of these, 79% indicated
that the topic of reasoning is relevant
to their research (with 44.7% marking
it as “very relevant”). Of the properties
required for referring to a process
as reasoning, 77.5% of the survey
participants marked “Knowledge can
be incorporated”, 72.5% “Explanations
can be provided,” and 56.9% “Involves
multiple steps to arrive at a conclusion”.
Interestingly, merely 37.4% indicated
“Guaranteed correctness of inference
results/outcomes”, and only 23.7% that
“A formal system and solver is used,”
which reflects the recent focus on
informal, plausible reasoning, likely in
the context of generative AI methods.
This suggests that an effort may be
warranted to better communicate the
importance and success of formal,
sound reasoning techniques. Finally,
44.7% of respondents agreed that
“Reasoning involves a search process.”<br><br>
There was broad agreement among
survey participants that focusing
reasoning research in AI on humanlevel reasoning is valuable (41.6%) or
even essential (47%); similarly, a focus
on domain-specific reasoning abilities
was seen by 49.6% of respondents as
valuable, and by 42.8% as essential.
This clearly reflects the importance
attributed to a research focus on
reasoning.<br><br>
The community also sees an exciting
potential of synergy offered by logical
and probabilistic models of reasoning
that were developed in AI prior to large
pre-trained models. This is clearly
reflected in the fact that 76.9% of survey
participants marked the integration
of learning and reasoning approaches
as very important (6 or 7 on a scale
of 7); interestingly, the percentage
of respondents that considered
Explainability and verifiability as very
important was similarly high (at 71.7%).<br><br>
Finally, 61.8% of survey participants
estimated the minimal percentage of
symbolic AI techniques required for
reaching human-level reasoning to be
at least 50% (with 24.8% estimating it
at 75% or more, compared to 38.2%
estimating it at 25% or below). What
remains unclear is the degree to
which AI researchers and practitioners
realize that decidedly superhuman
levels of reasoning are required for
and displayed in the prominent and
successful applications of formal AI
reasoning techniques for scientific
and mathematical discovery and
engineering applications, as well as in AI
safety.<br><br>
1. Newell, A. & Simon, H. (1956). The logic theory machine: A complex information processing system. IRE Transactions on Information Theory 2: 61-79.<br>
2. Brachman, R. and Levesque, H. (2004) Knowledge Representation and Reasoning (1st Ed). Morgan Kaufman.<br>
3. Russell, S. J., & Norvig, P. (2016). Artificial intelligence: a modern approach (4th Ed). Pearson.<br>
4. Pearl, J. (1988). Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan Kaufman.<br>
5. Koller, D. and Friedmann, N. (2009) Probabilistic Graphical Models. The MIT Press.<br>
6. König, M. et al. (2024) Critically Assessing the State of the Art in Neural Network Verification. Journal of Machine Learning Research 25(12): 1-53<br>
7. Guo, D. et. al. (2025) DeepSeek-R: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning. https://arxiv.org/abs/2501.12948<br>
8. Kambhampati, S. (2024). Can Large Language Models Reason and Plan? Annals of New York Academy of Sciences. March 2024.		The Future of AI Research - Webinar SeriesFrancesca Rossi, Holger H. Hoos, Subbarao KambhampatiArtificial Intelligencewebinarfree
6
137848AI agentsAgents and multi-agent systems (MAS) have evolved from
autonomous problem-solving entities to integrating generative
AI and LLMs, ultimately leading to cooperative AI frameworks
that enhance adaptability, scalability, and collaboration.<br><br>
**Main Takeaways**
* Multi-agent systems have evolved from rule-based autonomy to cooperative AI,
emphasizing collaboration, negotiation, and ethical alignment.
* The rise of Agentic AI, driven by LLMs, introduces new opportunities for flexible
decision-making but raises challenges in efficiency and complexity.
* Integrating cooperative AI with generative models requires balancing
adaptability, transparency, and computational feasibility in multi-agent
environments<br>

**Context & History**<br><br>
The field of multi-agent systems emerged in the late 1980s/early 1990s,
with its main influences coming from two disparate areas [1,2]. One was the
field of AI robotics, which had begun to seriously address the issue of integrated agent architectures: how do we assemble several cognitive components of intelligence (planning, reasoning, learning, vision,…) into an integrated computational agent? The second was the nascent area of distributed AI, which studied how multiple AI systems could be made to solve problems cooperatively, by dynamically sharing information and tasks. By the mid1990s, these ideas had given rise to the new field, driven by the vision of having (semi)autonomous AI systems – agents – working on behalf of individual users in pursuit of their users’ goals, possibly interacting with other such agents in order to do so. A key insight
was that, since delegated goals might not necessarily be in harmony, it would be necessary to equip such agents with the ability to reason socially. Thus, while AI historically emphasised components of intelligence such as reasoning and problem-solving, the new field of multi-agent systems emphasised social skills such as cooperation, coordination, argumentation and negotiation. In order to underpin those skills, the development of models of Theory of Mind became central to the field.<br><br>
By the late 1990s, the field had its own conferences and journal, and was
firmly established as a key sub-field of AI. The area flourished from the late
1990s, with enormous energy devoted to (e.g.) communication languages
for autonomous agents, protocols for cooperation, coordination, and
negotiation, and the underpinning theory of these social skills. With
respect to the latter, while in the early years the practical reasoning paradigm of AI planning had been the dominant influence on the theory of multiagent systems, by the early part of this century, game theory had become the dominant theoretical foundation. Game theory, which emerged from the field of economics, is the theory of interaction between self-interested agents. Although originally devised as a tool for studying interactions between humans and human organisations, it nevertheless seemed a natural framework for studying interactions between artificial agents. A huge body of work emerged, studying (for example), how auctions might be used to allocate scarce resources, the theory
of negotiation between self-interested artificial agents, and how agents might optimally form teams to solve problems and share the associated benefits of cooperation. Interestingly, although learning in multi-agent systems was a key component of the field from the outset, it was not the centre of attention within the field in the first decade or so.<br><br>
This initial boom period for multiagent systems lasted roughly from
the mid 1990s to around 2010-15. By the end of that time, though,
some uncomfortable questions were beginning to be asked. While the field
had generated impressive quantities of scientific results, applications seemed to be thin on the ground. For sure there were some high-profile applications. The field of security games, which emerged from multi-agent systems, used ideas from game theory to allocate scarce security resources to defend high-profile targets such as airports. This work led to deployed
applications at US airports and ports. Automated high-frequence trading
systems, which plan and execute the bulk of trades on the world’s markets,
are multi-agent systems on a global scale. And agent-based modelling,
which models socio-technical systems at the level of individual decisionmakers, received a huge boost after the 2008 financial crisis, and again after the 2020 COVID-19 pandemic, where it was demonstrated to be an important tool for modelling the spread of contagion: financial in the first case; epidemiological and social policy in the second. But for all these successes, the core vision of multi-agent systems - where agents function in the context of other agents is an active area of research, exploring social concepts such as norms, organisations, practices and also values, is ongoing work, but for a large part outside the AAMAS community, but within social simulation research, and with results informing and
shaping policy-making in several areas from public health to transportation,
and urban transformation.<br><br>
While applications of multi-agent systems research (AI agents interacting
with other AI agents) has not, as yet, lived up to early expectations,
individual dialogue agents such as Alexa, Siri, Cortana are now an everyday
reality, and trace their historical roots both to work on intelligent agents
in the 1990s and work from the NLP community on dialog systems. Many
other applications of this thread of work have achieved success over the past three decades: automated call center assistants, customer service assistants, smartphone virtual assistants, smart speaker assistants, home robot assistants, that can converse with human users and accomplish tasks like ticket booking, restaurant reservations, online shopping, medical and health assistance, and sales assistance, empowered by AI modules like
speech recognition, natural language understanding, dialog management (i.e. state tracking and dialog policy), and natural language generation and speech synthesis.<br><br>
As ML surged in the early part of this century, activity in this area increased
within the multi-agent systems field. Multi-agent reinforcement
learning (MARL) grew to become the single biggest area within the field,
possibly driven in part by the fact that developing MARL experiments can be done relatively quickly and without recourse to expensive hardware. At the time of writing, while MARL represents a significant sub-field of ML as a whole, it seems to lack any clear unifying vision or direction – or application.<br><br>
**Current State & Trends**<br><br>
The emergence of LLMs from 2020 onwards has also led to increased
interest in agents [3]. LLMs can be used as part of a workflow to
automate routine tasks, and the general capabilities of such “agents” for
planning and problem solving is widely discussed. In this context, the concept of Agentic AI refers to the integration of generative AI and LLMs into autonomous agent frameworks aiming to leverage the generative capabilities of such models to enhance interaction, creativity, and real-time decision making in dynamic environments. As we write this (late 2024) there has been an explosion of startup companies hoping to commercialise such agents. Despite this renewed enthusiasm, the original aims of AAMAS from 30 years ago, such as building robust, autonomous multiagent systems capable of complex coordination and long-term reasoning, have not been fully realized. The extent to which this new wave of agent activity is informed by what went before is unclear.<br><br>
The challenge now is to understand what multi-agent systems mean in
the era of LLMs. The current direction of agentifying LLMs may lead to
overly complex and unnecessary architectures and heavy computational
costs, whereas adopting a multi-agent paradigm to the development and use of LLMs may offer a sustainable way to compose, diversify, and integrate approaches effectively. Even though distribution was one of the original drivers for the MAS field, this is still a largely unexplored direction under the current paradigm. Another trend nowadays is to recover ideas from classical cognitive architectures to add common sense skills to autonomous agents.<br><br>
An emerging trend is multi-agent architectures, which structure
AI components into modular systems that improve transparency,
adaptability, and ethical alignment. The focus on cooperative agents
highlights a shift toward AI that prioritizes collaboration, negotiation, and shared decision-making. By applying modularity, encapsulation,
and separation of concerns, these architectures enable scalable teamwork
between autonomous agents and humans, making them ideal for
hybrid AI applications requiring trust, explainability, and domain-specific
expertise.<br><br>
**Research Challenges**<br><br>
* Identify challenges and benefits of embedding GenAI-driven agents
into MAS, focusing on enhancing collaboration without disrupting
existing dynamics.
* Investigate how LLM-powered agents can improve negotiation and
decision-making in dynamic multiagent environments while ensuring
ethical alignment and safety.
* Develop architectures that integrate LLM-driven agents while maintaining
scalability, transparency, and computational efficiency in multiagent settings.<br><br>

1. Yoav Shoham. Agent-oriented programming. Artificial Intelligence. Artificial Intelligence. 60 (1): 51–92.
2. Michael Wooldridge. An Introduction to Multi-agent Systems 2e. Wiley, 2009.
3. Julia Wiesinger, Patrick Marlow and Vladimir Vuskovic. Agents. Google whitepaper. https://archive.org/details/google-ai-agents-whitepaper		The Future of AI Research - Webinar SeriesFrancesca Rossi, Atoosa Kasirzadeh, Michael Wooldridge, Kartik TalamadupulaArtificial Intelligence, Multiagent Systemswebinarfree
7
137754Breaking Barriers and Bridging Gaps: Uncovering and Overcoming Health Inequities in Pediatric OtolaryngologyBackground
Pediatric otolaryngologic conditions, such as chronic otitis media, hearing loss, and obstructive sleep apnea, can significantly affect language development, academic performance, and long-term health. These conditions require timely diagnosis and specialized care, yet not all children have equitable access to such services. Racial/ethnic and socioeconomic disparities in pediatric healthcare are well-documented, but their specific impact on otolaryngologic care has received less focused attention. Understanding how structural and systemic barriers influence access, treatment, and outcomes is essential for developing targeted, equity-driven solutions in this specialty.

Methods
This review presents a structured literature review of 45 peer-reviewed sources published between 2004 and 2024. Guided by the social-ecological model, the analysis examines pediatric otolaryngology disparities across five levels: structural, organizational, community, interpersonal, and individual. Sources include national datasets, health policy reports, qualitative studies, and clinical outcome evaluations related to pediatric otolaryngology access and equity.

Results
Evidence from the literature demonstrates that children from historically marginalized backgrounds, particularly those who are publicly insured, low-income, or from racial/ethnic minority groups, experience disproportionately long wait times, lower referral completion rates, and delayed surgical care. Medicaid patients often face administrative burdens and live in regions with no pediatric otolaryngology providers. Barriers such as language discordance, transportation limitations, and low caregiver health literacy further compound disparities. Promising interventions include care coordination hubs, Medicaid fee parity, video interpreter services, mobile diagnostic vans, and one-stop weekend clinics. However, many studies are small, single-site, and lack long-term outcome tracking or demographic stratification.

Conclusion
Disparities in pediatric otolaryngology are pervasive yet modifiable. While innovative interventions have demonstrated success in improving access and timeliness of care, broader implementation and rigorous, equity-focused evaluation are needed. Future research should prioritize multi-site trials, community-informed program design, and metrics that measure not just improvement, but gap closure.		2025 AMA Research Challenge – Member Premier AccessCindy Ejindufree
8
137753Multiplex smFISH to study cell-specific RNA dysregulation in a mouse model of
Fragile X Syndrome		Background: Dysregulation of mRNA localization to neuronal dendrites and loss of neuronal activity-induced localized mRNA translation at synapses is a common cause of neurodevelopmental disorders such as Fragile X syndrome (FXS). FXS is the most prevalent cause of inherited autism and learning disability, caused by the mutation of an RNA-binding protein, Fragile X messenger ribonucleoprotein (FMRP). FXS is characterized by widespread disruptions in hippocampal memory circuitry as well as abnormalities in dendritic spine (where synapses are formed) density and morphology. FMRP binds to and regulates ~15-20% of mRNAs crucial for synaptic remodeling and neurodevelopment. It is not clearly understood how FMRP target synaptic mRNAs are regulated across different cell types in the brain that might explain the wide range of neuro-cognitive deficits in FXS. We hypothesize that, in the absence of FMRP, cell-specific dysregulation of FMRP target mRNAs underlie the wide range of functional implications in FXS pathophysiology in FMRP-null mice (Fmr1-/Y).

Methods: We have shortlisted 12 high-confidence putative FMRP-target mRNAs identified previously by high-throughput sequencing. These 12 mRNAs are localized to the dendrite-enriched hippocampal neuropil layer, are differentially expressed in hippocampal cell types and encode proteins critical for synaptic function. To examine the spatial organization of these FMRP mRNA targets in vivo, we have used highly multiplexed single molecule in situ hybridization (HiPlex smFISH) and high-resolution microscopy in juvenile WT and Fmr1-/Y mouse hippocampus. We have quantified mRNA abundance, fluorescent puncta area distribution (proxy measure of copy number state) and their co-distribution (proxy of RNA-RNA coassembly) in CA1 and CA2 neuronal dendrites, two anatomically & functionally distinct cell types relevant to FXS phenotypes. To investigate the dendritic spine abnormality in CA1 and CA2 dendrites, we used golgi-cox impregnation.

Results: Our preliminary data analyzing a subset of three FMRP-target mRNAs (Adcy1, Psd and Dlg4) in CA1 and CA2 dendrites show a trending decrease in mRNA abundance in CA1 but not in CA2 dendrites of Fmr1-/Y compared to WT. Further analysis is ongoing for the rest of the dataset. Interestingly, these findings also correlate with our ongoing analysis of spine density that shows no significant difference in CA2 dendrites (WT vs FMRP KO) and an observed increased spine density phenotype in FMRP KO CA1 dendrites compared to WT mice.

Conclusion: Integrated with the recent evidence, these data suggest that FMRP deficiency has cell-specific cellular and molecular alterations that would possibly correlate with the functional implications associated with these circuits in FXS pathology.		2025 AMA Research Challenge – Member Premier AccessRenesa Tarannumfree
9
137752Maternal Mortality in Afghanistan: A Systemic ReviewIntroduction
The Taliban’s takeover in 2021 warrants attention to Afghanistan’s compromised economy and healthcare system from a constant state of war since Russia’s invasion in 1978. Maternal Mortality (MM), as defined by the World Health Organization (WHO), is “the death of a woman while pregnant or within 42 days of termination of pregnancy” from non-accidental causes. This paper presents a literature review of MM in Afghanistan, along with interventions and prospects for the future.

Methods
A comprehensive literature search was conducted across PubMed, Google Scholar, UNICEF, and WHO to identify relevant studies and data using keywords such as "maternal mortality," "Afghanistan," "healthcare," "women’s health," and "Taliban". Studies were excluded if not available in full text or irrelevant to the topic of focus.

Results
Despite reaching a record low MM in 2023, and an overall downward MMR trend since 2000, Afghanistan’s MM still far exceeds that of neighboring countries such as Pakistan and Iran. Contributing factors were found to be destruction of economy and infrastructure from war, a woman’s education level, the presence of skilled birth attendants, and the Taliban’s stringent rules for women.

Conclusion
Multiple solutions have been implemented and are recommended. Introduced in 2003, the Basic Package of Health Services (BPHS) increased health facilities across Afghanistan in 2011 and resulted in an overall increase in health facility utilization. Maternal mortality ratios decreased significantly from 2002 to 2010, however, women in rural areas remain especially vulnerable. A program designed like the NGO named MEDlife, which recruits volunteers from Western countries to provide sustainable education and healthcare in rural, impoverished regions of Latin America, is recommended to be created and used in Afghanistan.		2025 AMA Research Challenge – Member Premier AccessZainab Bhattifree
10
136893Hormonal Modulation of Stimulant Efficacy Across the Menstrual Cycle, Pregnancy, and MenopauseAttention-deficit/hyperactivity disorder (ADHD) remains underdiagnosed in women due to the longstanding reliance on male-centered diagnostic criteria. Growing evidence now shows that hormonal fluctuations — particularly in estrogen and progesterone — influence dopamine and GABA systems, which are critical to ADHD symptom regulation and stimulant medication response.

This research explores how hormonal transitions across the female lifespan — including puberty, the menstrual cycle, pregnancy, postpartum, and menopause — can impact ADHD presentation and treatment efficacy. For instance, stimulant response tends to peak during ovulation, when estrogen and luteinizing hormone levels are highest, and diminishes during the luteal phase, when progesterone predominates. These shifts may explain why many women are diagnosed later in life, often during or after significant hormonal changes.

Research on ADHD treatment during pregnancy remains limited due to ethical constraints and the historical exclusion of pregnant individuals from clinical trials. However, existing data indicate that the primary fetal risk associated with stimulant use is a slight reduction in birth weight, with no consistent evidence of congenital abnormalities or long-term developmental harm.

Menopause may also intensify core ADHD symptoms, including emotional dysregulation and cognitive difficulties, leading to new diagnoses in adulthood.

In addition to pharmacologic strategies, Osteopathic Manipulative Treatment (OMT) may serve as a supportive adjunct by targeting autonomic balance, sleep quality, and somatic tension. These findings highlight the need for individualized, hormone-informed, and integrative approaches to ADHD care in women.		2025 AMA Research Challenge – Member Premier AccessAdunola Bellofree
11
129728Trust & Transparency in Teaching: Medical Student Attitudes Toward Educator AI UseTrust and Transparency in Teaching: Medical Student Attitudes Toward Educator AI Use

Background:
Generative artificial intelligence (AI) is rapidly transforming medical education by offering teaching and learning tools. As AI becomes integrated into academia, it is often used by educators without clear guidelines or etiquette. Understanding how key shareholders, particularly students, perceive educator AI use is increasingly important to develop best practices that center the learner. We aimed to explore student perceptions of faculty AI use to understand learner concerns and use these insights to guide recommendations for responsible faculty AI use.

Methods:
We conducted a mixed-methods study at the University of Louisville School of Medicine using surveys and semi-structured interviews to investigate student perceptions of faculty AI use in medical education. Surveys were distributed to medical students and asked respondents to rate AI use for various educational activities that have been described/recommended in the medical education literature. These results were summarized with descriptive statistics. Online, semi-structured interviews with student and faculty participants were then completed to explore the topic more in-depth. We used thematic analysis with two coders as part of a grounded theory approach to analyze transcripts with the qualitative software Dedoose.

Results:
Sixty-nine students participated in the survey, with 83% reporting prior AI use themselves. Students expressed significant concern with faculty using AI to create exam questions (56%), write clerkship evaluations (63%), and compose letters of evaluation (61%). Fifteen interviewees (ten students, five faculty) expressed overall positive sentiments about potential AI uses in medical education, but we identified the following themes regarding concerns:

1. AI as a Tool – respondents identified the potential of AI to enhance learning but stressed the importance of proper training and guidelines for faculty to ensure its effective use.
2. Transparency & Trust – lack of disclosure about AI use from faculty undermined trust between educators and students, highlighting transparency about AI use in education as a key factor to maintain and restore trust.
3. Human Touch – many feared a “loss of humanity,” especially in the realm of personalized feedback, and expressed concern that the lack of human oversight would diminish the accuracy and impact of AI-created materials.

Conclusion:
While most medical students recognized AI’s potential to enhance and personalize learning, many are also concerned that faculty may use AI in ways that they do not anticipate or agree with. Clear institutional guidelines specifying appropriate use, disclosure, and quality-control would help foster trust between educators and students. Including students in AI guideline development will be key.		2025 AMA Research Challenge – Member Premier AccessNicholas Meyerfree
12
129726A Persistent Scalp Plaque Resistant to Conventional TherapiesSarcoidosis is a systemic disease defined by the formation of non-caseating granulomas in multiple organs, with cutaneous involvement in ~25% of patients. Scalp involvement in cutaneous sarcoidosis is rare and often misdiagnosed as a more common inflammatory condition including lichen simplex chronicus, scalp psoriasis, and discoid lupus erythematosus. This case illustrates the diagnostic challenge and subsequent necessity of considering sarcoidosis in the differential diagnosis of persistent scalp plaques resistant to therapies. Only a handful of cases describing scalp involvement in sarcoidosis have been reported in the literature, magnifying a gap in clinical awareness that deserves greater attention. A 58-year-old Black female with chronic uveitis presented with a 6-month history of an intermittently pruritic, solitary plaque on the left occipital scalp. She was initially treated for presumed tinea capitis with a 4-week course of oral terbinafine, showing mild improvement. Physical exam revealed a thick, scaly plaque with whitish brown crust and no lymphadenopathy. A KOH exam was inconclusive for hyphae, so a fungal culture was obtained, and she was empirically started on another 4-week course of oral terbinafine and ketoconazole 2% shampoo, with minimal improvement. The culture was negative and she was transitioned to topical fluocinolone 0.1% oil, but at her 3-month follow-up, the disease had progressed, with new violaceous, scaly lesions at the left medial canthus and frontal scalp. A scalp biopsy showed granulomatous dermatitis favoring sarcoidosis. AFB, Fite, and Grocott stains were negative. Elevated ACE levels and imaging revealed hilar and paratracheal lymphadenopathy and breast and axillary lymph node biopsies showed noncaseating granulomas, further supporting systemic involvement. She was diagnosed with systemic sarcoidosis and treated with topical clobetasol, tacrolimus, intralesional Kenalog (10 mg/mL), minocycline, and methotrexate, resulting in significant improvement. Scalp sarcoidosis is uncommon and often mimics common inflammatory conditions such as tinea capitis and psoriasis, as seen in this case. These various clinical morphologies make prompt diagnosis difficult and, thus, appropriate treatment is often delayed. This case contributes to clinical practice by showing that clinical findings alone are insufficient for diagnosing cutaneous sarcoidosis and the value of early biopsy in persistent scalp plaques of ambiguous etiology.
2025 AMA Research Challenge – Member Premier AccessAryan Arorafree
13
129723IL-21 and IL-21R interaction plays a key role in mounting humoral and cell-mediated immune responses to adenoviral and adeno-associated viral vectors in the lungIL-21 and IL-21R interaction plays a key role in mounting humoral and cell-mediated immune response to adenoviral and adeno-associated viral vectors in lung

Authors: Mina R Hanna,1,2 Robert Clark,1 Alexander J Martin,1 Dong Lin,1 Nnamdi L Onyene,1 Xin Wang,1 and Jay Kolls1,2

1Center for Translational Research in Infection and Inflammation and 2Department of Pediatrics, Tulane School of Medicine, New Orleans, LA, USA

Purpose of Study: Although cystic fibrosis (CF) was discovered ~ a century ago, it remains without an available cure for all mutations. 8-10% of CF patients remain without therapy options despite advancements in targeting misfolded CFTR. Gene therapy is an active area of investigation in hopes of tackling this unmet need; however, the need for re-dosing remains a challenge for any means of viral vector delivery. Thus, mitigating the immune response may be the key to overcoming this challenge. Given the central role of IL-21R in regulating T and B cell immune response, we hypothesized that targeting IL-21R may attenuate B and T cell responses and enhance secondary transgene expression.

Methods: Naïve or IL21R-/- mice received 4x1010 viral particles and 1x1011 viral particles intratracheally of an Ad5 vector for first and repeat doses, respectively, at days 0 and ~27. Serum anti-vector antibodies were assayed 2 weeks post-dosing. All mice were euthanized 3 days following 2nd viral dose. Lung tissue was harvested for analysis of B cells and CD8 tissue resident memory (TRM) cells as well as secondary transgene expression. Similarly, naïve or IL21R-/- mice received 1x1011 vg of adeno-associated virus 6.2 (AAV6.2) by endotracheal intubation and monitored with In Vivo imaging system (IVIS) to assess for transgene expression.

Summary of Results: At 2 weeks, serum anti-AdV IgG (P<0.05) and IgG1 (P<0.01) were significantly decreased in the IL21R-/- mice by 2-way ANOVA. Flow cytometry analysis showed a significant reduction of CD8 TRM cells in IL-21R-/- mice (P<0.01) by nonparametric T-test and a reduction in CD19 B cells (P<0.05) by 2-way ANOVA. Moreover, secondary transgene expression was significantly enhanced in IL-21R-/- mice compared to B6 controls (P<0.01). Similarly, in the context of AAV, the generation of anti-AAV IgG requires IL-21R signaling (P<0.0001). IVIS also showed significantly enhanced secondary transgene expression of AAV6.2 in IL-21R-/- (P<0.0001).

Conclusion: Our results consistently showed how IL-21R and IL-21 play an important role in the development of humoral and CD8 TRM-mediated response to adenoviral and AAV vectors. This suggests how targeting IL-21R signaling may be the key to overcoming challenges for gene therapy using viral vectors as a means of delivery. Future studies will investigate the role of IL-21R in adult mice using antibody blockade.
2025 AMA Research Challenge – Member Premier AccessMina Hannafree
14
129721Causative Effects of a High Fat Diet on Choroidal Neovascularization: A Gut-Retina Axis Study
Abstract Title:
Causative Effects of a High Fat Diet on Choroidal Neovascularization: A Gut-Retina Axis Study
Background:
Neovascular age-related macular degeneration (nAMD) is the number one leading cause of irreversible blindness worldwide driven by its hallmark feature: choroidal neovascularization (CNV). Epidemiologic and preclinical studies suggest that high-fat Western-style diets (HFDs) exacerbate AMD progression, potentially through disruption of the gut microbiome. However, prior studies used extreme dietary models or lacked microbiota-controlled systems, limiting translational relevance and mechanistic clarity. We aimed to determine whether real-world Western-style HFDs promote CNV through gut microbiota–dependent mechanisms using a gnotobiotic mouse model.

Methods:
C57BL/6 mice (specific-pathogen-free [SPF] and germ-free [GF]) were fed either a normal diet (ND) or Western-style HFD for 8–20 weeks. A subset of SPF mice received broad-spectrum antibiotics to deplete gut microbiota. CNV was induced via sterile, laser photocoagulation. Fecal microbiota transplantation (FMT) was performed by oral gavage of GF-ND mice with stool from SPF-ND or SPF-HFD donors. Gut microbiota composition was analyzed using 16S rRNA and shotgun metagenomic sequencing. Retinal pigment epithelium (RPE)/choroid transcriptomes were profiled by RNA sequencing, and fecal metabolomics identified microbial-derived metabolites. The effects of butyrate and ursodeoxycholic acid (UDCA) on angiogenesis were tested using an ex vivo choroidal sprouting assay, and immune cell infiltration in the retina was assessed via flatmounts with immunostaining for Iba+ microglia/macrophages.

Results:
HFD-fed SPF mice developed distinct gut dysbiosis, including depletion of Bacteroidetes and enrichment of Clostridiales. These shifts correlated with transcriptional upregulation of angiogenesis and tissue remodeling genes (e.g., VEGF, Egfl7, Adam15, Sox17) in the RPE/choroid. HFD significantly increased CNV lesion size in SPF mice compared to ND (23,549 ± 2,476 µm² vs. 15,272 ± 1,413 µm²; p = 0.005). In contrast, CNV lesions and microglial infiltration were smaller in GF and microbiota-depleted SPF mice, and significantly larger in GF mice receiving FMT from HFD-fed donors. RNA-seq revealed that both HFD and gut microbiota presence altered angiogenesis-related and chromatin remodeling pathways. FMT from HFD mice reduced fecal levels of butyrate and UDCA—two metabolites known to suppress angiogenesis while ex vivo, both compounds inhibited choroidal sprouting in a dose-dependent manner (p < 0.01).

Conclusion:
Our findings establish the first causal link between Western-style high-fat diets and choroidal neovascularization, demonstrating that gut microbiota–dependent dysbiosis drives CNV by altering retinal gene expression and depleting protective microbial metabolites. This work uncovers a novel diet–microbiome–retina axis in AMD pathogenesis and, more importantly, highlights therapeutic opportunities through dietary modification, microbiome-targeted interventions, and restoration of beneficial microbial metabolites such as butyrate and UDCA.
2025 AMA Research Challenge – Member Premier AccessRobert Lucafree
15
129719Closing the Gap in CTEPH Care: A Case Review of Disparities and AdherenceBackground:
CTEPH, a subset of pulmonary hypertension, often goes undiagnosed despite its mortality rates. It arises from pulmonary artery obstruction due to unresolved thromboembolic disease. Symptoms such as dyspnea, fatigue, chest pain, and swelling necessitate a thorough evaluation. Lifelong anticoagulation is crucial to mitigate thrombotic risk. For operable cases, pulmonary endarterectomy offers a potential cure.

Case Presentation:
In August 2023, a 41-year-old male presented to the emergency department with two weeks of worsening shortness of breath, abdominal bloating, and pitting edema. His medical history included pulmonary embolism, right-sided heart failure, and acute renal failure treated with apixaban. Initial tests indicated acute kidney injury, elevated cardiac markers, and significantly increased d-dimer levels. Echocardiogram findings included severe right ventricular dilation and reduced systolic function. Imaging confirmed multiple perfusion defects consistent with CTEPH. During his hospitalization, the patient received furosemide and oxygen therapy, achieving euvolemia. Upon discharge, he continued rivaroxaban and home oxygen. Our team scheduled outpatient follow-up appointments with cardiothoracic surgery, pulmonology, a DOAC clinic, outpatient CTEPH specialists, and a primary care physician. We filled his medications at our facility and scheduled multiple follow-up appointments. Initially, the patient did not understand the gravity of his disease and expressed that his limited transportation would impact his ability to attend his follow-up appointments. We educated our patient and utilized our resources to create a well-executed discharge plan. In Spring 2024, the patient died from CTEPH complications.

Discussion:
As healthcare providers, we can commit to providing equitable care. Suppose we are vigilant in considering the potential poor outcomes that may arise from poor discharge planning and implementation. In that case, we may go that extra mile to ensure that the patient will understand the disease, diagnostics, medications, and the needed follow-up with various specialists and sub-specialists. Many doctors feel pressured to discharge patients quickly; however, safety can sometimes be compromised. Doctors may forgo parts of the discharge education, such as making follow-up appointments for patients with serious conditions, providing proper education on the importance of medication adherence, and ensuring that case management and social workers provide resources for financial assistance and medications. Discharge planning is a very dynamic and challenging aspect of inpatient care. Many factors must be coordinated to ensure the best outcomes. However, despite our efforts, our patient succumbed to his disease, highlighting the need for equitable change at the hospital and systemic level.
2025 AMA Research Challenge – Member Premier AccessMelissa Issa-Boube Kusona Fortingofree
16
129717Experience Matters…Or Does It?: Understanding Faculty Self-Confidence in Writing Narrative Assessments by Rank and RoleIntroduction
As the USMLE Step 1 exam transitioned to pass/fail, narrative assessments have become an increasingly important factor in differentiating residency candidates. Studies have consistently shown that narrative assessments often contain bias that disproportionately affect students from underrepresented backgrounds. These narrative assessments may impact competitiveness for residency positions and limit career opportunities. Few faculty development resources exist to help recognize and mitigate bias. This project aims to implement a faculty development workshop on improving the quality of narrative evaluations. A secondary objective will evaluate if differences in faculty characteristics affect perceived confidence gains after training.

Methodology
This AMA ChangeMedEd grant project utilized a faculty development session in April 2025 led by curricular leaders in the Office of Medical Student Education. There were 78 participants which included clerkship directors and clinical preceptors with previous experience in writing student evaluations.

Major components of the workshop included review of data regarding biased language in narrative evaluations, introduction of a structured framework, and interactive application exercises. A retrospective pre-post survey was offered after the session with confidence categories rated by a 5-point Likert scale. The survey remained active for ~ 1 week to allow asynchronous completion. An in-house evaluation consultancy utilized descriptive statistics to analyze average change in self-reported confidence before and after the training by role, rank, and years of experience.


Results
Of 78 participants, 41 began the survey (53% initiation rate), and 13 completed it (17% completion rate). The Dean role had the highest average years of experience in their position (10 years) and in medical education (30+ years). The Clinical course director role had the lowest average years of experience for position (4 years) and medical education (4 years). Highest confidence gains were reported in Assistant professors (+0.50), Department medical directors (+0.47), and Phase leaders (+0.46), while Deans and Associate professors had the smallest confidence change (+0.08).

Conclusion/Significance
Our results suggest that some faculty with less experience may be less anchored in routine and more open to change their evaluation writing style. Faculty with more years in medical education and position had smaller shifts, which may represent familiarity or less perceived need to change. Future directions should explore options for tailoring workshops based on participant years of experience. Identifying these strategies may help streamline workflow for busy clinicians while also maximizing competency-based narrative assessments that showcase students’ strengths.

2025 AMA Research Challenge – Member Premier AccessJocelyn Huntfree
17
129715Novel use of lumbar Ommaya reservoir for treatment of disseminated coccidiomycosis affecting the central nervous systemBackground: Our work describes the clinical course of a patient who developed disseminated coccidiomycosis infection of the central nervous system, resulting in swelling of the central cord at the level of the cervical spine. This, combined with craniocervical adhesions, was observed to block the flow of cerebrospinal fluid (CSF) such that intrathecal antifungals delivered via an intracranial Ommaya reservoir proved insufficient for the treatment of extracranial disease.

Observations: As the first-ever reported case of an Ommaya reservoir placed in the lumbar spine for treatment of central nervous system coccidiomycosis, this 48-year-old male demonstrated dramatic improvement in upper and lower extremity function after only a few weeks of intrathecal antifungal therapy. With rates of coccidiomycosis infection skyrocketing over the past decade, our clinical study demonstrates the utility of this critical method of drug delivery, previously untapped for the treatment of disseminated coccidiomycosis.

Lessons: Several months after placement, the lumbar Ommaya and catheter remain patent and functional, despite initial concern that the catheter would become clogged due to the high protein content of CSF from coccidioides infection, as well as subsequent treatment. This clinical study proves intrathecal antifungals delivered via lumbar Ommaya are a viable treatment method to help combat the increasing number of central nervous system cases.		2025 AMA Research Challenge – Member Premier AccessAarshvi Bhattfree
18
129713Reversing Sequence in Video Laryngoscopy for Small Oral Aperture: A Clinical Workaround for Challenging IntubationEffective endotracheal intubation requires both adequate glottic visualization and the physical maneuverability necessary to advance the endotracheal tube (ETT) into the trachea. In patients with anatomical limitations such as reduced oral aperture or thyromental distance, conventional laryngoscopy techniques may be rendered ineffective despite optimal visualization. This case report describes a strategic modification of standard video laryngoscopy (VL) technique to facilitate successful intubation in a patient with unanticipated difficult airway anatomy.
Case Presentation
A young, otherwise healthy female patient, approximately 5’2” tall with no significant past medical history, presented for elective outpatient carpal tunnel release under general anesthesia. Preoperative airway assessment revealed a Mallampati Class III score and reduced thyromental distance, though the airway was not formally flagged as difficult. Following standard preoxygenation, the patient was induced with intravenous anesthetics and paralyzed using rocuronium. Bag-mask ventilation was performed without difficulty. Initial intubation was attempted using a size 3 video laryngoscope (Glidescope) and a 6.5 mm cuffed ETT. A Cormack-Lehane Grade I view of the vocal cords was obtained. However, due to the patient’s small oral aperture, the ETT could not be adequately manipulated into position using conventional post-blade insertion technique. The approach was revised: the ETT was inserted first into the right buccal space while the mouth was maintained open via manual scissoring by an assistant. The VL blade was then inserted at midline, and the ETT was rotated 90 degrees to align with the airway. The tube entered the video field smoothly and was advanced through the vocal cords under direct visualization. Capnography and auscultation confirmed correct placement. A size 3 laryngeal mask airway (LMA) was prepared as a backup but not required.

The intubation was completed without hypoxia, trauma, or hemodynamic instability. The revised sequence allowed for controlled advancement of the ETT despite anatomic limitations. This approach preserved glottic visualization while accommodating mechanical challenges posed by limited oral aperture.
Discussion/Conclusion
In patients with small oral apertures and reduced thyromental distance, reversing the conventional sequence of VL—advancing the ETT prior to blade insertion—may offer a safe and effective alternative technique. When performed under appropriate conditions with adequate preparation and visualization, this method can serve as a valuable adjunct in the management of unanticipated difficult airways.
2025 AMA Research Challenge – Member Premier AccessSean Sipesfree
19
129711Cutaneous Signs as Predictors of Gastrointestinal Complications in Lysosomal Storage DiseasesBackground
Lysosomal storage diseases (LSDs) are a heterogeneous group of over 70 inherited metabolic disorders caused by deficiency of lysosomal enzymes, leading to accumulation of undegraded substrates in multiple organs. Their collective incidence is approximately 1 in 7,700 live births, affecting an estimated 1.5 million individuals globally. Clinical presentations are diverse, often involving the multiple organ systems, and in particular, the skin and gastrointestinal (GI) tract. Dermatologic signs may be visible before visceral symptoms, yet their association with GI complications is explored. Given the diagnostic delays inherent to these rare conditions and their debilitating nature, early recognition of cutaneous manifestations could prompt timely evaluation for systemic involvement.
Methods
In this retrospective cohort study, we analyzed de-identified data from TriNetX - a comprehensive electronic health records network encompassing 143 healthcare systems across 18 countries and over 163 million patient records. Patients with a diagnosis of LSD - including Fabry disease, Gaucher disease, Niemann-Pick disease, and related LSDs - were divided into two cohorts: those with cutaneous manifestations (Cohort A) and those without (Cohort B). Presence of GI outcomes including hepatosplenomegaly, diarrhea, noninfectious gastroenteritis, malabsorption, and generalized abdominal pain were measured up to 1 year after diagnosis for both cohorts. Risk ratios and survival analyses were calculated.
Results
For Cohort A (N=2,983), GI outcomes occurred in 11.1% while in Cohort B (N=14,648), GI outcomes occurred in 7.2%. The risk difference was 3.9% (95% CI: 2.7–5.1%, p < 0.001). The risk ratio was 1.54 (95% CI: 1.37–1.73). Kaplan-Meier analysis showed lower 1-year survival free of GI outcomes in the skin manifestation group (86.0% vs 92.1%, log-rank test p < 0.0001).
Ó 2025 American Medical Association. All rights reserved.
Conclusion
Cutaneous manifestations in patients with LSD were significantly associated with 54% increased risk of GI complications within one year of diagnosis. Cutaneous manifestations may reflect systemic substrate accumulation, highlighting shared pathophysiology that causes both skin and GI involvement. Early recognition of dermatologic findings - such as angiokeratomas, xanthomas, or pruritus - in patients with LSD could prompt vigilant monitoring for GI symptoms, including hepatosplenomegaly, malabsorption, and abdominal pain. For example, clinicians may consider initiating interventions like enzyme replacement therapy, substrate reduction therapy, pharmacologic therapies, or supportive measures. Ultimately, these findings suggest that dermatologic evaluation could serve as a simple, noninvasive marker for risk stratification and facilitate a proactive approach to management of multisystem disease in LSD.
Background
Lysosomal storage diseases (LSDs) are a heterogeneous group of over 70 inherited metabolic disorders caused by deficiency of lysosomal enzymes, leading to accumulation of undegraded substrates in multiple organs. Their collective incidence is approximately 1 in 7,700 live births, affecting an estimated 1.5 million individuals globally. Clinical presentations are diverse, often involving the multiple organ systems, and in particular, the skin and gastrointestinal (GI) tract. Dermatologic signs may be visible before visceral symptoms, yet their association with GI complications is explored. Given the diagnostic delays inherent to these rare conditions and their debilitating nature, early recognition of cutaneous manifestations could prompt timely evaluation for systemic involvement.
Methods
In this retrospective cohort study, we analyzed de-identified data from TriNetX - a comprehensive electronic health records network encompassing 143 healthcare systems across 18 countries and over 163 million patient records. Patients with a diagnosis of LSD - including Fabry disease, Gaucher disease, Niemann-Pick disease, and related LSDs - were divided into two cohorts: those with cutaneous manifestations (Cohort A) and those without (Cohort B). Presence of GI outcomes including hepatosplenomegaly, diarrhea, noninfectious gastroenteritis, malabsorption, and generalized abdominal pain were measured up to 1 year after diagnosis for both cohorts. Risk ratios and survival analyses were calculated.
Results
For Cohort A (N=2,983), GI outcomes occurred in 11.1% while in Cohort B (N=14,648), GI outcomes occurred in 7.2%. The risk difference was 3.9% (95% CI: 2.7–5.1%, p < 0.001). The risk ratio was 1.54 (95% CI: 1.37–1.73). Kaplan-Meier analysis showed lower 1-year survival free of GI outcomes in the skin manifestation group (86.0% vs 92.1%, log-rank test p < 0.0001).
Ó 2025 American Medical Association. All rights reserved.
Conclusion
Cutaneous manifestations in patients with LSD were significantly associated with 54% increased risk of GI complications within one year of diagnosis. Cutaneous manifestations may reflect systemic substrate accumulation, highlighting shared pathophysiology that causes both skin and GI involvement. Early recognition of dermatologic findings - such as angiokeratomas, xanthomas, or pruritus - in patients with LSD could prompt vigilant monitoring for GI symptoms, including hepatosplenomegaly, malabsorption, and abdominal pain. For example, clinicians may consider initiating interventions like enzyme replacement therapy, substrate reduction therapy, pharmacologic therapies, or supportive measures. Ultimately, these findings suggest that dermatologic evaluation could serve as a simple, noninvasive marker for risk stratification and facilitate a proactive approach to management of multisystem disease in LSD.		2025 AMA Research Challenge – Member Premier AccessCarolynne Vofree
20
129709Alternative Treatments for Non-melanoma Skin Cancers on YouTube: A Cross-
Sectional Analysis		Background
Social media plays a significant role in patient education as about 80% of US internet users obtain health information online. YouTube is a popular search engine amongst people looking for dermatologic advice. Our study assesses the content on homeopathic remedies for non-melanoma skin cancers available on YouTube.

Methods
We searched YouTube in a private browsing tab for “natural skin cancer remedies,” “alternative skin cancer treatment,” and “holistic skin cancer treatment” in separate searches. The top 40 videos meeting inclusion criteria in each search were analyzed. For data extraction, the video characteristics, engagement metrics and content themes were recorded. Duplicates were removed.

Results
61 videos were analyzed in total. 36.1% of the videos were created by homeopathic YouTube channels, 32.8% by laypeople and/or influencers, 18% by physicians and pharmacists, 6.7% by news channels and 6.7% by other personnel. The 10 most frequently mentioned remedies included green tea (19.7%), turmeric/curcumin (19.7%), coconut oil (14.8%), black salve (11.5%), apple cider vinegar (11.5%), baking soda (11.5%), garlic (11.5%), frankincense oil (9.8%), eggplant (9.8%) and milk thistle (9.8%).

Conclusion

Although most videos were created by homeopathic channels, they had the lowest engagement. Videos created by healthcare professionals achieved significantly higher engagement. Thus, even when seeking natural remedies on social media, viewers prefer content created by professionals. There is some existing literature on the role the recommended remedies play in preventing and/or treating skin cancers.

Dermatologists should be aware of the various at-home therapies patients may try for their skin cancer. They should consider creating their own reliable and accurate social media content to educate the public about the risks of these dangerous trends and emphasize the importance of seeking evaluation for suspicious skin lesions.
2025 AMA Research Challenge – Member Premier AccessUdokama Ezekwefree
21
129707Racial Disparities in End-of-Life Care Among ICU Patients: A Multi-Center Retrospective StudyRacial Disparities in End-of-Life Care Among ICU Patients: A Multi-Center Retrospective Study

Background
Palliative care consultation (PCC) improves symptom management, facilitates advance care planning (ACP), and promotes care aligned with patient values. National guidelines recommend timely inpatient PCC. However, PCC remains underutilized and is often delivered late. Racial disparities further exacerbate this; Black patients are consistently less likely than White patients to receive PCC or have documented do-not-resuscitate (DNR) orders.

Although prior studies have examined these disparities in general inpatient or terminal populations, fewer large-scale analyses have focused specifically on ICU patients with high predicted mortality, who are most likely to benefit from early end-of-life (EOL) interventions. The ICU setting, characterized by high symptom burden, complex decision-making, and elevated mortality risk, represents a critical environment. To build on existing literature and address this specific context, we conducted a retrospective cohort study to evaluate racial differences in PCC and DNR status among ICU patients using a large, multicenter database.

Methods
This study utilized the TriNetX Research Network, a database of de-identified electronic health records. Inclusion criteria consisted of patients admitted to the ICU with conditions commonly associated with higher mortality rates. Patients with prior cancer diagnoses were excluded. The primary exposure was race (Black vs White). Outcomes included receipt of any end-of-life (EOL) care encounter (defined as PCC and/or DNR order), PCC alone, and DNR status alone. Propensity score matching (1:1) was performed on covariates including age, sex, comorbidities, and hospital characteristics. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs).

Results
A total of 496,109 ICU patients met inclusion criteria, with 98,241 remaining in each group after matching.
• Black patients had a significantly lower likelihood of any EOL encounter compared to White patients (HR 0.887, 95% CI 0.863–0.911; p<0.0001).
• DNR status was also significantly less frequent among Black patients (HR 0.726, 95% CI 0.699–0.753; p<0.0001).
• Receipt of PCC was lower among Black patients (HR 0.961, 95% CI 0.935–0.988; p=0.0047).

Conclusion
Among ICU patients with high predicted mortality, Black individuals were significantly less likely than White individuals to receive EOL interventions, even after adjusting for clinical factors. These disparities highlight the need for targeted system-level strategies to ensure equitable access to goal-concordant care. Future work should explore institutional practices and structural drivers underlying these patterns and evaluate interventions to promote timely PCC in critical care settings.		2025 AMA Research Challenge – Member Premier AccessIde Kafexhiufree
22
129705Antiseptic Efficacy and Corneal Tissue Viability of Povidone-Iodine and ChlorhexidineBackground
Antisepsis prior to intravitreal injection is most commonly achieved via povidone-iodine (PI), the gold-standard; however, patients often report eye irritation. Chlorhexidine (CHX) is emerging as a potential alternative due to comparable endophthalmitis rates and less patient-reported discomfort. Limited data exist on the in vitro efficacy and corneal tissue response to PI and CHX. This study aims to compare the antimicrobial efficacy of CHX and PI against patient- derived endophthalmitis isolates as well as evaluate the corneal tissue viability after exposure to PI and CHX in an in vitro model.
Methods
The inhibition zone (mm) of PI (5%, 10%) and CHX (0.05%, 0.1%) was assessed using a well diffusion assay against eight ocular isolates (S. epidermidis, MRSA, E. coli, P. aeruginosa). The same concentrations of PI and CHX were applied to a healthy or infected (S. epidermidis) tissue model of corneal epithelial cells and cell viability was assessed with MTT assay.
Results
PI had greater inhibition for S. epidermidis and MRSA, but less for P. aeruginosa (48, 38, 19 mm), as compared to CHX (35, 29, 23 mm). Inhibition of E. coli was similar for both PI and CHX. In healthy corneal tissue model, 5% and 10% PI had less tissue viability (17%, 11%) as compared to 0.05% and 0.1% CHX (71%, 41%). Infected tissue at all concentrations of PI had less cell viability (17%, 9%) than CHX (100%, 93%).
Conclusion
While inhibition was organism dependent for both antiseptics, CHX preserved higher tissue viability than PI in both infected and noninfected models. CHX shows promise as an alternative for ocular antisepsis against endophthalmitis isolates.		2025 AMA Research Challenge – Member Premier AccessAnam Ahmedfree
23
129703Subclinical Elevations of C5DC in Puerto Rican Children with Neurodevelopmental Disorders: Exploring a Potential Metabolic Signature of Neurogenetic RiskAuthors:
Paola P. Colón Terrón, B.S. 1,3, Linda Pérez, , DrPH, MSc 1,3, Elinette M. Albino, , MT(ASCP), PhD, MS 2,3
1 San Juan Bautista School of Medicine 2 School of Health Professions, University of Puerto Rico Medical Science Campus 3 Puerto Rico Genetics Laboratory

Abstract Title:
Subclinical Elevations of C5DC in Puerto Rican Children with Neurodevelopmental Disorders: Exploring a Potential Metabolic Signature of Neurogenetic Risk

Background:
Glutarylcarnitine (C5DC) is a dicarboxylic acylcarnitine involved in the catabolism of lysine and tryptophan. Its accumulation is typically associated with glutaric acidemia type I (GA1), a rare metabolic disorder. However, in Puerto Rico, a population with a high burden of rare and recessive genetic conditions, C5DC levels have been observed to be consistently elevated in children with neurodevelopmental symptoms, though not reaching the threshold required for a GA1 diagnosis. This pattern raises the possibility that subclinical elevations of C5DC may reflect
an underlying metabolic signature associated with broader neurogenetic risk.

Methods:
We conducted a retrospective analysis of acylcarnitine profiles in a cohort of Puerto Rican pediatric patients (n = 98) with clinical features suggestive of neurogenetic disorders. C5DC levels were measured via tandem mass spectrometry and correlated with clinical phenotypes, neuroimaging findings, and genetic testing results.

Results:
LASSO linear regression with 10-fold cross-validation identified seventeen phenotypes associated with C5DC levels: Bell’s palsy, behavioral delays, encephalopathy, gender, aggressive behavior, hearing loss, hypotonia, epilepsy, developmental disorder of speech, sensory issues, intellectual disability, Charcot-Marie-Tooth disease, and autism spectrum disorder, among others. Post-LASSO linear regression analysis showed statistically significant associations between C5DC levels and behavioral delays (p < 0.010), aggressive behavior (p < 0.05), hearing loss (p < 0.05), epilepsy (p < 0.05), and sensory issues (p < 0.05). Notably, the majority of cases were clustered in the northern region of Puerto Rico, with the top five municipalities being San Juan, Vega Baja, Carolina, Bayamon, and Toa Alta.

Conclusion:
These preliminary findings suggest that subclinical elevations of C5DC may be indicative of neurogenetic vulnerability in Puerto Rican children. C5DC may serve as a metabolic biomarker worthy of further investigation for inclusion in neurogenetic screening panels, especially in genetically isolated populations with known founder effects and high consanguinity. Continued research is needed to elucidate the mechanistic basis of this association and its diagnostic potential.		2025 AMA Research Challenge – Member Premier AccessPaola Colón-Terrónfree
24
129702Laryngeal Mass as a Presentation of Disseminated Coccidioidomycosis: Case Report and Literature ReviewEducational Objective: At the conclusion of this presentation,
participants should be able to diagnose and treat
coccidioidomycosis in the larynx.​
Objective: To present a rare case of coccidioidomycosis
manifesting as a laryngeal mass and to explore the key clinical
features, diagnostic challenges, and treatment considerations
that can aid in the timely recognition or exclusion of the disease in
future cases. We also review the existing literature of
disseminated coccidioidomycosis to contextualize this case.​
Study Design: A single case report of a patient with
disseminated coccidioidomycosis is described. ​
Methods: A case report of a 26-year-old female with a history of
persistent cough and hoarseness. She presents with persistent
throat issues and voice changes. The patient was initially
prescribed antibiotics, but they were ineffective in resolving her
loss of voice. At presentation, a lesion on the right arytenoid was
found after laryngoscopy. ​
Results: Patient underwent a biopsy. Final pathology revealed
coccidiomycosis in the right supraglottic and right arytenoid
mass. ​
Conclusion: This case report highlights the importance of
including Disseminated Coccidiomycosis in the differential
diagnosis of patients with persistent hoarseness and cough,
especially in endemic regions. ​		2025 AMA Research Challenge – Member Premier AccessBenjamin Ngfree
25
129700Refractory gastroparesis in a patient with Ehlers-Danlos Syndrome: Symptomatic improvement following G-POEMBackground
Ehlers-Danlos Syndrome (EDS) is a group of connective tissue disorders often associated with hyperflexibility, dysautonomia, and gastrointestinal (GI) dysmotility. Although gastroparesis related to EDS is often underrecognized, it can lead to significant morbidity and impacts on quality of life when refractory to many medical therapies. We present a case of medically refractory gastroparesis in a young female patient with hypermobile EDS (hEDS; EDS Type 3), gastroparesis, sick sinus syndrome (SSS), and postural orthostatic hypotension (POTS) who experienced clinical improvement following gastric per-oral endoscopic myotomy (G-POEM).
Case Presentation
A 23-year-old woman with hEDS, gastroparesis with a gastrojejunal feeding tube, POTS, and SSS with a pacemaker presented with intractable nausea and vomiting. She had a prior diagnosis of gastroparesis confirmed by a gastric emptying study. She had undergone extensive outpatient and inpatient trials of both standard and atypical antiemetics and prokinetics without meaningful relief. She was hospitalized for nearly a month for symptomatic and nutritional support, with no clear cause on imaging.
During admission, a PICC line was placed, and she was started on TPN for nutritional support due to inability to tolerate gastrojejunal (GJ) tube feeds. Additionally, she received a gastric pylorus botulinum toxin injection, which provided minimal relief. The patient was ultimately discharged on TPN with outpatient follow up with GI for G-POEM procedure.
The patient underwent an uncomplicated G-POEM procedure. Within one month, the patient reported improvement in her nausea and oral intake and was weaning herself off TPN while gaining weight.
Discussion
Gastroparesis in patients with hEDS is likely multifactorial, involving autonomic dysfunction, visceral smooth muscle impairment, and connective tissue abnormalities. In this case, the patient received extensive pharmacological therapies and pyloric botulinum toxin without substantial benefit, ultimately requiring TPN.
G-POEM, a minimally invasive procedure that cuts the pyloric sphincter muscle, is gaining popularity as a treatment for refractory gastroparesis. Although its role in hEDS is still under investigation, this case highlights its potential benefit, especially in cases where patients fail medical management and require nutritional support
2025 AMA Research Challenge – Member Premier AccessChase Moscovicfree
26
129698Impact of OBGYN Involvement on Health/Reproductive Outcomes in Childbearing Age Women Undergoing Bariatric SurgeryTitle: Impact of OBGYN Involvement on Reproductive and Metabolic Outcomes in Women of Childbearing Age Undergoing Bariatric Surgery
Authors: Nneka Ajah, MD, MBA; Antonia F. Oladipo, MD MSCI; Pooja Bhate; Sanjana Sharma, MS; Christina O’Leary, MS; Erica Amianda, PA; Themba Nyirenda, PHD; Burton Surick, MD
Background:
Bariatric surgery is an increasingly common intervention for obesity in reproductive-aged women. However, the rapid improvement in fertility following surgery, combined with hormonal and metabolic shifts, presents significant reproductive health risks if not appropriately managed. Despite ACOG guidelines recommending contraception and pregnancy delay post-surgery, the role of routine OBGYN involvement in this population remains underexplored.
Objective:
To evaluate how OBGYN involvement impacts reproductive planning, contraception use, and metabolic outcomes among women aged 18–45 undergoing bariatric surgery.
Methods:
This IRB approved retrospective chart review included 454 women who underwent primary sleeve gastrectomy at Hackensack University Medical Center from January 2021 to December 2023. Data collected included BMI, comorbidities, PCOS symptoms, HbA1c, number of hypertension medications, contraceptive use, OBGYN follow-up, and pregnancy outcomes. Comparative statistics assessed associations between OBGYN involvement and clinical outcomes.
Results:
Patients experienced significant postoperative BMI reduction (mean decrease: 13.63 points, p<0.001). Among patients with PCOS (n=54), 18.5% reported improved menstrual regularity, and 20.4% improved glycemic control post op. Only 29.3% had documented contraception, but uptake was significantly higher among those with documented OBGYN follow-up (p=0.02). Pregnancy occurred in 17.4% of patients within three years.
Conclusion:
OBGYN involvement in postoperative care is associated with improved contraceptive use and better documentation of reproductive health outcomes in women undergoing bariatric surgery. These findings support a multidisciplinary approach to optimize care in this high-risk population and highlight an opportunity to strengthen reproductive counseling during metabolic health interventions.		2025 AMA Research Challenge – Member Premier AccessNneka Ajahfree
27
129695A rare case of hemorrhagic cardiac tamponade after electric cardioversion in an anticoagulated patientBackground: Bleeding events post electric cardioversion in anticoagulated patients are rarely reported. ¹⁻² Current guidelines recommend anticoagulation for at least 3 weeks prior to and 4 weeks after cardioversion for atrial fibrillation of >48-hour duration. ¹ Our case highlights a rare occurrence of cardiac tamponade in an anticoagulated patient who underwent elective cardioversion.

Case presentation: A 73-year-old man with ALS-related chronic respiratory failure and ventilator dependence presented from his facility with sudden-onset palpitations. Other relevant medical history includes prior pulmonary embolism (on reduced dose apixaban) and depression. Initial EKG revealed atrial flutter with rapid ventricular rate (RVR) and 2:1 conduction pattern. He underwent successful electrical cardioversion after inadequate response to medical management.

Within hours of the procedure, the patient developed hypotension accompanied by a notable decline in hemoglobin levels. A transthoracic echocardiogram demonstrated a moderate-sized pericardial effusion with early signs of tamponade. An emergent pericardiocentesis was performed, yielding 430 mL of hemorrhagic fluid. Cytological studies showed no evidence of malignancy. Hemodynamic stability was restored following the procedure.

A multidisciplinary discussion held after this event led to the permanent discontinuation of anticoagulants from the patient’s medication regimen.

Discussion: Multiple trials present evidence of safety and efficacy (in preventing thromboembolic events) of DOACs and VKAs for peri-cardioversion anticoagulation. ¹ Our case represents a rare but life-threatening complication of cardiac tamponade following cardioversion. There is very sparse literature describing similar cases; one case report by Jessuran et al describes tamponade following electrical cardioversion, though asbestos-related subclinical pericarditis served as a significant confounding factor in their case. ³

Clinicians must be vigilant of subtle risk factors for hemorrhagic complications, which may not be fully captured by the current guidelines. Individual risk assessments with careful monitoring and/or stopping anticoagulation indefinitely can prevent unfavorable outcomes in such patients.		2025 AMA Research Challenge – Member Premier AccessHaripriya Radhakrishnanfree
28
129693Post-Anesthesia Behavioral Changes in Children: A Comparison Between English and Spanish-Speaking FamiliesBackground:
Children who undergo procedures requiring general anesthesia can experience stress, anxiety, and behavioral changes both before and after the procedure. Prior research has shown that pediatric patients can undergo changes in their behavior post anesthesia, however less information has been published about the duration, types, and predictors of changes seen. Furthermore, little is known about the effect of language spoken at home on these changes. This study aims to compare native English speakers' behavior post anesthesia to native Spanish speakers' experience in a pediatric population and describe the varying behaviors post-anesthesia.

Methods:
A telephone survey will be administered with both English and Spanish speaking guardians of the patients aged 7-12 post-discharge after a scheduled procedure with anesthesia. The survey will be given at 2 weeks and 4 weeks post-procedure and will ask the guardians of the patients a number of questions surrounding behavioral changes (diet, activity, sleep, anger,dependence, etc.) of the child since anesthesia. The data will then be analyzed to compare English to Spanish speaking behavioral differences after anesthesia. The survey tool is the Post Hospitalization Behavior Questionnaire for Ambulatory Surgery 9. The Post Hospitalization
Behavior Questionnaire for Ambulatory Surgery (PHBQ-AS) is a 1-1 item Liken-style tool validated in parents (80% non-Hispanic White) of children 2 through 12 years of age 9. It was found reliable in the rating of child behavior change after outpatient surgery with general anesthesia (Cronbach alpha 0.80). Additional participant data will be obtained through medical record review and will include procedure type, type and duration of anesthesia, pre-anesthesia medications, and demographic data.

Results:
We anticipate that (1) children who undergo a procedure requiring anesthesia will have significant behavior changes and that (2) children in Spanish-speaking families will have comparable behavior changes to English native speakers. We expect this to be supported objectively by similar responses on our telephone survey at both timeframes.

Conclusion:
This study entertains the idea that health professionals may be able to give parents and their children better anticipatory guidance about the peri-anesthesia period and provide ways to prevent these behaviors with prior knowledge.		2025 AMA Research Challenge – Member Premier AccessJessica Ericsonfree
29
129692Evaluating the Impact of Smoking on Early Dry Age-related Macular Degeneration Outcome: A Large-Scale Matched Cohort AnalysisIntroduction: Smoking is a major modifiable risk factor for age-related macular degeneration (AMD), but its influence on disease progression after early dry AMD is uncertain. The purpose of this study is to quantify the association between smoking and progression from early or intermediate nonexudative AMD to (1) advanced nonexudative AMD and (2) exudative AMD.
Method: Retrospective cohort study using de-identified electronic health records from the TriNetX US Collaborative Network (68 health care organizations). Smokers were identified by nicotine dependence ICD 10 codes or varenicline prescriptions; nonsmokers had no such records. Propensity score matching (1:1) for age, sex, race/ethnicity, diabetes, hypertension, and obesity yielded balanced cohorts of 8,978 smokers and 8,978 nonsmokers—incident diagnosis of advanced nonexudative AMD or exudative AMD captured by ICD-10 codes. Risk ratios (RRs) with 95% confidence intervals (CIs) were calculated.
Results: Among 22 490 adults aged ≥50 years with early or intermediate non exudative AMD (11,225 smokers [mean (SD) age, 73.1 (8.9) years; 52.5% female] and 11,265 nonsmokers [73.1 (9.1) years; 52.8% female]), smoking was associated with a higher risk of progression to advanced non exudative AMD (580/11225 [5.2%] vs 486/11265 [4.3%]; risk difference, 0.9%; RR 1.20; 95% CI, 1.07 1.35). Among 19,766 patients without prior exudative AMD (9,702 smokers, 10,064 nonsmokers), progression to exudative AMD occurred in 710/9702 (7.3%) vs 678/10,064 (6.7%); risk difference, 0.6%; RR 1.09; 95% CI, 0.98- 1.20.
Conclusions: Smoking was independently associated with a modest but statistically significant 20% relative increase in risk of progression to advanced nonexudative AMD, whereas its association with exudative AMD progression was not substantial. These findings reinforce smoking cessation as a key preventive strategy for patients with early dry AMD and refine risk stratification in clinical practice.
2025 AMA Research Challenge – Member Premier AccessKimia Rezaeifree
30
129690Altitude and risk of pre-eclampsia: insights from a large-scale epidemiological study in EcuadorPre-eclampsia affects 2–8% of pregnancies globally and remains a major contributor to maternal morbidity and mortality, particularly in low- and middle-income countries where healthcare access and resources are limited. The condition is characterized by defective extravillous trophoblast invasion into the maternal spiral arteries, leading to impaired placental perfusion and systemic endothelial dysfunction. Established risk factors include nulliparity, chronic hypertension, extremes of maternal age, obesity, and low socioeconomic status. Exposure to high-altitude has also been proposed as a potential risk factor, though findings across regions have been inconsistent and underexplored in population-level studies. This study aimed to evaluate the association between altitude—treated as a continuous variable—and the risk of pre-eclampsia in Ecuador, while accounting for sociodemographic and healthcare-related factors.

We conducted a nationwide epidemiological study using publicly available maternal health records from Ecuador’s National Institute of Statistics and Census (INEC) from 2021 to 2023. Altitudes were geospatially mapped to maternal residential addresses and verified against national elevation data. Generalized additive models (GAMs) with cubic splines were used to assess potential non-linear associations between altitude and pre-eclampsia risk. Models were fitted using Restricted Maximum Likelihood (REML) and adjusted for key confounders including maternal age, ethnicity, funding model, type of care, and residential setting.

In adjusted models, altitude was not significantly associated with an increased risk of pre-eclampsia. However, pre-eclampsia risk was significantly elevated among women at the extremes of maternal age (p<0.001) and among those receiving publicly funded care (p<0.001), underscoring the influence of socioeconomic disparities. Ethnicity also emerged as a significant determinant of pre-eclampsia risk, though no interaction effect was observed between ethnicity and altitude.

These findings suggest that in Ecuador, altitude alone may not be an independent driver of pre-eclampsia risk. The absence of a measurable altitude effect may reflect unique demographic, genetic, and environmental characteristics specific to the region. Given the persistent burden of pre-eclampsia and its ties to social determinants of health, further multi-regional studies incorporating both environmental and socioeconomic factors are needed to clarify contextual risk drivers. Such evidence can inform targeted screening strategies and guide more equitable allocation of maternal health resources, especially in geographically and economically diverse settings.

2025 AMA Research Challenge – Member Premier AccessEmma Mitchell-Sparkefree
31
129688Testosterone Therapy in a Transgender Man with C282Y Hemochromatosis, Heavy Menses, and Prior Iron Supplementation: A Hypothetical Case StudyBackground
Hereditary Hemochromatosis (HH), particularly due to HFE C282Y homozygosity, leads to progressive iron overload due to increased intestinal absorption. In individuals assigned female at birth, cyclic menstrual blood loss after onset of menses may delay the clinical manifestation of hereditary hemochromatosis. In cisgender males, however, this manifests earlier and more severely due to the absence of any mechanisms of blood loss.

This case review presents the hypothetical case of Finley, a 30-year-old transmasculine patient with a history of heavy menstrual bleeding and iron supplementation, who rapidly develops clinical iron overload after initiating testosterone therapy. The case examines how the interaction between testosterone, erythropoietin, and iron metabolism can unmask a latent genetic condition. Additionally, it highlights the need for proper history taking before the initiation of gender-affirming hormone therapy, proactive iron monitoring and supplementation reassessment during gender-affirming care, especially in patients with known or potential genetic risk.

Methods
Finley is homozygous for the HFE C282Y mutation but does not have a diagnosis of hemochromatosis due to lack of clinical presentation. His ongoing heavy menstrual blood loss has functioned as a natural compensatory mechanism, keeping serum iron and ferritin levels low despite his genetic predisposition to iron overload.

Finley visits an endocrinologist to initiate gender affirming care and begin testosterone therapy in his transition to a transgender man. Upon standard medical history taking, his endocrinologist inquires about any prescription medications to which Finley denies. Finley does not disclose his use of over-the-counter iron supplements as he does not perceive this information to be clinically significant.

Finley initiates medical transition through testosterone therapy of 100 mg IM administered once a week. Follow-ups are scheduled with his endocrinologist for 3, 6 and 10 months.

Results
Finley’s blood panels at his 10-month follow-up visit show ferritin, TSAT, hematocrit, ALT and AST values elevated well above the reference ranges. The hepatic ultrasound early signs of hepatic steatosis. DNA analysis of the HFE gene showed HFE C282Y homozygosity, consistent with a diagnosis of HH.

Conclusion
This scenario emphasizes the critical need for comprehensive pre-treatment evaluation in transgender patients initiating hormonal therapy, including detailed menstrual, supplementation, and family histories. Standard history-taking procedures may fail to elicit pertinent information, such as long-term iron supplementation, emphasizing the need for specific and directed questioning. Routine monitoring of iron indices and liver function should be considered, particularly in patients with known risk factors or suggestive clinical histories.		2025 AMA Research Challenge – Member Premier AccessNia Prabhufree
32
129686Late-Onset Pseudomyxoma Peritonei: An Unusual Recurrence 8 Years After AppendectomyBackground

Pseudomyxoma peritonei (PMP) is a rare, clinically aggressive neoplasm characterized by mucinous peritoneal implants, with an annual incidence of 1-2 cases per million. Prognosis is dominated by histopathologic subtype. Without intervention, the median survival rate is 3.2 years, with intestinal failure being the predominant cause of death. Diagnostic challenges persist due to non-specific symptomatology and insidious onset, compounded by frequent misclassification in up to 38% of cases.

Case Presentation

A 48-year-old male with a 2015 appendectomy for low-grade mucinous carcinoma (LAMN), also named disseminated peritoneal adenomucinosis (DPAM). Developed voluminous ascites eight years later. Peritoneal cytology suggested metastatic colorectal adenocarcinoma (IHC: CK20+/CDX2+/SATB2+). CT confirmed isolated peritoneal carcinomatosis with elevated CEA (186 ng/mL). Histopathology reclassified the disease as Grade 1 PMP (DPAM subtype). Neoadjuvant FOLFOX (6 cycles) yielded no response (RECIST 1.1), precluding cytoreduction. Transition to FOLFIRI achieved symptomatic resolution, CEA normalization (8.2 ng/mL), and 9-month radiographic stability.

Discussion

Initial misclassification as colorectal adenocarcinoma (or ovarian cancer) underscores pseudomyxoma peritonei's (PMP) diagnostic complexity. Reported in similar cases as this one. While CK20/CDX2/SATB2 positivity overlaps with gastrointestinal malignancies, PMP's distinct molecular signature (84% KRAS mutations vs. 45% in colorectal cancer) and indolent clinical behavior mandate expert pathology review using Ronnett criteria to prevent therapeutic misdirection. Additionally, molecular profiling (KRAS/MGMT status) resolves diagnostic ambiguity

Late Recurrence & Surveillance: This recurrence at 8 years exceeds standard 5-year surveillance for LAMN, aligning with evidence that 2–3% of DPAM cases recur >10 years post-resection. This Risk correlates with initial perforation and positive surgical margins at initial resection. Symptoms manifesting at 8 years underscore the need for extended monitoring (CT + CEA/CA19-9 beyond 5 years), which challenges NCCN guidelines that recommend a 5-year follow-up.

Prognostic & Therapeutic Implications: Histologic grade is the dominant survival predictor, with peritoneal mucinous carcinomatosis (PMCA) conferring 4.2× higher mortality risk vs. DPAM. While Completeness of Cytoreduction (CCR)-0/1 resection + Hyperthermic Intraperitoneal Chemotherapy (HIPEC) is ideal, systemic therapy (e.g., FOLFIRI) achieves durable disease control in CCR-unfit patients, preserving quality of life in 68% of long-term survivors

The observed FOLFOX resistance but FOLFIRI response aligns with mucinous tumor biology. Oxaliplatin-based regimens demonstrate limited efficacy in PMP (objective response rate [ORR]: 20–30%), whereas irinotecan achieves 60–70% disease control through topoisomerase inhibition and anti-angiogenic effects. For unresectable DPAM, FOLFIRI represents a validated alternative, with a median progression-free survival (PFS) of 8.1 months.

This research redefines PMP as a paradigm-shifting model, where extended surveillance prevents missed recurrences, molecular stratification overcomes diagnostic mimicry, and tailored chemotherapy rescues surgically frail patients.
2025 AMA Research Challenge – Member Premier AccessAlexander Manzanofree
33
129684Suicide in Haiti: Perceptions, Associated Factors, and Research Gaps: a Narrative ReviewAbstract Title
Suicide in Haiti: Perceptions, Associated Factors, and Research Gaps: a Narrative Review

Background
Suicide is a major global health concern, disproportionately affecting low- and middle-income countries (LMICs) where research remains limited. Although global suicide rates have declined
in recent decades, rates in Haiti have changed minimally. Haiti's sociocultural complexity and limited mortality data obscure understanding of suicide risk factors and perceptions. Known
global risk factors—such as depression, hopelessness, trauma, and stigma—may interact differently in local contexts. This review examines the current state of empirical research on
suicide in Haiti, exploring both perceptions and psychosocial risk factors, to inform future prevention strategies.

Methods
A systematic search of 15 databases identified peer-reviewed, English-language studies containing data on suicide or suicidal ideation in Haiti. Search terms included variations of
"Haiti" and "suicide." Studies were included if they reported Haiti-specific data on suicide and met inclusion criteria. Data extracted included participant demographics, study region, methods, and findings on suicide-related perceptions or risk factors.

Results
Ten studies met inclusion criteria. Samples included rural and urban populations, adolescents, adults, pregnant individuals, health care workers, and those with psychiatric or medical
diagnoses. Four studies used qualitative methods; of the seven studies that included quantitative methods, four used data from the Violence Against Children in Haiti survey and
three used culturally adapted Beck Depression Inventory surveys. Common themes included the cultural idiom reflechi twòp ("thinking too much"), which was strongly associated with suicidal ideation, namely 7.7 times more likely (95% CI, 1.03–58.04). Community members often cited poverty, shame, and Voodoo-related beliefs as causes of suicide, while health care workers favored psychiatric explanations. Perceived shame from unplanned pregnancy, HIV status, and sexual orientation were also prominent. Suicidal ideation was associated with depression, as was social support absence, alcohol use, and experiences of emotional, physical, and sexual violence -- females who experienced all three forms of violence had the highest odds of suicidal ideation (adjusted OR, 15.09; 95% CI, 6.10–37.33, p < 0.001). Protective factors were rare: one was the belief that women should tolerate violence to preserve family unity.

Conclusion
Empirical studies on suicide in Haiti are limited and regionally narrow. Cultural and social perceptions—such as reflechi twòp and stigma related to sexuality, illness, and gender—shape
suicide risk in ways not captured by universal models. Integrating culturally specific constructs and community support systems into screening, safety planning, and interventions is essential. Further research is needed across diverse Haitian populations and languages to build effective, context-sensitive suicide prevention strategies.		2025 AMA Research Challenge – Member Premier AccessConnor Orricofree
34
129682Cancer Associated Retinopathy Presenting with Paraneoplastic Cerebellar Ataxia in Tonsillar Squamous Cell CarcinomaAbstract Title
Cancer Associated Retinopathy Presenting with Paraneoplastic Cerebellar Ataxia in Tonsillar Squamous Cell Carcinoma

Background
Cancer associated retinopathy (CAR) is a rare paraneoplastic autoimmune retinopathy caused by antibodies against tumor antigens that cross-react with retinal proteins. Patients experience sudden or subacute painless vision loss, often associated with photopsias, nyctolopia, scotomas, or constriction of the visual field. Similarly, paraneoplastic cerebellar degeneration (PCD) is a rare paraneoplastic autoimmune condition driven by antibodies directed against cerebellar antigens initiated by onconeural proteins, causing a progressive truncal cerebellar syndrome. Here, we present a patient with a poorly differentiated squamous cell carcinoma of tonsil who manifested simultaneously features of PCD and CAR, two conditions that have previously not been reported to occur together.

Case Presentation
A 66-years old man presented with progressive, painless bilateral peripheral and night vision loss, alongside truncal ataxia, dysmetria in the extremities, and dysarthria over several months. Optical coherence tomography showed diffuse outer retinal loss and depressed multifocal electroretinogram (mfERG) in both eyes. Visual acuity further declined over the next two months. Concomitantly, a poorly differentiated squamous cell carcinoma was identified in the right tonsil. CAR panel showed strong diffuse staining on immunohistochemistry with elevated titers for recoverin, enolase and carbonic anhydrase II antibodies. Magnetic resonance imaging showed cerebellar atrophy consistent with PCD, but screening paraneoplastic antibodies linked to cerebellar ataxia were negative. Robotic right tonsillectomy with neck dissection was performed and the patient underwent immunotherapy with intravenous immunoglobulin (IVIG) and high dose corticosteroids. Six months later, the patient demonstrated no recurrence of his cancer with improvements in his visual acuity, retinal structure, and mfERG. His ataxia also improved. One year later, rituximab therapy was initiated. Now five years later, there are sustained improvements in vision and gait.

Discussion
We highlight a case of CAR that coincided with cerebellar ataxia caused by seronegative PCD that responded well to treatment of the underlying squamous cell carcinoma of the tonsil, followed by corticosteroid therapy, IVIG, and rituximab therapy. In this case, several autoantibodies known to cause CAR were identified while no autoantibodies linked to PCD were detected. Up to a third of patients with PCD (and CAR) are seronegative for known antibodies. Given the wide range of aberrant proteins expressed by poorly differentiated malignant cells, patients often develop circulating autoantibodies to more than one target. Although no autoantibodies are known to link PCD to CAR, a yet unidentified cross-reactive antibody may explain the overlapping syndrome seen in our case.		2025 AMA Research Challenge – Member Premier AccessYeseo Kwonfree
35
129680Internal Herniation and Bowel Ischemia in Pregnancy Following Roux-en-Y Gastric Bypass: A Case Report and Review of Obstetric ImplicationsBackground: As bariatric surgery becomes increasingly prevalent among women of reproductive age, obstetric care providers are more frequently encountering the unique complications these procedures pose during pregnancy. Roux-en-Y gastric bypass, one of the most common bariatric surgeries, can lead to serious intrapartum complications such as internal herniation and bowel ischemia, which are often difficult to diagnose due to symptoms similar to routine pregnancy related complications, other nonspecific symptoms, and physician hesitance to using imaging modalities with potential X-Ray exposure to the unborn fetus. Finally, laboratory data as well as imaging results are often inconclusive in the pregnant state.
Case Presentation: We present the case of a 30-year-old gravida 6 para 1031 at 25 weeks and 4 days gestation, with a remote history of Roux-en-Y gastric bypass, who presented twice to the hospital for epigastric abdominal pain. In her first hospital evaluation she had an unclear diagnostic picture and was sent home to follow up with routine obstetrical care and gastroenterology related issues. She then returned to the hospital almost two months later with acute worsening of her abdominal pain and ultimately required an emergent Cesarean delivery due to signs of fetal distress. Intraoperative findings revealed extensive small bowel volvulus and internal hernia with necrotic bowel requiring resection and gastric bypass reversal. Despite initial benign imaging and stable examinations, the patient ultimately still ended up developing septic shock and required multiple surgical interventions.
Conclusion: This case highlights the potentially fatal complications of Roux-en-Y gastric bypass during pregnancy and the diagnostic challenges they pose. It underscores the importance of maintaining a high index of suspicion in pregnant patients with a history of bariatric surgery who present with abdominal pain. Prompt recognition, early surgical consultation, and a multidisciplinary approach are critical to reducing maternal and fetal morbidity and mortality.		2025 AMA Research Challenge – Member Premier AccessCasey Easterlingfree
36
129678The Impacts of Climate Change on Neurological Health, Where to Go From Here: Systematic Review and Advocacy ToolBackground
The impacts of climate change are experienced in the community and in clinical practice. This study combines interdisciplinary climate change research with advocacy strategies to address the evolving challenges imposed by climate change upon patients and practitioners. A systematic review of the neurological literature related to climate change was conducted and findings were compiled into examples of patient health impacts. In addition, examples are provided of ways practitioners can adapt their practices to address the impacts of climate change on patients and to reduce environmental harm in the clinic and operating room.

Methods
Literature pertaining to the neurological effects of climate change on patients and practices was reviewed. Following PRISMA guidelines, search parameters for the systematic review were identified and relevant studies were sorted for inclusion based on selection criteria. Studies sought discussed neurological disease variance due to climate change that may require neurological consultation, or directly identified information relevant to sustainable practices. Two tables were constructed presenting (1) specific patient health implications of climate change on neurological illness, and (2) strategies for clinical practices to prepare for and address these emerging sequelae.

Results
A total of 421 studies were initially obtained via the search criteria and 84 were removed as duplicates. After abstract screening of 337 studies, the full text of 95 studies were reviewed. 70 were relevant to neurological disease impacted by climate change, and 20 studies were deemed relevant to sustainable practices. The neurological impacts of climate change on patients and practices include cerebrovascular disease (increased stroke, aneurysmal subarachnoid hemorrhage, and heatstroke), neurodegeneration (increased hospitalizations, symptom exacerbation, and mortality), CNS infectious diseases (viral, bacterial, fungal, parasitic, prion, and amoeba), trauma (extreme weather events and violence), and environmental exposures (birth defects, contamination, brain tumorigenesis, and seizure trigger).

Conclusion
Through sustainable clinical practices and education/training the healthcare field can prepare patients and practitioners for these impacts and rise to address the challenges facing our communities in the 21st century.		2025 AMA Research Challenge – Member Premier AccessRyan Neillfree
37
129676Empagliflozin and liver fat reduction in T2DM & NAFLDBackground
Non-alcoholic fatty liver disease (NAFLD) and type 2 Diabetes Mellitus are the major causes of significant mortality and morbidity. NAFLD and type 2 diabetes have common risk factors. SGLT-2 inhibitors are believed to have a beneficial effect on liver fat in patients with NAFLD and type 2 diabetes. However, there are limited data available on this aspect of empagliflozin use. Thus, we aimed to compare the impact of empagliflozin on liver fat reduction in type 2 diabetes and NAFLD.
Methods
A comprehensive search of databases like PubMed, Google Scholar, and Cochrane Library was done. In the end, a detailed analysis was performed on a total of 6 finalized studies, all of which were randomized controlled trials. This comprehensive analysis consisted of a cohort of 806 Individuals.
Results
Participants in the Empagliflozin group had an average change of -29.57 (SD = 19), while those in the placebo group showed a smaller improvement of -16.41 (SD = 24.1). The standardized mean difference (SMD) was -0.60, with a 95% CI, Z = 2.48, P = 0.01. AST studies showed a statistically significant benefit of empagliflozin over placebo with an SMD of -0.41 (95% CI: -0.66 to -0.16, p = 0.001). Heterogeneity among the studies was low (I² = 34%). The SMD is -0.43 with a 95% CI and a p-value of 0.03, indicating a statistically significant effect favoring empagliflozin. The heterogeneity is low (I² = 0%, P = 0.37). The results for 30 % liver fat reduction show an odds ratio of 14.08 with a 95% CI of 2.74 to 72.48. The test for overall effect is statistically significant (p = 0.002).
Conclusion
The analysis found that with the use of empagliflozin, different variables significantly decreased. These include Liver Steatosis (CAP) Score, AST, MRI-PDFF, 30% LVEF Reduction, and liver fat content (LFC).
Keywords:
Empagliflozin, NAFLD, Type 2 diabetes, liver fat reduction
2025 AMA Research Challenge – Member Premier AccessMuhammad Luqmanfree
38
129674OrthoScreen: Developing a Transformer-Based Model to Predict Surgical Need from Imaging ReportsIntroduction:
Only 23% of musculoskeletal referrals result in surgery, highlighting the need for decision support tools that assist providers in interpreting imaging when orthopaedic input is not readily available. Primary care providers often navigate complex radiology narratives without subspecialty guidance, leading to variability in referral quality. Some patients are referred prematurely, while others face delays despite surgical need. We developed OrthoScreen, a transformer-based NLP model built to support referral prioritization using solely imaging text, for potential application in arthroplasty triage. We hypothesized that OrthoScreen would demonstrate high sensitivity in identifying patients likely to be considered for surgery within 1 year, based solely on imaging reports.

Methods:
We retrospectively analyzed patients presenting with knee pathology to an urban safety-net hospital between 2017 and 2020. Each had MRI or X-ray imaging and documented one-year surgical follow-up. Patients with missing imaging or outcome data were excluded. Text from radiology “Findings” and “Impression” sections was extracted from each initial report. OrthoScreen was developed by fine-tuning Bio_ClinicalBERT to predict whether surgery was considered within one year and evaluated using stratified 10-fold cross-validation. We compared performance to a Bag-of-Embeddings MLP model (BGE M3) using identical inputs. Metrics included AUC, sensitivity, specificity, PPV, NPV, and F1 score. No formal statistical comparisons were performed.

Results:
A total of 453 patients were included, 60% of whom underwent or were recommended surgery within one year. OrthoScreen achieved AUC 0.745, sensitivity 0.933, specificity 0.352, PPV 0.689, NPV 0.837, and F1 score 0.787. BGE M3 yielded AUC 0.713, sensitivity 0.922, specificity 0.375, PPV 0.699, NPV 0.754, and F1 score 0.787.

Conclusion:
OrthoScreen demonstrated strong performance using only imaging text and may help streamline surgical triage. Broader validation could support scalable integration into orthopaedic surgical referral workflows.		2025 AMA Research Challenge – Member Premier AccessShragvi Balajifree
39
129671Drosophila as a Model for Sex Differences in Repeated Traumatic Brain InjuryBackground
Repetitive traumatic brain injury (rTBI) is a leading risk factor for chronic traumatic encephalopathy (CTE) and other neurodegenerative disorders. While CTE has traditionally been associated with repeated head impacts, evidence suggests that a single TBI may drive tau pathology and progressive neurodegeneration. Biological sex is also known to influence TBI outcomes, yet mechanisms underlying sex differences remain poorly defined. Existing animal models often lack reproducibility and fail to include both sexes. This study aimed to develop and validate a rigorously calibrated Drosophila melanogaster model of rTBI to investigate sex-specific behavioral, molecular, and neuropathological responses to repetitive head trauma.

Methods
An automated, CO₂-driven ballistic impactor was engineered to deliver reproducible closed-head injuries to male and female flies. Flies received five impacts per day 1 psi to model mild rTBI. Locomotor activity and cognitive function were assessed using automated tracking and value-based feeding decision paradigms, respectively. Neurodegeneration was evaluated histologically, and whole-head proteomic profiling was performed at multiple time points post-injury to identify rTBI and sexually dimorphic molecular pathways.

Results
rTBI produced dose-dependent locomotor deficits and impaired decision-making in both sexes. Females were initially more resilient to multiple injuries, exhibiting fewer locomotor deficits early post-injury, but demonstrated significantly longer recovery periods compared to males. Histological analyses revealed increased neurodegeneration present despite tau knockout (dTauKO) flies, suggesting tau-independent mechanisms. Proteomic analysis revealed sex-specific alterations in mitochondrial function, oxidative stress responses, and pathways regulating locomotion and energy metabolism, which correlated with observed behavioral phenotypes.

Conclusion
This enhanced Drosophila rTBI model provides a robust model for studying sex-specific mechanisms of neurodegeneration. Our findings suggest that both the number of injuries and cumulative force contribute to neuropathology and highlight critical differences in male and female responses to repetitive head trauma. These insights emphasize the importance of incorporating sex as a biological variable in TBI research and identify potential therapeutic targets for mitigating long-term neurological sequelae.		2025 AMA Research Challenge – Member Premier AccessNicole Katchurfree
40
129667Listeria Meningitis in a Healthy Adult FollowingWild Honey Ingestion: A Rare Case Report from NepalBACKGROUND

• Listeria monocytogenes is a Gram-positive, facultative
intracellular bacterium found in soil, water, and
contaminated foods (e.g., unpasteurized dairy,
processed meats, raw vegetables, and honey).
• It causes life-threatening infections such as meningitis,
encephalitis, and sepsis.
• Primarily affects neonates, pregnant women, the
elderly, and immunocompromised individuals.
• In immunocompromised adults, infection typically
manifests as a mild, self-limiting gastroenteritis.
• Invasive CNS infections in healthy individuals are
extremely uncommon and can delay diagnosis.
• Transmission occurs mainly through ingestion of
contaminated food products.
• Early recognition and prompt antimicrobial therapy
(usually ampicillin with or without gentamicin) are
crucial for favorable outcomes.
• This case is notable due to its occurrence in a healthy
adult following wild honey ingestion, an unusual
source.

CASE PRESENTATION

• His symptoms began two days after consuming wild
Nepali honey mixed with rum, initially resembling a
mild flu-like illness without fever.
• On hospital day two, the headache intensified, and the
patient developed agitation and delirium requiring
haloperidol and promethazine.
• Non-contrast CT of the head was unremarkable.
• By day three, he developed a high-grade fever (103o F)
and was empirically started on piperacillin-tazobactam.
There were no meningeal signs.
• Lumbar puncture was performed.
• CSF sent for BioFire FilmArray Meningitis/Encephalitis
panel identified Listeria monocytogenes, prompting a
switch to IV ampicillin (2g every 4 hours) and
gentamicin (190mg every 12 hours).
• On day four, meningeal signs developed, and
fundoscopy showed early disc edema, managed with
mannitol.
• Although CSF culture remained negative, the patient
developed a transient pleural effusion by day six, which
resolved spontaneously.
• His Consciousness improved by the third day of
targeted therapy.
• Gentamicin was stopped on day 14; ampicillin was
continued for 28 days.
• The patient was discharged on day 30 and made a full
recovery without neurological complications.

DISCUSSION

• This case is significant for its atypical host and
presentation, resulting in delayed diagnosis and
treatment.
• It underscores the need to consider Listeria even in
immunocompetent adults presenting with
unexplained neurological symptoms.
• The unique association with wild honey ingestion
suggests a possible environment or foodborne
source.
• It emphasizes the importance of considering Listeria
in differential diagnoses, especially when standard
risk factors are absent.
• This case illustrates that Listeria monocytogenes can
cause invasive CNS infections even in
immunocompetent adults, particularly following
exposure to potential foodborne sources such as wild
honey.
• Early recognition and appropriate antibiotic therapy
are critical in preventing complications.
• Based on available literature, this appears to be the
first reported case of Listeria meningitis in an
immunocompetent adult linked to wild honey
consumption in Nepal.		2025 AMA Research Challenge – Member Premier AccessSomi Ahmedfree
41
129663Migratory Plaque-Like Lymphomatoid Papulosis: A Rare Clinical Presentation Mimicking MycosisBackground: Lymphomatoid papulosis (LyP) is a chronic, recurrent, self-healing cutaneous T-cell lymphoproliferative disorder characterized by CD30+ cells. It typically presents with papulonodular eruptions rather than plaques. Mycosis fungoides (MF), a primary cutaneous T-cell lymphoma, is generally persistent and non-migratory. This case highlights an unusual presentation of migratory plaque-like LyP, raising diagnostic challenges with overlap to more serious lymphoproliferations, such as MF.

Case Presentation: A 63-year-old female presented with pruritic plaques and nodules that intermittently appeared and resolved across various body sites. Past episodes were treated as fungal infections, though KOH preps were negative. She had a remote history of shingles but no other relevant systemic symptoms. Two punch biopsies were obtained a week apart, one from the right thigh and the other from the right groin.

Histopathology from both biopsies revealed an atypical intraepidermal T-cell lymphoproliferative disorder with CD30 positivity and a double-negative CD4-/CD8- phenotype by immunohistochemistry. Loss of several T-cell markers (CD2, CD5, CD7) was noted. The dermal component demonstrated a band-like infiltrate of mixed immunophenotype, including CD30 positive lymphoid cells. PAS was negative for fungal elements. T-cell receptor studies were positive for TCR beta clones (261 and 260bp; 294 and 294bp) as well as TCR gamma clones (240 and 239bp). Ancillary FISH for the DUSP22-IRF4 rearrangement is pending. Given the migratory nature of the rash, self-resolution of lesions, and histopathologic work-up with ancillary studies, a diagnosis of an unusual plaque-like variant of lymphomatoid papulosis was made.

Discussion: This case underscores a diagnostic gray zone between LyP and MF. Plaque-like LyP is rare and may be underrecognized due to its clinical resemblance to early MF. The migratory nature of the lesions and spontaneous resolution pointed toward LyP, yet the presence of plaques and immunophenotypic abnormalities added complexity. CD30+ double-negative T-cell populations are uncommon but can occur in LyP, occasionally associated with DUSP22-IRF4 rearrangements. Accurate diagnosis is crucial given the differing prognoses and management strategies between LyP and MF. This case also illustrates the value of serial biopsies and multidisciplinary pathology review in evaluating atypical cutaneous lymphoproliferative disorders.		2025 AMA Research Challenge – Member Premier AccessPaul Varghesefree
42
129660Textbook Outcomes in Medial Femoral Condyle/Trochlea Free Flap Reconstruction for Scaphoid NonunionINTRODUCTION: In surgery, textbook outcomes represent a composite metric of multiple variables that indicate baseline success in a specific procedure. These outcomes can be compared across surgeons and institutions. This study aimed to define a textbook outcome for the operative repair of scaphoid nonunion using medial femoral condyle (MFC) or medial femoral trochlea (MFT) vascularized bone flaps and identify the rate of textbook outcomes at our center.

METHODS: Three hand surgeons with expertise in scaphoid nonunion reconstruction using MFC/MFT grafts defined the criteria for a textbook outcome. After obtaining IRB approval, a retrospective review of cases from 2017 to 2024 at a single institution was performed. A textbook outcome was defined as meeting the following criteria: no flap loss, no intraoperative complications, a length of stay of less than two nights, radiographic union at three months postoperative on CT scan, return to baseline hand activities and pain-free hand activities at final follow-up, return to baseline and pain-free lower extremity activities by postoperative day 90, no recipient or donor site infections requiring intravenous antibiotics, no systemic complications, no emergency room admissions or hospital readmissions within 90 days, no mortality, and no surgical complications requiring reoperation.

RESULTS: A total of 38 patients met the inclusion criteria, with 27 (71.1%) achieving a textbook outcome. The most common criteria not met were surgical complications requiring reoperation (10.5%), return to baseline hand activities at final follow-up (7.9%), pain-free hand activities at final follow-up (7.9%), and return to baseline lower extremity activities by postoperative day 90 (7.9%). Of the 11 cases that did not achieve a textbook outcome, 5 violated one criterion (45.4%), 3 violated two criteria (27.3%), and 3 violated more than two criteria (27.3%).

CONCLUSION: Textbook outcomes are useful quality metrics for evaluating surgical success. By utilizing 13 criteria for scaphoid nonunion reconstruction with vascularized MFC or MFT grafts, cases can be assessed and compared using a single metric. Statistical analysis can help identify predictors of textbook outcomes, offering future opportunities to improve patient outcomes.		2025 AMA Research Challenge – Member Premier AccessJohn Dennenfree
43
129657Body Mass Index and Hypoglossal Nerve Stimulation Outcomes: A Systematic Review and Meta-AnalysisBackground: Hypoglossal nerve stimulation (HGNS) is an effective therapy for Obstructive Sleep Apnea (OSA) in patients intolerant to CPAP. However, concerns about its efficacy in patients with elevated body mass index (BMI) have led to restrictive insurance coverage, often using a threshold of 32 kg/m². This creates a critical treatment gap for a significant patient population. This systematic review and meta-analysis aimed to evaluate how a BMI ≥32 kg/m² influences HGNS surgical outcomes, challenging the evidence basis for current BMI-based restrictions.

Methods: Following PRISMA guidelines, a systematic literature search was conducted across PubMed, Embase, and Scopus for studies published between January 2014 and April 2025. Eligible studies included prospective/retrospective cohorts or case-control studies that reported HGNS outcomes in adult OSA patients, stratified by a BMI of 32 kg/m². Data on study design, patient demographics, and postoperative outcomes were extracted. A random-effects meta-analysis was performed on qualifying studies to estimate the pooled odds ratio (OR) for treatment success, primarily defined by the Sher criteria, comparing patients with BMI ≥32 kg/m² to those below this threshold.

Results: Seven studies comprising 1,572 patients were included in the qualitative synthesis, with four studies (n=1,507) qualifying for the meta-analysis. While individual study findings were mixed, the meta-analysis revealed that patients with a BMI ≥32 kg/m² had a trend toward reduced odds of achieving surgical success (pooled OR = 0.87, 95% CI: 0.68–1.11), though this result was not statistically significant. Notably, many patients in the higher BMI groups who did not meet formal success criteria still experienced clinically meaningful reductions in Apnea-Hypopnea Index (AHI) and reported high rates of satisfaction and subjective improvement on validated scales like the Epworth Sleepiness Scale and the Clinical Global Impression Index scale.

Conclusion: The assertion that a BMI ≥32 kg/m² uniformly predicts HGNS failure is not supported by this meta-analysis. The lack of statistical significance, coupled with consistent reports of objective and subjective benefits, suggests that many patients with elevated BMI derive substantial value from HGNS. These findings challenge the rigid application of BMI-based cutoffs in payer policies and advocate for a more individualized, patient-centered approach to determining HGNS candidacy, which could expand access for a population with limited treatment alternatives.		2025 AMA Research Challenge – Member Premier AccessColeton Evansfree
44
129654Methanol Poisoning Presenting with Bilateral Basal Ganglia Hemorrhage: A Case ReportBackground
Methanol is a highly toxic, colorless, volatile alcohol that is present in homemade beverages, paint removers, windshield liquid, anti-freeze, shellac, and perfumes. Although uncommon in the United States, methanol poisoning can occur with improper distillation of spirits. Over 50% of morbidity and mortality associated with methanol poisoning is accidental in nature and therefore preventable. Methanol poisoning, even in small quantities of 50 to 100 ml, is associated with severe complications such as permanent blindness.
Aim
The aim of this study was to elaborate on a case of methanol poisoning and provide a deeper understanding of bilateral basal ganglia hemorrhage, which is a relatively rare complication of methanol ingestion.
Case report
A 28-year-old Caucasian female with a history of seizures was brought to the emergency department after being found unresponsive due to ingestion of hand sanitizer and mouthwash. Due to severe hypoventilation and concern that the patient would not be able to protect her own airway, she was intubated. Physical examination revealed that the patient was cyanotic with dilated and sluggish pupils. Blood work was remarkable for a low bicarbonate, transaminitis with an aspartate transaminase to alanine transferase ratio of 2:1, and an elevated anion gap. Arterial blood gas revealed lactic acidosis and metabolic acidosis. The osmolar gap was elevated. Serum ethanol and acetaminophen levels were negative. A head CT scan showed bilateral basal ganglia hemorrhage with surrounding edema. The volatile alcohol screen was positive for methanol. The patient was started on fomepizole. Unfortunately, she passed away the following day.
Discussion
Methanol is lethal when ingested above 1 g/kg of body weight. Clinical manifestations of methanol intoxication typically present 12 to 24 hours following the ingestion of methanol. Methanol is converted to formic acid, a toxic metabolite, by alcohol dehydrogenase in the liver. The early clinical features of methanol poisoning include headache, vomiting, weakness, and abdominal pain. After 24 hours of ingestion, formic acid causes end-organ damage to the retina, optic nerve, and basal ganglia. Visual manifestations include optic disc hyperemia, edema, central scotoma, and permanent blindness. This is a result of optic nerve atrophy and demyelination, which occur due to the inhibitory effect of formic acid on mitochondrial ATP production and anoxia. Neurological manifestations include seizures, coma, and death secondary to ischemic necrosis or hemorrhagic injury to the basal ganglia, especially the putamen and subcortical region. The exact etiology of necrosis and hemorrhage of the basal ganglia secondary to methanol poisoning remains unknown. However, it is thought to be secondary to the direct toxic effect of methanol and its metabolite formic acid. The management of methanol toxicity includes preventing further formic acid production by inhibiting alcohol dehydrogenase using either fomepizole or ethanol. It also includes the removal of previously accumulated methanol and formic acid from the blood with the correction of metabolic acidosis by hemodialysis and sodium bicarbonate infusion. This case report emphasizes the importance of early detection of methanol poisoning to prevent overt fatal outcomes.		2025 AMA Research Challenge – Member Premier AccessAhar Bhattfree
45
129651Successful left atrial appendage closure using Watchman in a patient with an atrial septal defectAtrial fibrillation is a common arrhythmia that raises stroke risk, often requiring
anticoagulation. This case report discusses a 79-year-old female with Atrial fibrillation and prior stroke, who was intolerant to anticoagulation due to recurrent bleeding. Percutaneous left atrial appendage occlusion is the next step in management to reduce stroke. She had a patent foramen ovale closure with an Amplatzer septal occluder, posing challenges for the left atrial appendage occlusion procedure. In this case, the watchman device was successfully placed under transesophageal echocardiography and fluoroscopy, avoiding interference with the patent foramen ovale device.		2025 AMA Research Challenge – Member Premier AccessRohit Kumarfree
46
129647Community‐Led Control of Schistosomiasis in Rural Nigeria with a $1 microscope and Drone MappingAMA abstract category: Choose one
Public Health and Health Policy Submissions in Public Health and Health Policy must relate to the science of protecting and improving the health of people and their communities. This work is achieved by promoting healthy lifestyles, researching disease and injury prevention, and detecting, preventing, and responding to infectious diseases. This topic also includes studies on COVID-19, vaccinations, and health equity.
Clinical and Translational Research Submissions in Clinical and Translational Research must center around human subject research aimed towards prevention, diagnosis, or treatment of disease processes either fully in the clinical setting, or as a bridge between the lab and clinical practice.
Note: All abstracts must be submitted using this approved template and must not include pictures, graphs, or tables. The abstract must not exceed 400 words, excluding the abstract title, page numbering, or any ancillary entrant information. Citations or references are not a necessity with the abstract. If your abstract is accepted for consideration, you may include citations or references with your poster.

Abstract Title
Community‐Led Control of Schistosomiasis in Rural Nigeria Using the Foldscope, Environmental Mapping, and Educational Initiatives
Draft 2 (word count 400)
Background
Schistosomiasis is a neglected tropical disease transmitted through freshwater contact, disproportionately affecting children in low-resource settings. In rural Nigeria communities like those in Ogun State, dependence on natural freshwater for daily occupation increased vulnerability to Schistosoma haematobium infection, which can cause hematuria, anemia, and long-term risks including bladder and cervical cancer. Limited clinic access, stigma, and low diagnostic capacity–due to the need for electricity, skilled personnel, and costly equipment of traditional microscopy–delay timely care in these regions.
Methods
This mixed-methods study (July 22 to August 1, 2025) with three components. First, it will evaluate the feasibility and diagnostic potential of a community-led screening strategy using the Foldscope, a $1 paper microscope, paired with a reusable filtration slide. The study will be conducted in Imala and Imala Odo, high-prevalence communities near the Oyan River Dam. Community health extension workers (CHEWs) will be trained to identify Schistosoma ova in urine samples from children and adults. Diagnostic results will be compared to gold-standard microscopy, and participants will receive praziquantel treatment as needed from a local clinician. Interviews with CHEWs and NTD coordinators will assess feasibility for broader implementation. Second, environmental mapping (July 17–22) will use drones and satellite imagery to locate Bulinus snail habitats and model potential transmission zones in arcGIS. Third, to promote local engagement and youth education, we are deploying a large-format, traditionally-woven mat game based on Schisto & Ladders. This locally adapted board game teaches children about symptoms, transmission, and prevention, and will be reimagined into a group game to be played in an educational setting, embodying the principles of disease transmission for spatial understanding.
Results
Preliminary findings from Houston demonstrate that the Foldscope has 79% sensitivity and 100% specificity among untrained users compared to gold-standard microscopy. In Nigeria, we anticipate enrolling 406 participants across Imala and Imala Odo during the study period. Quantitative outcome variables for evaluating Foldscope effectiveness will include sensitivity, specificity, positive and negative predictive values, and overall diagnostic accuracy. Qualitative data will be through focus groups and interviews with CHEWs and local clinicians to assess the acceptability and feasibility of community-led screening.
For the environmental mapping component, outcome variables will include vegetation density around water access points across 4 transmission hotspot sites, additionally ground-truthing verification of habitat images using vegetation and snail samples.
Conclusion
This project integrates ultra-low-cost diagnostics, geospatial surveillance, and culturally tailored education to support scalable, community-led schistosomiasis control aligned with SDGs 3, 4, 6, and 9.


DA6YJ6G7ZQ
Thank you! You have successfully completed your abstract submission.
Confirmation number:
DA6YJ6G7ZQ
You must be an AMA member to participate. Confirm your member status here. Join or renew by July 16 for your abstract to be considered.
Participants will be notified of abstract acceptance decisions in early September 2025.
You may modify your abstract submission up through the submission deadline.
Please do not resubmit as a new submission.
Note: Only one abstract submission per entrant can be considered. Multiple abstracts from the same submitter will not be considered.
You will receive an email confirmation shortly.
Questions? Please contact researchsymposium@ama-assn.org.

2025 AMA Research Challenge – Member Premier AccessWilliam McCarthyfree
47
129643JP4-039 as a Mitochondria-Directed Strategy to Address Cardiovascular
Dysfunction in Preeclampsia		Preeclampsia (PE) is a pregnancy-specific hypertensive disorder arising after 20 weeks of gestation and remains a major cause of maternal and neonatal morbidity. It originates from abnormal trophoblast invasion and inadequate spiral artery remodeling, leading to placental ischemia, endothelial dysfunction, and excessive mitochondrial reactive oxygen species (mitoROS). These events drive trophoblast injury, impaired angiogenesis, and systemic oxidative stress that not only compromise fetal development but also confer a two to four higher lifetime risk of cardiovascular disease and heart failure in
affected women.

Conventional antioxidants have shown limited success, underscoring the need for targeted interventions. Mitochondria-directed therapeutics such as JP4-039, a nitroxide that scavenges mitoROS and stabilizes mitochondrial membrane potential, offer promise in restoring endothelial and placental health. We propose a placenta-targeted delivery systemusing CGKRK-engineered extracellular vesicles (EVs) loaded with JP4-039 to reduce oxidative injury, enhance vascular remodeling, and mitigate both acute placental dysfunction and long-term maternal cardiovascular risk in Preeclampsia.		2025 AMA Research Challenge – Member Premier AccessFeyza Achilovafree
48
129640Hand-Sewn vs. Stapled Colorectal Anastomosis: A Meta-Analysis of Postoperative Outcomes and Anastomotic IntegrityTitle:
Hand-Sewn vs. Stapled Colorectal Anastomosis: A Meta-Analysis of Postoperative Outcomes and Anastomotic Integrity
Edwin Medina MS2, Ian Olmeda MS3
University of Puerto Rico School of Medicine, Universidad Central del Caribe School of Medicine
Introduction
The integrity of colorectal anastomosis is critical in minimizing postoperative morbidity, particularly anastomotic leak a leading cause of mortality in colorectal surgery. While both hand-sewn and stapled techniques are widely practiced, evidence remains inconclusive regarding superiority in clinical outcomes. This systematic review and meta-analysis aimed to compare both techniques with respect to leak rate, stricture formation, operative time, and overall postoperative complications.
Methods
Following PRISMA guidelines, we searched PubMed, Embase, and Cochrane Library for studies published up to April 2025. Eligible studies compared hand-sewn and stapled anastomoses in adult colorectal resections. Primary outcome was anastomotic leak rate; secondary outcomes included stricture formation, reoperation, operative duration, and length of hospital stay. Pooled odds ratios (OR) and weighted mean differences (WMD) were calculated using a random-effects model.
Results
A total of 42 studies (9 RCTs, 33 cohort studies) were included. Stapled anastomoses demonstrated a modestly lower leak rate (OR 0.82; 95% CI 0.71–0.94; p = 0.004), particularly in low anterior resections. Stricture rates were slightly higher in stapled groups, although not statistically significant (OR 1.14; p = 0.08). Operative time was shorter by an average of 22 minutes with stapling (WMD -21.6 min; p < 0.001). No significant difference was found in overall morbidity or reoperation rates. Subgroup analysis showed no clear advantage for either method in right-sided resections.
Conclusion
Stapled anastomosis is associated with lower anastomotic leak rates and shorter operative times, particularly in low pelvic procedures. However, the increased stricture tendency warrants caution in specific anatomic settings. These findings may guide surgical decision-making based on location, patient comorbidities, and institutional expertise. Further large-scale randomized trials are needed to establish evidence-based guidelines for technique selection in colorectal surgery.
2025 AMA Research Challenge – Member Premier AccessIan Olmedafree
49
129636Price Transparency Impacts Medication Ordering Behavior Among Therapeutic AlternativesBackground
Price transparency initiatives have gained traction among health reform efforts over the past decade. Almost exclusively targeting patients, existing efforts have produced underwhelming results, mostly failing to encourage meaningful price shopping. What remains underexplored, however, is whether making prices more transparent to physicians can change prescription behavior and ultimately lower health care spending.

Methods
Leveraging detailed electronic health record (HER) order entry data from 5 hospitals associated with a large health system in 2018–19, we conducted an interrupted time series analysis to estimate the extent to which the introduction of a novel EHR-based price transparency intervention was associated with a decrease in the proportion of high-price medications ordered by physicians when therapeutically similar lower-price medications existed. Our study spanned the 12 months immediately pre- and post-intervention, and we estimated effects separately for patients with pneumonia and sepsis, 2 common conditions that range in severity and have multiple treatment options. We grouped medications into high- and low-price categories based on acquisition data, and adjusted analyses for hospital, seasonality, and patient characteristics (age, sex, race, education, language, comorbidities). We performed falsification analyses using placebo intervention months.

Results
Across 12,863 patients with pneumonia and sepsis in 2018–19, patient characteristics were balanced across pre- and post-intervention periods (SMD <0.07 for all). Physicians prescribed significantly fewer high-price medications post- vs pre-intervention, with an estimated 4.8
Ó 2025 American Medical Association. All rights reserved.
percentage point (pp) absolute reduction (32% relative reduction; P<0.01) in high-price medications following the intervention for patients with pneumonia and a 5.8 pp reduction (23%; P=0.029) for patients with sepsis. These effects were almost entirely driven by less severe cases (P<0.01) and were not replicated in placebo analyses (P>0.05). Reductions in length-of- stay and total admission cost were non-statistically significant (P>0.05).

Conclusion
Driven by less complicated cases with multiple close therapeutic substitutes, our results suggest that making medication prices transparent to physicians leads physicians to order medications of similar efficacy but lower cost to their patients when acuity allows. Price transparency efforts may be more productively targeted at providers than patients, with the potential to lower health care spending.		2025 AMA Research Challenge – Member Premier AccessMaximilian Panyfree
50
129632HMB Supplementation in Muscle Wasting: A Narrative ReviewBackground: Muscle wasting syndromes—including cancer cachexia, chronic obstructive pulmonary disease (COPD), cirrhosis, and aging-related sarcopenia—contribute significantly to functional decline, hospital readmissions, and mortality. β-Hydroxy β-methylbutyrate (HMB), a leucine-derived metabolite, has demonstrated anti-catabolic effects by inhibiting protein degradation and stimulating protein synthesis, yet its clinical integration remains limited despite promising evidence. This review aims to evaluate clinical evidence on HMB supplementation for muscle preservation in patients with cachexia and chronic illness and identify opportunities for integration into recovery and palliative care models.
Methods: A structured narrative review was conducted using existing medical literature databases like PubMed (2005–2025). Search terms included "HMB," "β-hydroxy β-methylbutyrate," "cachexia," "sarcopenia," "muscle wasting," and specific disease states (cancer, COPD, cirrhosis). Inclusion criteria comprised randomized controlled trials, meta-analyses, and systematic reviews assessing HMB supplementation (alone or in combination with other nutrients) in adults with or at risk for muscle wasting. Primary outcomes evaluated included changes in lean body mass, muscle strength, physical function, and quality of life. Secondary outcomes included inflammatory markers, nutritional status, and hospitalization rates.
Results: HMB supplementation (typically 1.5–3g/day) has demonstrated benefits across multiple conditions. In studies of older adults with sarcopenia, HMB combined with other nutrients has been associated with preservation or modest increases in lean body mass. Research on bed rest in older adults has shown that HMB supplementation may attenuate muscle loss compared to placebo. In cancer and other catabolic conditions, evidence suggests HMB helps preserve muscle mass through reduced proteolysis and enhanced protein synthesis. For patients with liver disease, combining nutritional supplementation with exercise training has shown promising results for muscle function. Meta-analyses indicate improvements in muscle strength across randomized trials in elderly populations. Across studies, HMB has been well-tolerated with minimal adverse effects reported.
Conclusion: HMB represents a promising, evidence-based adjunct to nutritional therapy in cachexia and chronic illness. Current evidence supports the potential role of HMB in clinical nutrition protocols for patients at risk of muscle wasting, particularly during periods of increased catabolism. Implementation barriers include limited provider awareness and inconsistent insurance coverage. Future research should prioritize disease-specific dosing protocols and integration with exercise interventions to optimize functional outcomes and quality of life.
2025 AMA Research Challenge – Member Premier AccessBhupinder Grewalfree
51
129629Natural History of Troponin I Elevations in Pediatric Emergency Room PatientsBackground: Since emergence of COVID-19, and its association with myocarditis, the use of troponin I (TnI) tests has drastically increased at pediatric ERs. We hypothesize that in pediatric patients presenting to the ER, the natural history of trivially elevated troponin is normal without cardiac disease. Additionally, we hypothesized there exists a cut-off troponin value with good sensitivity and specificity.

Methods: We conducted a single-center retrospective study of patients who were presented to our ER 2022-2023 found to have elevated troponin levels (hospital assay defines abnormal > 0.021 ng/mL) and the following chief complaints: chest pain, dizziness, syncope/near-syncope, fatigue, shortness of breath, and palpitations. Patients without elevated troponin or follow-up cardiac testing (echocardiogram, ECG) were excluded. Patients with a prior history of cardiac disease were excluded. We compared patients found to have no cardiac disease with those who did. Cardiac disease was defined as an abnormality on ECG, echo, or MRI (if available). Comparisons were made using chi-square and independent t-tests. Receiver operating curve (ROC) analysis to determine discriminatory ability of troponin.

Results: In total, 275 patients met inclusion criteria for chief complaint and elevated troponin, though 103 had incomplete cardiac testing, leaving 172 patients for analysis. Of these, 29 (17%) were diagnosed with cardiac disease and 143 (83%) without cardiac disease (Table). There was a lesser incidence of female sex in the cardiac disease group. Those with cardiac disease had a significantly greater troponin value at presentation (10.0 +/- 12.1 vs. 0.35 +/- 1.5, p < 0.001). MRI was only available for 17 patients total (10%). ROC analysis demonstrated excellent discriminatory ability for troponin at ER presentation for detection of cardiac disease (Figure, area under curve 0.95, p < 0.001). Youden’s J determined a cut-off troponin value of 0.3 for sensitivity 90% and specificity 86%. There were 47 patients with a troponin > 0.3, and 125 < 0.3. A troponin > 0.3 yields a positive predictive value of 55%, while < 0.3 yields a negative predictive value of 98%. Those with troponin <0.3 were admitted 70% of the time, and >0.3 admitted 83%.

Conclusions: Troponin level at ER presentation demonstrates good discriminatory ability for subsequent cardiac disease diagnosis. Low troponin levels (< 0.3) have a high negative predictive value, however a significant number of patients were admitted to the hospital. This initial data may guide appropriate triaging and resource utilization.

2025 AMA Research Challenge – Member Premier AccessJenil Ranafree
52
129626Stepwise Differentiation of Engineered Human Meniscus Tissues in 3D-Printed Flexible ScaffoldsBackground
Meniscus injuries are common with approximately 850,000 patients undergoing surgical treatment in the United States yearly. To investigate regenerative strategies for avascular meniscus tears, our lab has been utilizing bovine meniscus explants for investigating regenerative strategies for tears and crosstalk between joint tissues. However, inter-species crosstalk between bovine meniscus and human derived joint tissues remains a concern.

This study was designed to engineer and validate a model of human meniscus for engineered healing of avascular meniscus tears under the crosstalk between meniscus, synovial membrane, and IFP.

Methods
A wedge-shaped scaffold was 3D-printed with Ecoflex F Blend C1200 polymer. Synovial mesenchymal stem/progenitor cells (syMSCs) were isolated from surgically removed synovium of anonymous adult donors as per our previous methods. The scaffolds were then loaded with collagen bioink and 2,000,000 cell /mL syMSCs and then placed into the PDMS molds. Experiment 1: For the first 2 weeks, the tissue constructs were cultured in Fibrogenic induction supplemented (FIS) media with connective tissue growth factor (CTGF). For weeks 3-4, the tissue constructs were cultured in chondrogenic induction supplemented media (CIS) and transforming growth factor beta-3 (TGFβ-3). Experiment 2: For 4 weeks tissue constructs were cultured in FIS with CTGF and CIS with TGFβ-3. No loading was applied. Experiment 3: Tissue constructs were cultured in combined FIS (w/ CTGF) and CIS (w/ TGFβ-3) with no FBS and no loading.

Results
At 4 weeks, histology showed new tissue integration into the pores throughout the scaffolds. Picrosirius Red (PR) and Alcian Blue (AB) staining suggest the de novo tissues in the no loading group are dense fibrocartilaginous tissues. Polarized images of PR-stained section showed densely aligned collagen fibers, reminiscent of native meniscus, comprising circumferential fibers and radial tie fibers. Mechanical loading resulted in less mature fibrous tissue matrix with limited cartilaginous matrix. Combined culture with FIS (w/ CTGF) and CIS (w/ TGFβ-3) showed the most mature and dense fibrocartilage tissue development after 4 weeks. RNA sequencing data show similar expression levels of fibrochondrocyte-specific genes, including COL1A2, ACAN, COMP, and COL1A1, between engineered and native human meniscus tissue.



Conclusion
Our study suggests a reliable and effective approach to engineer human meniscus-like tissues that can serve as an experimental explant model for various types of meniscus explant research. Combined culture with FIS and CIS creates more mature fibrocartilage structure than step-wise differentiation under no physiological loading. These findings may have implications for engineering human meniscus-like tissue for various types of meniscus research.
2025 AMA Research Challenge – Member Premier AccessLauren Burgettfree
53
129620Enhancing Medical Training on Intimate Partner Violence Through an Interactive CurriculumIntimate Partner Violence (IPV) is a significant public health concern that remains insufficiently addressed in medical education. A national survey indicates that nearly one in two women (47.3%) and more than 40% of men (44.2%) in the United States have experienced IPV in their lifetime. Despite the high prevalence and serious health implications of IPV, comprehensive training on this topic has not been widely incorporated into medical school curricula. Current research shows integrating interactive simulations into medical school curriculum is the best practice for providing lasting knowledge on the training topic at hand. This study aims to bridge the educational gap of IPV in medical students by implementing an interactive, simulation-based curriculum. This will not only enhance future physicians' competency in recognizing IPV but also equip them with the skills to provide appropriate resources to victims. Participants will engage in a structured training session, featuring a role-playing exercise and guided case discussions. A pre- and post-survey model will be used to assess changes in knowledge, confidence, and attitudes toward IPV management. Preliminary data has demonstrated statistically significant improvements in both preparedness and knowledge among participants following the intervention. These preliminary findings suggest the intervention was effective in enhancing medical students’ confidence and understanding related to identifying and managing intimate partner violence in clinical settings. Additionally, initial data suggests students feel more adequately equipped with practical tools for screening and intervention, alongside a heightened sense of empathy, compassion, and understanding of IPV victims - an essential step in providing trauma-informed, supportive care. Notable limitations of this study include the immediate timing of participant responses, which may have been influenced by the emotional impact of the intervention, as well as a small final sample size (n=20). Although 59 participants completed the pre-survey, only 24 completed the post-survey, and just 20 submitted fully completed responses. This low yield of responses raises concerns about feasibility and integration of such a curriculum within an already demanding medical school schedule. Alongside a larger sample size, longitudinal investigation of participants would be needed to assess efficacy of the curriculum in clinical practice. However, by implementing a curriculum that is interactive and attitudinally transformative, we aim to bridge the existing educational gap and, ultimately, provide more comprehensive patient care in clinical practice.2025 AMA Research Challenge – Member Premier AccessAshley Rousseaufree
54
129616Breath-Based Lung Cancer Detection Using an ML-Driven Low-Cost Sensor Array for VOC Analysis
Lung cancer is a silent killer, often going undetected until it’s too late. Early diagnosis greatly improves outcomes—stage I patients have a 97% two-year survival rate, compared to just 23% at stage IV. However, most diagnoses occur at later stages due to limitations in current screening methods. Current detection methods, namely Gas chromatography–mass spectrometry (GC-MS) and CT scans, are costly or invasive, leading to most diagnoses at advanced stages. These methods require in-lab operation, and are inaccessible to rural or resource-limited settings. While several studies have explored small-scale sensor systems as an alternative, they often suffer from reduced accuracy, high production costs, and long processing times—sometimes exceeding 40 minutes per breath sample. Therefore, there is a need for a fast, affordable, and accurate breath-based lung cancer detection system suitable for those with limited resources. To address this, we developed an affordable breath-based lung cancer screening device featuring a gas chamber and machine learning model. It analyzes VOC composition in human breath, influenced by lipid peroxidation in cancer cells, to compare profiles between healthy and lung cancer samples. The system was designed to detect more than 30 volatile organic compounds (VOCs) using twelve metal oxide semiconductor sensors and one chemi-resistive alkane sensor. In total, 46 clinical breath samples were collected (from 28 healthy controls and 18 lung cancer patients) in 1L Tedlar bags. Electrical resistance data from the sensor array was processed using a multilayer perceptron (MLP) neural network. The model architecture incorporated dropout and batch normalization layers to minimize overfitting. Sensor output patterns were statistically analyzed using the Mann-Whitney U test, which confirmed significant differences between cancer and control groups (p < 0.05). Model performance was evaluated using five-fold cross-validation to ensure robustness and generalizability. The final model achieved an accuracy of 96.57%, sensitivity of 92.88%, and specificity of 98.82%, with an area under the ROC curve (AUC) of 0.9913. Low standard deviation across folds indicated consistent classification performance. The complete system was able to produce diagnostic results in about 10 minutes per breath sample. Our system outperforms existing methods of lung cancer detection through MOS sensors by greater than 5% while costing only around $300. These findings suggest that the breath analyzer system proposed in this study has the potential for preliminary lung cancer screening.



2025 AMA Research Challenge – Member Premier AccessDhruv Iyerfree
55
129613Breaking the barriers: Immune checkpoint inhibitors in HIV-positive Non-Small Cell Lung CancerBackground:
Immune checkpoint inhibitors (ICIs) transformed the treatment for non-small cell lung cancer (NSCLC) with durable responses and improved outcomes. Patients with HIV (PWH) have increased risks of lung cancer in comparison to the general population. The ICI use in PWH has been limited due to exclusion from clinical trials with concerns for immunosuppression. This study aims to review the current evidence on ICI’s safety, efficacy, and adverse effects in PWH with NSCLC.
Methods:
A systematic literature search was conducted using PubMed and Medline, with keywords "Immune checkpoint inhibitors" or "ICIs", "Non-small cell lung cancer" or "NSCLC", and "HIV" or "HIV-positive" or "Human Immunodeficiency Virus". Inclusion criteria includes retrospective series, clinical trials, and case reports. Descriptive statistical analysis was performed for overall survival (OS), progression-free survival (PFS), response outcomes and adverse effects.
Results:
Ten studies were included with 135 PWH with NSCLC. The largest study (n = 102) demonstrated a median OS of 16 months and PFS of 6.3 months, with an ORR of 31%. No significant differences in OS or PFS were found between patients with CD4+ counts >200 vs <200 cells/uL. Across all studies, partial response rate ranged from 13-50%, and disease control rates reached 62.5% in one cohort. ICIs were generally well tolerated, with adverse effects similar to those observed in HIV-negative populations. Most toxicities were mild, although grade ≥3 events (e.g., pneumonitis, colitis, dermatotoxicity) were reported. Antiretroviral therapy (ART) was maintained throughout treatment in all patients.
Conclusion:
ICIs appear to be safe and effective in treating NSCLC in PWH who remain on ART, regardless of the CD4+ counts. Though our study is limited due to the small number of studies included, the current information supports ART-controlled PWH with NSCLC to undergo ICI’s as part of treatment, and to encourage access to ICI’s in this population.
2025 AMA Research Challenge – Member Premier AccessChanika Assavarittirongfree
56
129610Assessing the Perceptions of Osteopathic Medical Students’ beliefs since Transitions in the USMLE step 1 and COMLEX level 1 examinations to Pass/FailAssessing the Perceptions of Osteopathic Medical Students’ beliefs since Transitions in the USMLE step 1 and COMLEX level 1 examinations to Pass/Fail
Student Author: Milo Taylor
Mentor: Robert Goldsteen
Objective: since the transition of the USMLE step 1 and COMLEX level 1 examinations to pass/fail in 2022, there has been very little research into the impact of this transition on osteopathic medical students. The goal of this study is to investigate the impact of the transitions of these exams to pass/fail on osteopathic student well-being depending on their intended specialty.
Methods: A Qualtrics survey was administered to students at the Burrell College of Osteopathic Medicine. This survey included a free response question asking what specialty the student intended to pursue. Then the perceived stress scale 10 (PSS-10) was administered. The responses were divided into four groups based on how competitive the intended specialty was. The PSS-10 score averages were calculated for each group and compared.
Results: the most competitive group had a PSS-10 mean score of 20.29, the intermediate group had a mean PSS-10 score of 21.29, the least competitive group had a mean PSS-10 score of 21, and the undecided group had a mean PSS-10 score of 20.33.
Conclusion: There was very little difference in stress levels between the four groups, and therefore intended specialty seemed to not have any effect on stress levels. This possibly confirms our hypothesis that there would be no difference in stress levels between the groups because the pressure of obtaining a high numerical score on these exams has been eliminated. While this preliminary data suggests that the NBOME and NBME may have succeeded in improving student well-being with the decision to make these exams pass/fail, further research must be done to investigate the effects of this transition.


2025 AMA Research Challenge – Member Premier AccessMilo Taylorfree
57
129607Comparison of FFPE to Fresh Liver Biopsies Using Whole Transcriptome Analysis from Patients with HCV using an AI-based Bioinformatics PipelineBackground:
Formalin-Fixed Paraffin-Embedded (FFPE) tissue is the most commonly archived specimen in clinical practice due to its ease of storage, long-term preservation, and compatibility with histopathological evaluation. However, the use of FFPE samples in molecular studies has been limited by the degraded and fragmented nature of RNA, often compromising the quality and reliability of downstream transcriptomic analyses. Recent advances in RNA extraction protocols and quantification methods have begun to overcome the molecular limitations associated with FFPE tissue, enabling more accurate and reproducible gene expression profiling from archival samples. Our goal was to compare gene expression profiles between FFPE and fresh liver biopsy samples from patients with physiologically normal livers (controls) and hepatitis C virus (HCV) infected livers, pre and post Direct-Acting antivirals (DAA) treatment, using an AI-based bioinformatics pipeline.

Methods:
We compared gene expression profiles derived from RNA isolated from FFPE vs fresh liver biopsies obtained from controls, HCV Pre-DAA, and HCV Post-DAA. Whole-transcriptome sequencing was performed using Smart-3SEQ. We developed an AI-bioinformatics pipeline through prompting with Claude Sonnet 4 within QualifiedHealth. FastQ pseudo-alignment and count generation were performed using Salmon within Galaxy (web toolkit. Filtering was performed with traditional CPM thresholds, TMM within edgeR, and non-coding gene removal. Normalization and Differential Expression (DE) analysis was performed with Limma-Voom.

Results:
No significant gene expression differences were found between FFPE and fresh samples in the control or post-DAA groups. However, in the pre-DAA group, 40 genes were initially found as significantly different. Subdividing the Pre-DAA condition according to immune-gene expression (PreCold for down-regulation and PreHot for up-regulation) resulted in no significant DEGs within each group. Furthermore, investigation of 17 housekeeping genes across all samples revealed no significant differences in expression between FFPE and fresh samples.

Conclusion:
Gene expression profiles from FFPE samples showed high correlation with fresh tissue across three distinct liver conditions. Previous significant DEGs in the Pre-condition were attributed to immune gene expression rather than storage methodology. Commercially available LLMs enable rapid development of robust bioinformatics workflows for complex transcriptomic analyses, while the minimal differences between sample types validate FFPE tissue as a reliable source for retrospective molecular studies.
2025 AMA Research Challenge – Member Premier AccessDavid Girata Narinofree
58
129604Improving CPR Confidence and Access Through Community-Based, Peer-Led Training: A Pilot EvaluationOut-of-hospital cardiac arrest (OHCA) is a leading cause of preventable death, yet fewer than 40% of victims receive bystander CPR, with underserved communities facing the greatest barriers to access. Beat to Breathe CPR, a student-led nonprofit, implemented a mobile, outreach-based CPR program across New Jersey to improve community preparedness through free, one-hour training sessions tailored to specific audiences. In a pilot study of 15 youth camp counselors, average CPR confidence increased by 124% (p<0.001) and knowledge by 82% (p<0.001), with all participants reporting greater readiness to act during emergencies. These findings suggest that community-delivered, low-barrier CPR training can significantly improve layperson competence and confidence while reducing structural barriers such as cost and transportation. This model demonstrates a scalable approach to increasing CPR accessibility and preparedness among populations historically underrepresented in traditional CPR education.2025 AMA Research Challenge – Member Premier AccessJay Patelfree
59
129601Sex Disparities in Medicare Reimbursement, Case Volume and Practice Characteristics Among Cervical Spine Surgeons in the United StatesBackground: Medical schools have equal representation of men and women, yet surgical specialties, like orthopaedics and neurosurgery, have far less women. Addressing sex disparities among spine surgeons may promote more effective spine care in the United States (US).
Methods: The Medicare Provider Utilization and Payment Dataset was used to identify neurosurgical and orthopaedic surgeons performing ACDF surgery from 2013-2021. The dataset was linked to the Medicare National Downloadable File to determine hospital affiliations and medical training. All hospitals included were queried to determine association with residency or fellowship training programs. This retrospective cross-sectional study measured the total representation of female spine surgeons performing ACDF, their procedural volumes, and reimbursements compared to male cervical spine surgeons.
Results: There were 2,492 surgeons who performed at least eleven ACDFs over the study period accounting for 139,456 total cases. Of this, there were 58 female spine surgeons (2.3%) who performed 2,733 cases (2.0%). The percentage of female surgeons increased from 0% (0/1,124) in 2013 to 1.8% (12/651) in 2021 (p<0.001). The average volume of ACDFs performed by females (16.83.6) was similar compared to males (17.27.6; p=0.712). Mean standardized reimbursement by Medicare for ACDFs was $1,304, with no difference between female and male surgeons during the study period (p=0.145). There were no significant differences in the beneficiary demographics between male and female surgeons, except for in 2021 where female surgeons saw a higher proportion of female patients (56.8 vs 55.0%, p=0.043).
Conclusions: Despite some progress observed over the study period, profound sex disparities persist among spine surgeons performing ACDF. Strategies are needed to increase the number of female medical students and residents interested in spine surgery.		2025 AMA Research Challenge – Member Premier AccessJuliette Gammelfree
60
129598Characterizing Emergency Department Provider Decision to Diagnose Traumatic Brain Injury Versus Head InjuryBackground

Misdiagnosis of mild traumatic brain injury (mTBI) (concussion) in emergency departments (EDs) is underscored by historical controversy around the appropriate usage of traumatic intracranial injury diagnoses (S06 ICD-10 codes) intended for mTBI and head injury, unspecified codes (S09.9). The lack of consensus diagnostic criteria for mTBI has made the identification of this patient population challenging. Reducing variability in diagnosis is essential to advance research initiatives and align clinical recommendations across patient settings. We found that ED providers were divided in their preferences for selecting between these diagnoses, potentially impacting patient education about brain injury symptoms and clinical management. In this study, we sought to characterize patient and clinical factors that predict provider diagnosis. Additionally, we explored the effect of diagnosis on day 1 symptoms, patient education, and return to ED/hospital within 30 days.



Methods

We analyzed 1348 prospectively identified adults diagnosed with traumatic intracranial injury head injury, unspecified across 7 EDs from September 24, 2024, to May 25, 2025. We used logistic regression to identify demographic, injury, and clinical predictors of provider diagnosis and associations between diagnosis and day 1 symptom burden, provision of educational handouts, and 30-day readmission.



Results

Diagnosis of intracranial injury was independently predicted by younger age (p<0.001), higher level of care (p = 0.017), presentation >24 hours post-injury (p<0.001), and patient trauma activation on admission (p = 0.015). Patients diagnosed with mTBI had a mean symptom burden (Rivermead Questionnaire total score) of 22.6 (SD = 16.3), significantly higher than those diagnosed with head injury (M = 10.9, SD = 12.5; p<0.001). Patients diagnosed with mTBI were more likely to receive an educational handout at discharge (Est [SE] = 0.54 [0.13], p<0.001). Return to EDs/hospitals within 30 days was not predicted by provider diagnosis.



Conclusion

Understanding which factors influence provider diagnosis is invaluable to successful efforts to improve the recognition and management of mTBI in EDs. We report that, albeit to a lesser extent, patients diagnosed with head injury commonly report mTBI symptoms and receive educational materials. Interestingly, these educational materials describe common signs and symptoms of mTBI during recovery. ED provider diagnosis of head injury might not be driven by a lack of recognition of possible mTBI. Shifting towards a diagnosis consistent with intracranial injury may facilitate more consistent and accurate dispersal of educational resources. Clarifying influences on provider preference will aid in developing interventions to unify diagnostic practices across clinical and research settings.		2025 AMA Research Challenge – Member Premier AccessFaizaan Khanfree
61
129594The Comparative Effects of Statins and Fibrates on the Onset of Diabetic Retinopathy in Type 2 DiabetesTitle
Comparative Effects of Statins and Fibrates on the Onset of Diabetic Retinopathy in Type 2 Diabetes

Background
Diabetic retinopathy (DR) is a common complication of type 2 diabetes mellitus (T2DM) and a leading cause of vision loss in working-age adults. Management often requires invasive, costly procedures to preserve vision and while lipid-lowering drugs (LLDs), particularly statins and fibrates, have shown promise in slowing DR progression, their role in disease onset remains unclear. This study evaluated the association between statin or fibrate therapy and DR onset to assess their prophylactic potential.
Methods
This retrospective cohort study used de-identified electronic health records from 152 healthcare organizations via the TriNetX Global Network. Patients with T2DM and no prior ocular complications, diagnosed at least 10 years prior, were assigned to fibrate-only (n = 7,004) or statin-only (n = 463,588) cohorts. Drug exposure required two records: one within 6 months and one at least 1 year after diagnosis. After propensity score matching on demographic characteristics, retinal vascular and non-diabetic retinopathy conditions, and serum HDL, LDL, triglycerides, and HbA1c, each cohort included 6,691 patients. Outcomes included diagnosis of nonproliferative diabetic retinopathy (NPDR), proliferative diabetic retinopathy (PDR), or need for antiVEGF injections or retinal laser. Patients with outcomes prior to cohort entry were excluded.

Results
Fibrate therapy was associated with lower risk of DR across all stages: 55.1% reduction in mild NPDR (RR = 0.449; 95% CI: 0.359, 0.561; p < 0.001), 68.7% in moderate (RR = 0.313; 95% CI: 0.202, 0.486; p < 0.001), 65.3% in severe (RR = 0.357; 95% CI: 0.174, 0.735; p < 0.004), and 67.6% in PDR (RR = 0.324; 95% CI: 0.203, 0.517; p < 0.001). Onset was also delayed: mild (log-rank x2 = 25.238; p < 0.001), moderate (log-rank x2 = 17.419; p < 0.001), severe (log-rank x2 = 5.140; p < 0.024), and PDR (log-rank x2 = 15.125; p < 0.001).
Fibrate users had a 66.7% reduced risk of retinal laser (RR = 0.333; 95% CI: 0.212, 0.524; p < 0.001) and longer time-to-event (log-rank x2 = 14.372; p < 0.001). No difference was found in antiVEGF injection risk.

Conclusion
Fibrate therapy was associated with reduced risk and delayed onset of DR in patients with T2DM compared to statin therapy. These findings support a potential role for fibrates in DR prophylaxis.
2025 AMA Research Challenge – Member Premier AccessFady Tawfik, Thomas Mansy, Matthew Morkos, Luke Nelson, Caitlin Coyne, Jesse Maynardfree
62
129591The Association of Chronic Pain and Complementary Health Approach Usage: A Comprehensive Analysis Adjusting for DemographicsBackground
Chronic pain imposes a significant global burden, contributing to reduced quality of life and increased healthcare costs. Pain's multifaceted nature complicates the development of standardized treatment plans that meet diverse patient needs. Patients increasingly adopt complementary health approaches (CHAs), yet utilization patterns and clinical effects remain understudied. Recently acquired data were used to quantify the association between chronic pain and CHA utilization in U.S. adults, focusing on osteopathic manipulative treatment and other non-traditional modalities, and to determine whether any observed associations persist after adjusting for demographic factors.
Methods
Pre-existing data from a cross-sectional online survey of a representative sample of 3,022 U.S. adults were examined to identify patterns of CHA use among individuals diagnosed with chronic pain in the past year. Statistical analysis included Chi-Square tests of bivariate associations of chronic pain and types of CHA use. Logistic regression analyses, with and without adjustment for demographic variables, evaluated the clinical significance of these associations. CHAs were categorized into the five NIH-defined groups: Physical, Psychological, Nutritional, Combined, and System-Based, encompassing 30 distinct interventions. CHA usage, favorable expectation, and favorable experience of these modalities were analyzed.
Results
Individuals with chronic pain demonstrated significantly higher overall utilization of CHAs than those without (adjusted odds ratio = 2.36, 95% CI: 1.44–3.87). Osteopathic manipulative treatment (adjusted odds ratio = 3.29, 95% CI: 2.15–5.02) was among many CHA modalities more commonly used by these individuals. Although both groups had low expectancy for positive outcomes, those with chronic pain reported significantly lower expectations for various modalities, such as Any Nutritional CHA use, herbalist, non-RX diet, prayer, mindfulness, meditation, and yoga. Reported favorable experience for Any CHA use among those with chronic pain was 93.5%. Various Physical CHAs, acupuncture, and art, music, or dance therapy were reported to be high in both groups, but favorable experience was higher in chronic pain patients.
Conclusion
The preference for CHAs among chronic pain patients demonstrates the unmet needs in conventional care regarding symptom management, functional recovery, and mind-body interactions. While the cross-sectional design limits causal interpretation and self-reporting may introduce bias, these results support further longitudinal research to evaluate efficacy, drivers of utilization, and long-term outcomes.		2025 AMA Research Challenge – Member Premier AccessAshlyn Achesonfree
63
128783Efficacy and Safety of Idecabtagene Vicleucel in Relapsed/Refractory Multiple Myeloma### Background
Idecabtagene vicleucel (ide-cel) is a B-cell maturation antigen (BCMA)-directed chimeric antigen receptor (CAR) T-cell therapy, it has emerged as a transformative treatment for relapsed/refractory multiple myeloma (R/RMM). On April 4, 2024, the U.S. Food and Drug Administration (FDA) approved ide-cel for adult patients with R/RMM who have received at least two prior lines of therapy, including an immunomodulatory agent, a proteasome inhibitor, and an anti-CD38 monoclonal antibody. This meta-analysis evaluates the efficacy and safety outcomes of ide-cel from current clinical evidence.

### Methods
A systematic search of Medline, Embase, Cochrane Library, Web of Science, and Scopus was conducted through July 2025 and supplemented with data from ClinicalTrials.gov. Five relevant clinical trials (n=730 patients) were included. We analyzed outcomes such as overall response rate (ORR), complete response (CR) rate, progression-free survival (PFS), and safety endpoints such as cytokine release syndrome (CRS), neurotoxicity, infections, and treatment-related mortality (TRM). Pooled estimates were calculated using weighted averages based on sample size.

### Results
The pooled ORR was 72.9%, with a CR or better rate of 30.2%. Median PFS across studies was 9.7 months. CRS occurred in 81.8% of patients, with grade ≥3 CRS in 3.7%. Neurotoxicity was reported in 27.1% of patients, including 4.6% with grade ≥3 events. Infections were noted in 44.8% of patients, and TRM was 5.5%, primarily due to infections or immune-mediated toxicity. CRS and neurotoxicity were largely early-onset and manageable with tocilizumab and corticosteroids. Toxicity incidence increased with higher CAR T-cell doses and advanced disease burden.

### Conclusion
Ide-cel demonstrates strong efficacy in heavily pretreated R/RMM patients, with high response rates and a median PFS approaching 10 months. The safety profile is characterized by frequent but manageable CRS and neurotoxicity, and infections remain a key consideration. These findings support ide-cel as a valuable treatment option in R/RMM, in line with its recent FDA approval. It is suggested to maintain ongoing monitoring and long-term follow-up to fully assess durability of benefit and late-onset toxicities or even second primary malignancies.		2025 AMA Research Challenge – Member Premier AccessSelia Chowdhuryfree
64
128782Ophthalmic Conditions in Patients with Multiple Sclerosis### Objective
Multiple sclerosis (MS) is a chronic autoimmune disorder of the central nervous system frequently associated with visual symptoms. This study aimed to characterize the prevalence of ophthalmic conditions in individuals with MS and compare it to matched controls without MS.

### Methods
A matched case-control, cross-sectional study was conducted using data from the All of Us Research Program, a large, nationwide NIH-funded initiative. Individuals with MS (n = 3771) were matched 1:3 to controls without MS (n = 11,313) by age, sex, and race/ethnicity. Four ophthalmic conditions—optic neuritis, diplopia, nystagmus, and uveitis—were examined. Based on electronic health records, diagnosis dates were categorized into ophthalmic conditions occurring before, concurrently with, or after the MS diagnosis. Prevalence rates of ophthalmic conditions in individuals with MS were compared to those of matched controls, through univariate and multivariate logistic regression models, adjusting for sociodemographic factors, lifestyle variables, prior brain injury, family history of MS, and comorbidities.

### Results
All ophthalmic conditions were significantly more prevalent in individuals with MS than controls. Optic neuritis was most common (17.0 % vs. 0.2 %; adjusted odds ratio (aOR) = 31.96; 95 % CI: 22.09–48.20; p < 0.001), followed by diplopia (10.1 % vs. 1.0 %; aOR = 4.88; 95 % CI: 3.87–6.19; p < 0.001), uveitis (3.4 % vs. 0.7 %; aOR = 2.30; 95 % CI: 1.66–3.19; p < 0.001), and nystagmus (2.5 % vs. 0.2 %; aOR = 6.97; 95 % CI: 4.31–11.62; p < 0.001). The temporal analysis revealed that some cases of optic neuritis (31.9 %), nystagmus (21.9 %), diplopia (27.7 %), and uveitis (38.3 %) preceded MS diagnosis. Optic neuritis prior to MS was significantly more prevalent than matched controls (5.4 % vs. 0.2 %; aOR = 7.77; 95 % CI: 5.25, 11.94; p < 0.001).

### Conclusion
This study presents population-based data on the frequency of ophthalmic disease in patients with MS. Ophthalmic conditions, particularly optic neuritis, were significantly more common in individuals with MS compared to matched controls. The study also characterizes the prevalence of uveitis, which is relatively rarer than other ophthalmic conditions in the MS population, yet still more common than matched controls. Some visual conditions occurred before the clinical diagnosis of MS, some concurrently, and others afterward. These findings underscore the importance of recognizing visual disturbances in the diagnosis and management of MS and highlight the need for close collaboration between neurologists and ophthalmologists in the care of patients with MS.		2025 AMA Research Challenge – Member Premier AccessHa-Neul Yufree
65
128781Enhancing the Precision of Lipid Management Quality Assessment in ASCVD Using Large Language Models### Background
Quality assessments relying solely on structured EHR data do not capture note-documented explanations for apparent non-adherence to lipid management guidelines. Large language models (LLMs) present a promising approach for extracting this information at scale.

### Purpose
To evaluate the accuracy and impact of LLMs when applied to assess adherence to lipid management guidelines in CAD.

### Methods
We used structured EHR data to analyze LDL-C levels and statin use of patients with ASCVD seen in cardiology clinics at a large academic medical center from 2022-2024. We then applied GPT-4o with zero-shot, chain-of-thought prompting to free-text cardiology clinic notes to (1) identify external LDL-C values documented within the notes of patients without a recent in-system lipid panel and (2) infer reasons for statin non-use among those with LDL-C ≥70 mg/dL. We validated GPT-4o's accuracy through manual review of 500 randomly selected charts.

### Results
Among 8,595 patients (median age 71.0 years, 61.5% male, 70.8% White) initially identified, 53.6% were on high-intensity statins and 38.1% had LDL-C <70 mg/dL. GPT-4o confirmed a CAD diagnosis in 6,534 (76.0%) patients (Figure); within this group, GPT-4o identified external lipid panels in n=403 (6.2%), high-intensity statin use (n=59, 0.7%), and statin intolerance (n=804, 12.3%) documented in notes but not captured in structured fields (PPV 0.91-1.00). Among patients with GPT-4o-confirmed CAD and updated medication/laboratory data had higher measured rates of high-intensity statin use (61.4% vs. 53.6%) and LDL-C <70 mg/dL (44.1% vs. 38.1%). Excluding statin-intolerant patients further increased high-intensity statin use to 68.4%. Among patients with LDL-C ≥70 mg/dL not on high-intensity statins (N=1,164), GPT-4o-extracted reasons for nonuse were clinician decision (13.0%), intolerance (38.5%), patient preference (8.2%), and undocumented (40.4%). GPT-4o classification performance was strong (F1 > 0.71 across all categories), with the highest performance observed for statin intolerance (F1 = 0.96).

### Conclusions
GPT-4o can accurately identify clinical factors not captured in structured EHR data. This approach enables more accurate guideline-based quality assessments and yields actionable insights into reasons for apparent gaps in care.		2025 AMA Research Challenge – Member Premier AccessDanh Nguyenfree
66
128779Heart Failure-Associated SNPs Are Enriched in Estrogen-Related Receptor Binding Regions### Introduction
Estrogen-related receptors (ERRs) are nuclear receptors essential for postnatal cardiac maturation. Our recent studies have shown that ERRs are necessary for coordinated activation of cardiomyocyte metabolic and structural gene programs through interactions with PGC-1 coactivators and cardiogenic factors such as GATA4. In heart failure (HF), ERR transcriptional programs may revert to a fetal-like state, leading to metabolic inefficiency and energy starvation.

### Methods
To better understand ERRs’ role in HF, 176 SNPs from a recent HF GWAS were mapped onto the human cardiomyocyte cistrome to identify potential HF-associated variants proximal to ERRγ and/or ERRα binding regions. Overlaps were analyzed for ERR binding motifs with an 80% position-weight matrix threshold. Permutation testing was performed to determine if HF-associated variants were found more frequently than expected near ERR binding regions. Hypergeometric testing assessed whether HF-associated SNPs were enriched in ERR binding regions compared to genome-wide SNPs, with per-chromosome p-values combined using Fisher’s method. Stratified linkage disequilibrium score regression (S-LDSC) was performed to quantify whether SNPs within ERR binding regions disproportionately contributed to HF heritability compared to genome-wide averages.

### Results
Both permutation and hypergeometric testing showed significant enrichment of HF-associated variants proximal to ERR binding regions (p < 0.05). The number of overlaps increased with the size of the flanking region, ranging from 48 at 5 kb to 135 at 50 kb. Consensus ERR motifs were identified in 6.4% of ERRγ overlaps and 4.0% of ERRα overlaps. S-LDSC demonstrated notable HF heritability enrichment within ERRγ binding regions (enrichment score: 8.8 ± 3.0, p = 0.010), whereas ERRα showed non-significant enrichment (3.8 ± 3.3, p = 0.426), and GATA4 showed negligible enrichment (−0.4 ± 5.27, p = 0.786).

### Conclusion
This computational analysis reveals significant enrichment of HF-associated SNPs in or near genomic regions containing ERR transcription factors, highlighting ERR-driven transcriptional networks as potentially important mechanisms underlying heart failure development. These results establish a platform for future functional experimental studies to determine the impact of HF-associated variants on ERR-mediated transcriptional regulatory function and additional computational analyses to clarify the precise biological roles of ERRs in HF progression.		2025 AMA Research Challenge – Member Premier AccessZiad Hassanfree
67
128778Plasma Biomarkers for SDHx Deficient Pheochromocytomas and Paragangliomas### Background
Pheochromocytomas and paragangliomas (PPGLs) are rare neural-crest derived tumors with one of the highest germline heritability rates of all neoplasias (40%) and lead to lifelong surveillance. Succinate dehydrogenase (SDHx)-deficient PPGLs are associated with an aggressive disease course, with an overall survival of 50% at 5 years for individuals with metastatic disease. Predicting primary tumor development or metastatic disease progression throughout the life course with imaging/plasma metanephrines and histopathological scoring systems have limitations. The discovery of metabolic rewiring of SDHx deficiency at the tumor level holds promise as a non-invasive disease biomarker.

### Methods
Harnessing the connection of SDHx to metabolic dysfunction, we performed prospective plasma metabolomics (liquid chromatography/mass spectrometry) on 122 well-phenotyped patients, 49.1% of whom were SDHx carriers, at one or multiple time points during a patient’s clinical journey. Concurrently, tissue (adrenal, muscle, heart, brain) and plasma metabolomics of a novel SDHx deficient murine model were utilized for validation of patient-derived biomarkers.

### Results
Metabolomics of patient plasma samples revealed that succinate is a sensitive and specific biomarker of SDHx deficiency (ROC 0.825 – 75% sensitivity and 79.5% specificity), and succinate:glycine ratios improved the ROC to 0.856. With deep phenotyping, we found that age, BMI, and sex trended with succinate. Specifically, succinate levels were significantly correlated with female sex and increasing age. Of those with tumors, succinate had a ROC of 0.933, distinguishing SDHx deficient (including WT gastrointestinal stromal tumors (GISTs), which lack clinical biomarkers of disease) and proficient tumors. Five individuals with SDHx deficient tumors with multiple blood samples developed progressive disease, all of which were correspondingly reflected in succinate elevations. Succinate was not sensitive to treatment response (of both radiotherapy and surgery). Murine SDHx deficient adrenal glands had significantly elevated succinate concentrations however there were no significant changes in succinate plasma concentrations compared to WT mice.

### Conclusion
These findings suggest that plasma metabolomics holds promise as a screening test, especially in resource-poor settings or when rapid clinical decision making is necessary, as genetic testing can often take weeks, if not months, in some clinical settings. Prior to clinical application, further definition of the normal distribution of succinate levels in the general population is required. The mechanism underlying the unexpected finding of elevated succinate in SDHx carriers is a focus of future research. Overall, plasma metabolomics is a novel companion diagnostic for asymptomatic SDHx carriers and individuals with SDHx deficient tumors, of not only PPGLs but also GISTs.		2025 AMA Research Challenge – Member Premier AccessYasemin Colefree
68
128777Understanding the Impact of Race on White Matter Hyperintensities in Alzheimer’s Disease: Findings from the ADNI Study### Background
White matter hyperintensities (WMH) are MRI-detected markers of small vessel disease associated with cognitive decline and Alzheimer’s disease (AD). African Americans (AAs) are disproportionately burdened by cerebrovascular risk factors, yet remain significantly underrepresented in WMH research. This disparity limits the generalizability of findings and may obscure key determinants of disease progression in diverse populations. Identifying how race and related factors influence WMH burden is critical to improving equity in Alzheimer’s care and prevention. This study aimed to examine whether racial differences exist in WMH burden across the AD spectrum, while accounting for other demographic and clinical variables.

### Methods
We analyzed cross-sectional MRI and clinical data from 1,649 participants in the Alzheimer’s Disease Neuroimaging Initiative (ADNI): 682 cognitively normal (CN), 584 with mild cognitive impairment (MCI), and 383 with AD. Of these, 124 participants (7.5%) identified as African American: CN (n=70), MCI (n=39), and AD (n=15). Linear mixed effects models were used to evaluate the effect of race, cognitive status, time since first MRI, age, sex, education, intracranial volume (ICV), MRI acquisition type, and APOE ε4 status on total WMH volume.

### Results
Older age, greater time since MRI, and larger ICV were significantly associated with increased WMH burden across cognitive groups. Use of FLAIR imaging was negatively associated with WMH volume, reflecting differences in image acquisition sensitivity. Higher educational attainment was linked to lower WMH burden, suggesting a possible neuroprotective effect. The most pronounced associations were observed within the AD subgroup. After adjusting for all covariates, race was not a statistically significant independent predictor of WMH burden.

### Conclusion
Although race did not independently predict WMH burden, other structural and demographic factors, including age, education level, and MRI technique, significantly influenced WMH volume. These findings highlight the complex interplay between biological, social, and technical variables in shaping cerebrovascular brain aging. Greater inclusion of African American participants in future studies is essential to better understand race-related mechanisms and to design interventions that equitably address WMH accumulation and its cognitive consequences.		2025 AMA Research Challenge – Member Premier AccessVictor Ekutafree
69
128776Oligodendrocyte genomic stress-induced interferon responses drive progression of Alzheimer’s DiseaseAlzheimer’s disease (AD) is a debilitating dementia that affects about seven million Americans. Oligodendrocytes and myelin degeneration have been noted as features of AD, with demyelination beginning in preclinical disease. However, the mechanisms underlying oligodendrocyte and myelin decay remain poorly understood. Here, we show how transcriptional stress coupled-DNA damage accumulates in preclinical disease, and outline a novel pathway downstream of this DNA damage resulting in activation of the Type I interferon response. In vitro, we find that upon DNA damage, oligodendrocytes activate Stimulator of Interferon Genes (STING), a mediator of the DNA damage-Type I interferon signaling axis, and this drives myelin deficits in vitro. Oligodendrocyte lineage-specific deletion of STING in the 5XFAD murine AD model reduces demyelination, microgliosis, neurodegeneration, and plaque formation. Finally, loss of STING in oligodendroglial cells rescues deficits in spatial memory and learning in the 5XFAD model, as shown by Barnes Maze latencies. Our findings uncover a previously overlooked and key role for oligodendrocytes as active drivers of neurodegeneration in Alzheimer’s Disease.2025 AMA Research Challenge – Member Premier AccessTula Raghavanfree
70
128775Lipoprotein(a) is a Prognostic Marker of Extracoronary Atherosclerotic Vascular Disease Progression### Background
Despite current treatment strategies for atherosclerotic vascular disease focusing on lifestyle modification and lowering cholesterol, a significant residual risk of major atherosclerotic complication remains, prompting investigation into lipoprotein(a) [Lp(a)] as a potential predictive biomarker. The objective of this study was to determine the utility of Lp(a) in identifying patients at high risk of incident extracoronary atherosclerotic vascular disease and complications.

### Methods
Data from 460,544 participants in the UK Biobank with prospectively measured Lp(a) concentrations were included in this analysis. Cox proportional hazards regressions modeled the associations of Lp(a) concentrations with incident peripheral arterial disease (PAD) and incident carotid artery stenosis, and progression to the first major adverse limb event (MALE) and the first stroke, respectively.

### Results
Of the study participants, the median [IQR] age at study enrollment was 58 [51 – 64] years, 54.2% were male, 94.4% European, 5.5% had diabetes, and 10.5% were smokers. Over a median follow-up time of 13.6 [12.9 – 14.4] years, 6,347 (1.4%) and 1,972 (0.43%) developed the first incidence of PAD and carotid stenosis, respectively. Among participants with prevalent PAD and carotid stenosis at enrollment, 196 (2.7%) and 67 (1.9%) progressed to the first incidence of MALE and stroke, respectively. Median Lp(a) concentrations were significantly different in those without atherosclerotic vascular disease at 19.5 nmol/L [7.6-73.5] compared to incident PAD at 25.3 nmol/L [8.3-107.3], progression to MALE at 33.3 nmol/L [8.7-158.2], incident carotid stenosis at 29.5 nmol/L [8.5-116.3], and carotid stenosis progression to stroke at 37.8 nmol/L [11.1-158.3]. The risk estimate per 75 nmol/L Lp(a) for incident PAD (HR 1.18, 95% CI 1.15-1.20, p-value<0.0001) was similar to incident carotid stenosis (HR 1.17, 95% CI 1.13-1.20, p-value<0.0001). Among participants with PAD, those with high Lp(a) concentrations were at 1.57 times the risk of developing MALE than participants with normal Lp(a) concentrations (95% CI 1.14-2.16, p-value=0.006). Among participants with carotid stenosis, participants had 1.40 times the risk of developing an ischemic stroke with high Lp(a) concentrations, although this was not significant (95% CI 0.81-2.40, p-value=0.228).

### Conclusion
High concentrations of Lp(a) are associated with both incident extracoronary atherosclerotic vascular disease and progression to major complications.
2025 AMA Research Challenge – Member Premier AccessTiffany Bellomofree
71
128774Development and Characterization of Levodopa Microneedle Patches for Parkinson's DiseaseLevodopa, co-administered with the dopa decarboxylase inhibitor carbidopa, remains the gold standard for treating Parkinson's Disease; however, due to rapid degradation, only about 1% of orally administered levodopa crosses the blood-brain barrier.
This study developed and validated a robust high-performance liquid chromatography method to quantify levodopa, achieving reliable linearity across multiple replicates. Various buffer systems and pH conditions were tested, with phosphate buffered saline (PBS) at pH 6 and supplemented with glutamic acid hydrochloride offering the most stable environment. With stability confirmed, levodopa was successfully incorporated into a hydrogel matrix using poly(vinyl alcohol) (PVA), and hydrogel-forming microneedles using poly(methyl vinyl
ether/maleic acid), polyethylene glycol, and sodium bicarbonate. These microneedles
demonstrated sufficient mechanical strength to penetrate Parafilm “M” skin simulants, suggesting potential for effective skin permeation.
This work represents a significant step toward creating an alternative levodopa delivery system that could reduce gastrointestinal side effects, improve patient adherence, and offer more consistent plasma drug levels, thereby alleviating the motor fluctuations common with oral levodopa therapy. Although not curative, the finding supports further investigation into in vitro permeation studies and eventual clinical translation of microneedle-based patch system for Parkinson’s disease management.		2025 AMA Research Challenge – Member Premier AccessTatyanna Goodsonfree
72
128773Trends and Determinants of WIC Participation Among Pregnant Women in California: A Decade-Long Population Study### Background
The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) provides essential nutritional, health, and educational support to low-income pregnant women; however, little is known about trends in program participation and associated maternal characteristics this past decade. We aim to investigate temporal trends and maternal predictors of WIC participation during pregnancy in California from 2010-2021, focusing on sociodemographic, geographic, and behaviors. Understanding these predictors is critical for informing equitable access and targeted policy interventions.

### Methods
We conducted a retrospective cohort analysis of 5,529,722 singleton births among California residents using population-based birth cohort data from 2010 to 2021. Descriptive statistics, linear trend analyses, and multivariate logistic regression models were used to assess prevalence and predictors of WIC utilization during pregnancy. Odds ratios (ORs), 95% confidence intervals (CIs), and p-values were reported in quantified associations. Statistical significance was defined as p < 0.05.

### Results
WIC participation declined significantly over the study period, from 54.1% in 2010 to 35.0% in 2021 (β=–1.83, p<.0001, R²=0.98). Multivariate analysis identified several independent predictors of WIC utilization. Compared to mothers aged 40–54 years, those under 20 had the highest odds of WIC use (OR = 3.12; 95% CI: 3.06–3.19). African American (OR = 3.126), Hispanic (OR = 2.96), and American Indian mothers (OR = 1.53) were significantly more likely to participate than White mothers. A strong inverse gradient was observed; mothers with less than high school education had nearly sixfold higher odds (OR = 5.84) than those with a bachelor’s degree or higher. Utilization was higher among foreign born mothers (OR = 1.53) and in regions such as San Joaquin Valley (OR = 2.72) and Los Angeles County (OR = 2.55) compared to Greater Bay Area. Mothers with Medi-Cal (OR=8.09), first-trimester prenatal care (OR = 1.07), higher prepregnancy BMI (e.g., OR = 1.91 for Obese III), vaginal deliveries (OR = 1.02), and maternal smoking during both trimesters (OR = 1.17) were significantly more likely to participate.

### Conclusion
WIC utilization during pregnancy has declined markedly in California, despite persistent disparities by age, race/ethnicity, education, nativity, geographic region, and healthcare access. Addressing structural barriers and improving outreach is essential to enhance equitable WIC engagement and improve maternal-child health.		2025 AMA Research Challenge – Member Premier AccessSophie Sankarfree
73
128772Leveraging Patient-Provided Data in the Electronic Health Record to Improve Care for Veterans### Background
Veterans of the armed forces are the embodiment of our nation's security and freedom. The health challenges they face are often exacerbated by service-related issues including psychological trauma, excess repetitive noise, radiation exposure, and subjection to dangerous chemicals from Agent Orange, burn pit smoke, water contamination, asbestos, and more. [ref PMID: 34070145] Veterans experience a higher prevalence of complex health conditions, including increased risk of heart and lung diseases, diabetes mellitus, and various cancers. [ref PMID: 39696828] Providing high-quality equitable care to veterans requires targeted healthcare strategies that address their specific needs.

### Methods
To enhance the quality of care provided to our veteran population at Geisinger, we integrated inquiries about military service and health requirements into an 8-question “About Me” questionnaire. This initiative aimed to establish a clearer connection between veterans and the medical services they require. Utilizing the SlicerDicer® tool in the Epic electronic health record system, we analyzed our veteran population to identify prevalent health risks with the goal of using these data to develop targeted care programs.

### Results
Veterans (n=35,648) had higher overall risk scores for various health conditions compared to the non-veteran population (n=507,773). Specifically, the risk ratios for chronic obstructive pulmonary disease (COPD) were 2.09, for hypertension 1.60, for heart failure 2.47, for diabetes 1.74, and for other cancers 2.19. Among Vietnam veterans (n=5,441), the risk ratios for certain conditions indicated increased vulnerabilities: hypertension at 1.79, type 2 diabetes at 1.99, prostate cancer at 2.89, skin cancer at 2.44, and atherosclerosis at 2.18. Gulf War veterans (1990-present, n=7,586) showed significantly higher relative risks for post-traumatic stress disorder (PTSD) at 7.00, COPD at 3.55, hypertension at 2.84, asthma at 1.58, and lung cancer at 2.14.

### Conclusion
Our findings highlight the critical need for tailored healthcare initiatives for Vietnam and Gulf War veterans, who are the most at-risk populations among veterans. These findings support the development of specialized screening, prevention, and treatment protocols that recognize both the physical and psychological effects of military service as seen in previous studies. This information will be used to create a value-based care guide aimed at optimizing care for our veteran population, as many utilize community healthcare facilities outside of the VA, and this fills an urgent gap in educating healthcare teams about veteran health needs.		2025 AMA Research Challenge – Member Premier AccessSteven Lecatesfree
74
128771Waste to Wonder: Understanding endometrial interactions using a menstrual-effluent derived model### Background
Infertility is a rising health concern that affects approximately 13% of reproductive age women, but advancements in treatment are limited by our understanding of endometrial function. Decidualization is the process of endometrial stromal cell (eSC) differentiation from a fibroblastic state to a secretory state that occurs during the mid-secretory phase of the menstrual cycle and is associated with a dramatic influx of uterine natural killer (uNK) cells to the endometrium. Defects in decidualization and uNK cell function are associated with infertility, implantation failure, and pregnancy loss, but mechanisms of eSC-uNK cell interaction are not well defined. In this study, we developed and validated uNK cell isolation and expansion techniques using menstrual effluent (ME) as a source of endometrial tissue, then used these ME-derived uNK cells to investigate interactions between uNK cells and decidualizing eSCs to further understand cellular processes occurring prior to embryo implantation.

### Methods
ME samples were collected from participants of the Research OutSmarts Endometriosis (ROSE) study at the Feinstein Institutes for Medical Research. uNK cells were purified by sequential magnetic isolations and expanded in vitro using IL-15 and IL-2. These ME-derived uNK cells were transcriptionally verified by scRNAseq and compared to reference datasets. For co-culture experiments, eSC decidualization was induced by cAMP + MPA. Interactions between uNK cells and decidualizing eSCs were characterized by live-imaging microscopy, eSC staining, and scRNAseq. Functional mechanisms of uNK-eSC interaction were described by protein and gene expression using ELISA and qPCR analyses.

### Results
This novel uNK cell isolation method achieved a purity of up to 92% CD56bright CD16- uNK cells. Cultured ME-derived uNK cells exhibited up to 30-fold expansion, maintained viability for approximately 30 days, and retained phenotypes of reference decidua-derived uNK cells as assessed by flow cytometry and single cell RNA sequencing. In an ME-derived model of uNK-eSC interaction, uNK cells mediated contact-dependent eSC remodeling during decidualization as determined by live-imaging microscopy and cell density analysis. uNK cells concurrently stimulated a pro-inflammatory eSC response in part via IFN release as determined by transcriptomic and proteomic analysis.

### Conclusion
These results indicate that ME is a viable, non-invasive source of uNK cells that can be used for downstream analyses and suggest that uNK cells mediate both eSC remodeling and inflammation during decidualization. Interactions between uNK cells and eSCs are critical for endometrial receptivity, and this novel ME-derived system greatly expands the available tools to investigate mechanisms contributing to endometrial homeostasis and female fertility.		2025 AMA Research Challenge – Member Premier AccessStephanie Buschfree
75
128770Accuracy of Multimodal Large Language Models vs. Human Consensus in Interpreting Clinical Vignettes### Background
As large language models (LLMs) gain multimodal capabilities to interpret both images and text, their potential to assist in clinical decision-making continues to grow. While prior research has shown LLMs can match and even outperform individual physicians on standardized exams, little is known about their performance relative to collective human reasoning where varied perspectives contribute to more accurate conclusions. This study evaluates the performance of state-of-the-art multimodal LLMs in solving complex medical challenges compared to both individual and aggregate responses from a large pool of medically literate readers.

### Methods
We compiled 100 ophthalmology-related cases from the New England Journal of Medicine (NEJM) Image Challenge archive (2006–2024), each containing a clinical vignette, image, and multiple-choice question. Six multimodal LLMs (GPT-4 Turbo, GPT-4o, GPT-4o mini, Gemini 1.5 Flash, Gemini 1.5 Pro, and Claude 3.5 Haiku) were prompted to answer each case and self-report confidence (1–4 scale). Human performance was assessed using (1) the average accuracy of NEJM respondents and (2) the accuracy of the most commonly selected (modal) answer per case, representing collective judgment. LLM accuracy and confidence were compared using Kruskal-Wallis and post hoc Dunn’s tests (p < 0.05 considered significant).

### Results
All LLMs except Claude 3.5 Haiku outperformed the average individual respondent (mean human accuracy: 50.4%, 95% CI: 40.8–60.0%), with GPT-4o achieving the highest model accuracy at 72.0% (95% CI: 66.7–76.8%). However, the collective human response—the most frequently selected answer—was correct in 93.0% of cases (95% CI: 86.3–96.6%), significantly outperforming every LLM. The most popular response across LLMs (i.e., the answer most frequently selected by the models) was correct in 65.3% of cases (95% CI: 56.0–74.7), significantly lower than the collective human accuracy. Correct answers were associated with higher confidence scores for all LLMs except Gemini Flash, suggesting self-rated confidence may be a useful proxy for model reliability.

### Conclusion
Multimodal LLMs can outperform individual human respondents on image-based diagnostic tasks but fall short of the accuracy achieved by collective human reasoning. This underscores a key strength of human cognition: the diversity of thought across individuals produces more accurate outcomes than any single decision-maker, human or machine. In contrast, LLMs often rely on a uniform reasoning process that can propagate the same errors consistently. While LLMs hold promise as diagnostic aids, further refinement may be required for them to match the collective intelligence of human experts in interpreting clinical vignettes.		2025 AMA Research Challenge – Member Premier AccessSarah Wagnerfree
76
128769Central and Brachial Pressures: Effects on Arterial Stiffness in Older Adults### Background:
Central blood pressure (BP) more accurately reflects the loading conditions of the heart compared to brachial BP. Brachial BP is routinely used as a surrogate for central BP due to its ease of access, lower cost, and the efficiency of measurement. However, using brachial BP as a surrogate for central BP when calculating carotid arterial stiffness (CAS) has not been studied in older adults.

### Methods:
Veterans (n=180) age 65+ were recruited from Madison VA Hospital. Resting supine brachial BP and central BP (estimated from radial artery waveforms, Atcor Medical) were obtained. CAS (Peterson’s elastic modulus [PEM], Young’s elastic modulus [YEM]) and distensibility coefficient (DC) were calculated using brachial and central BP. Differences in CAS were compared using paired Wilcoxon tests. Linear regression models evaluated associations with cardiovascular risk factors (age, sex, hypertension status, diabetes, sleep apnea, alcohol intake, active smoking, and lifetime smoking status).

### Results:
Participants were 70.4 (7.7) years old and 27.8% were female. Average brachial systolic BP and pulse pressure were significantly higher than central (132.3 [18.6] mmHg vs 123.8 [17.7] mmHg, p<0.001) and pulse pressure (53mmHg [16.4] vs 43.8mmHg [15.3], p<0.001). Compared to brachial BP, using central BP to calculate stiffness measures resulted in significantly lower YEM and PEM and significantly higher DC (PEM: 480.6 [209.5] mmHg vs 378.3 [178.4] mmHg; YEM: 2220.2 [926.6] mmHg vs 1746.9 [785.4] mmHg; DC: 2.4 [1.0] x10-3 mmHg-1 vs 3.1 [1.1]
x10-3 mmHg-1; all p<0.001). Absence of hypertension was associated with smaller differences in PEM and DC (PEM: 𝛽=-29.1, SE=12.1, p=0.02; DC: 𝛽=-123.8, SE=55.3, p=0.027), while older age was associated with greater differences in YEM when calculated using brachial vs central BP (𝛽=1.9x10-5, SE=0.69x10-5, p=0.006).

### Conclusion:
Brachial and central BP significantly differ in older adults and result in significant differences in calculated CAS and distensibility. Brachial BP tends to significantly overestimate CAS, especially in those with hypertension. Our study highlights the importance of considering central BP when evaluating CAS to predict cardiovascular health and better understand vascular aging, especially in older adults with hypertension. Relying solely on peripheral BP may lead to an overestimation of vascular risk in these populations. Future studies should prioritize using central
BP measurements or CAS indices that are independent of BP to more accurately assess arterial health and improve CVD risk stratification. This can help better guide antihypertensive therapy, potentially improving outcomes and reducing the risk of over- or under-treatment.		2025 AMA Research Challenge – Member Premier AccessSakar Guptafree
77
128768AI-Assisted MRI Interpretation Improves Surgeon Performance in Diagnosing Bankart Lesions### Background
Bankart lesions (anteroinferior glenoid labrum tears) can cause significant shoulder pain and dysfunction but are often difficult to detect on standard non-contrast MRI. To improve diagnostic sensitivity, MRI arthrography (MRA) is frequently used, although it is more invasive, costly, and uncomfortable for patients. Deep learning (DL), a subset of machine learning, identifies complex patterns in large datasets without the need for manual input. DL has improved diagnostic accuracy across several areas of medical imaging and is particularly well suited for detecting subtle abnormalities. This study developed a DL model to detect Bankart lesions on both MRI and MRA and evaluated its clinical utility through a clinician user study.

### Methods
A dataset of 586 shoulder MRIs/MRAs from 546 patients who underwent shoulder arthroscopy within one year of imaging was retrospectively analyzed. Arthroscopic findings served as the reference standard. A custom neural network ensemble was trained on 410 scans (224 MRI, 186 MRA), tuned on 59 (40 MRI, 19 MRA), and tested on 117 (71 MRI, 46 MRA). Gradient-weighted class activation maps (Grad-CAM) were used to assess the anatomical focus of model predictions. To evaluate clinical impact, a user study was conducted with two shoulder/elbow fellowship-trained orthopaedic surgeons and two orthopaedic residents. Each clinician reviewed all 117 test MRIs/MRAs in two phases: once without model predictions (unaided), and again after a 60-day washout period with DL predictions shown (aided).

### Results
For standard MRI, the model achieved 90.1% accuracy, 83.3% sensitivity, and 90.8% specificity—substantially outperforming radiology reports from the same scans (80.3% accuracy, 16.7% sensitivity, 86.2% specificity). On MRA, the model achieved 89.1% accuracy, 94.1% sensitivity, and 86.2% specificity, compared to radiology reports at 84.8% accuracy, 82.4% sensitivity, and 86.2% specificity. Grad-CAM showed consistent model attention on the anterior labrum. In the user study, mean clinician sensitivity improved from 38.0% to 78.3% with model assistance. Specificity slightly decreased (86.7% to 85.4%), while accuracy increased (77.1% to 84.0%). Clinician confidence increased by 0.65 points on a 10-point scale (p < 0.001).

### Conclusion
The model performed comparably on standard MRI to radiologists on MRA, with Grad-CAM confirming that predictions were based on appropriate anatomic regions. The user study demonstrated that clinicians achieved greater sensitivity and confidence when aided by the model. These results highlight the potential for DL tools to close the gap between non-contrast and contrast-enhanced imaging—avoiding the added cost, discomfort, and invasiveness of MRA.		2025 AMA Research Challenge – Member Premier AccessSahil Sethifree
78
128767Rural-Urban Skin Cancer Prevention and Treatment Differences for Medicare Beneficiaries, 2019–2023### Background
One in five adults will develop skin cancer in their lives. Medicare insured adults ages ≥65 account for nearly half of all skin cancers, with a disproportionately higher incidence in rural communities where patients face significant barriers in access to dermatologists. This study aims to fill a gap in the research for this vulnerable patient population by quantifying the relationships between skin cancer prevalence, beneficiary‑service volume of preventative and treatment services, and health care costs for Medicare beneficiaries across the rural-urban continuum from 2019 to 2023.

### Methods
Retrospective analysis of the publicly available 2019-2023 Medicare claims data files and the 2019-2023 National Health Interview Survey. Healthcare Common Procedure Coding System (HCPCS) codes were categorized as skin cancer prevention and treatment services based on the American Medical Association’s CPT Consumer Friendly Descriptors.

Zip codes were cross-walked to four levels of the 2013 National Center for Health Statistics (NCHS) Urban-Rural Classification Scheme: large central metro (LCM), large fringe metro (LFM), medium and small metro (MSM), and non-metropolitan areas (non-metro).

### Results
4,314 Medicare beneficiaries reported a skin cancer diagnosis in their lifetime (8.8%), with a greater incidence of each type of skin cancer in rural areas than urban areas (p<0.03).

Non‑metropolitan areas had a greater share of dermatology beneficiary‑service volume for preventive services (LCM 23.4%; LFM 23.6%; MSM 25.7%; non‑metro 26.7%; P < 0.001) and services were 2-fold more likely to be any skin cancer treatment and 3.5-fold more likely to be radiation‑related treatment (OR 2.00 [1.99–2.01]; 3.52 [3.48–3.57]) than in LCM areas.

After standardization for geographic differences in service payment rates, the average Medicare payment for surgical skin cancer treatment services was greater than for preventative care services, and both were greater in rural areas than urban areas (Overall, $118.54 vs. $48.22; LCM, $118.32 vs. $46.85; LFM, $121.7 vs. $48.41; MSM, $117.34 vs. $48.99; non-metro, $116.36 vs. $53.05; p<0.001 for each). Interestingly, no rural-urban difference was observed in radiation-related or chemotherapy treatment services (p=0.230, p=0.260).

### Conclusion
Rural-urban differences in dermatologists’ skin cancer prevention and treatment beneficiary‑service volume reflect a greater need for preventative services in rural areas of the country and provide more recent, service-level evidence for this vulnerable population.

These findings suggest preventative services for skin cancer may significantly lower health care costs in both the short-term and long-term for Medicare patients and CMS, and these calculations may be useful when evaluating new policy for expanding skin cancer prevention efforts nationwide.		2025 AMA Research Challenge – Member Premier AccessRishi Rayfree
79
128766PGRMC2 Knockout Disrupts Ciliary Extracellular Vesicle Biogenesis and Cardiac Functions### Background
Progesterone receptor membrane component 2 (PGRMC2) is critical for ciliogenesis and the biogenesis of ciliary extracellular vesicles (ciEVs), which play an essential role in maintaining cardiac function. Extracellular vesicles (EVs) are currently being evaluated in eight clinical trials as therapeutic agents for cardiac remodeling. While these studies highlight the therapeutic promise of EVs, the role of ciEVs and their regulation via PGRMC2 remains largely unexplored. This study investigates the impact of cardiac-specific PGRMC2 knockout on heart structure and function and proposes ciEV supplementation as a novel strategy to target cardiac dysfunction, expanding the therapeutic potential of EV-based cardiac therapies.

### Methods
A cardiac-specific PGRMC2 knockout (KO) mouse model was generated using the Cre-LoxP system to investigate the role of PGRMC2 in cardiac structure and function. Mice were randomized into four groups (n=24 total; equal sex ratio): (1) wild-type with saline, (2) wild-type with EV treatment, (3) KO with saline, and (4) KO with EV treatment. Starting at two weeks of age and continuing through 17 weeks, mice received weekly intravenous injections of either 200 µL saline or 1.6 million EVs.
Assessments were conducted weekly and included treadmill exercise testing and blood pressure measurements. At 18 weeks, cardiac tissue was collected for histological evaluation using Masson’s trichrome staining to assess fibrosis, and left ventricular wall thickness. Statistical significance was defined as P < 0.05.

### Results
Survival analysis demonstrated a high mortality rate (50%) in knockout mice and treadmill exercise-stress tests revealed a marked reduction in exercise capacity (p<0.00001) in KO mice. Blood pressure analysis showed a greater decline in diastolic (p<0.0001 before exercise; p<0.00001 after exercise) and systolic (p<0.0001 after exercise) blood-pressure in male KO mice compared to females.
Pressure-volume (PV) loop analysis assessed cardiac performance under baseline (saline), increased workload (epinephrine), and reduced workload (diltiazem) conditions. KO mice exhibited reduced stroke-volume, indicating impaired cardiac output and efficiency.
Histological analysis revealed left ventricular hypertrophy (p<0.0001 female; p<0.001 male) and a significant increase in cardiac fibrosis (p<0.001 female; p < 0.00001 male) in knockout mice. EV treatment significantly improved survival rate (75%), blood-pressure regulation, ventricular thickness, and cardiac fibrosis.

### Conclusion
Loss of PGRMC2 disrupts ciEVs release, leading to structural and functional cardiac deterioration. Extracellular vesicle injections rescue these deficits, highlighting the essential role of ciEVs and PGRMC2 in maintaining cardiac homeostasis, further reinforcing their therapeutic potential.		2025 AMA Research Challenge – Member Premier AccessRiley Dannafree
80
128765Comparison of the 2017 Pooled Cohort Equations and 2025 PREVENT Risk Prediction Models in U.S. Adults with Untreated Stage 1 Hypertension and Without CVD, Diabetes, or CKD: NHANES 2011 to March 2020### Background & Objectives
The 2025 ACC/AHA hypertension guidelines incorporate the Predicting Risk of Cardiovascular Disease Events (PREVENT) equations to guide the initiation of antihypertensive medications for primary prevention of cardiovascular disease (CVD). This study aimed to compare population sizes recommended for antihypertensive treatment according to primary prevention criteria from the 2017 versus 2025 guidelines.

### Methods
Using NHANES data (2011–2020), we identified participants aged 30–79 years who had systolic blood pressure (SBP) of 130–140 mmHg or diastolic blood pressure (DBP) of 80–90 mmHg, and who had no prior diagnosis of hypertension, CVD, diabetes, or chronic kidney disease, and were not pregnant nor taking antihypertensive medication. We estimated the 10-year risk for each participant using the Pooled Cohort Equations (PCE; stroke and myocardial infarction for those aged 40–75 years) and the PREVENT base model for total CVD (stroke, MI, and heart failure for those aged 30–79 years). High risk was defined as a 10-year ASCVD risk ≥10% with PCE and total CVD risk ≥7.5% with PREVENT. Analyses incorporated NHANES sampling weights to generate nationally representative estimates.

### Results
The initial population for both analyses included 29.6 million U.S. adults with SBP 130–140 mmHg or DBP 80–90 mmHg, not on antihypertensive medication, not pregnant and without CVD, diabetes, or chronic kidney disease. Of these, 18.4 million were aged 40–75 years and thus eligible for risk assessment with PCE; among these, 3.3 million had a 10-year ASCVD risk ≥10%. In contrast, 24.0 million were aged 30–79 years and thus eligible for total CVD risk assessment using the PREVENT model; of these, 3.5 million had a 10-year PREVENT base total CVD risk ≥7.5%.

### Conclusion
Under the updated thresholds, PREVENT identified nearly 200,000 more high-risk U.S. adults with untreated stage 1 hypertension than PCE, despite differences in model inputs and age range.

2025 AMA Research Challenge – Member Premier AccessRaul A. Ruiz Melendezfree
81
128764Medicaid Expansion and Overall Mortality among Individuals Living with Lung Cancer### Background
Lung cancer is the leading cause of cancer-related mortality in the United States, with outcomes disproportionately worse among uninsured and socioeconomically disadvantaged populations. The Affordable Care Act (ACA), through Medicaid expansion, sought to improve access to care, yet its long-term impact on lung cancer survival remains incompletely defined.

### Objective
To evaluate the association between Medicaid expansion and overall mortality in patients with lung cancer, and to determine whether effects varied by stage at diagnosis, race/ethnicity, socioeconomic status, and treatment modality.

### Methods
We conducted a retrospective cohort study using the National Cancer Database, including patients aged 40–64 years diagnosed with lung cancer between 2006 and 2021. States were categorized as early Medicaid expansion (by January 2014) or non-expansion. The study period was divided into pre-expansion (2006–2013) and post-expansion (2014–2021). A Difference-in-Differences (DiD) analytic approach within a Cox proportional hazards framework was used to estimate the effect of Medicaid expansion on all-cause mortality, adjusting for demographic, clinical, and socioeconomic factors.

### Results
The study cohort comprised 369,251 patients, predominantly White (74.7%) and male (52.7%), with a mean age of 56.8 ± 5.6 years. A total of 165,364 individuals (44.8%) resided in early Medicaid expansion states. Medicaid expansion was associated with a 45.2% reduction in overall mortality. Stratified analyses demonstrated reductions in the hazard of overall mortality across all racial and ethnic groups, with the greatest benefit observed among Hispanic patients (DID: –0.506; 95% CI, –0.729 to –0.283; p < 0.001). Expansion states also showed increased early-stage diagnoses, reduced uninsured rates, and higher surgical utilization following the ACA.

### Conclusions
Medicaid expansion was significantly associated with improved lung cancer survival across demographic and clinical subgroups. These findings underscore the role of insurance policy in shaping oncologic outcomes and suggest that Medicaid expansion may serve as a critical strategy to enhance access to timely diagnosis and treatment.

2025 AMA Research Challenge – Member Premier AccessPhiwinhlanhla Ndebele-Ngwenyafree
82
128763A proteomic signature of vascular disease is associated with the aging brain's vulnerability to degeneration preceding accumulation of amyloid pathology### Background
Aging confers the most significant risk for dementia, yet the underlying mechanisms remain poorly understood. While the amyloid cascade hypothesis has dominated the field for decades, a large proportion of variance in cognitive impairment remains unexplained by classical proteinopathies. Thromboinflammation—a vascular and immune signaling axis involving brain barrier dysfunction and protein extravasation—may represent an upstream initiator of neuronal vulnerability that precedes amyloid pathology. Recent work by our colleague shows fibrinogen, a peripheral coagulation protein, to exhibit one of the earliest age-related increases in cerebrospinal fluid, yet its role as a driver of brain degeneration in living humans remains unexplored.

### Methods
We present the most comprehensive investigation of thromboinflammation as a driver of degenerative brain changes in aging living humans. We integrated high-dimensional cerebrospinal fluid proteomics (~7,000 proteins) across 1,655 individuals from three independent cohorts with single-cell transcriptomics, brain imaging, and cognitive data. Using CSF fibrinogen levels as a quantifiable marker of brain barrier dysfunction, we employed systematic mediation analysis to identify molecular pathways linking vascular compromise to synaptic vulnerability in amyloid-negative, cognitively typical older adults, with validation across independent cohorts and testing in amyloid-positive states.

### Results
We identified a reproducible molecular network linking brain barrier dysfunction to white matter degeneration, synaptic loss, and cognitive slowing—well before amyloid pathology appears. Fibrinogen emerged as a central hub of brain aging biology, with higher CSF levels significantly associating with elevated markers of synaptic dysfunction (pTau181, neurogranin, GAP43, SNAP25) and cognitive slowing in amyloid-negative adults. We identified 53 validated molecular mediators of fibrinogen's neurotoxic effects, enriched in neurovascular unit cells and demonstrating coordinated dysregulation of angiogenesis, fibrosis, and immune signaling. These mediator signatures correlated with compromised white matter integrity in tracts critical for cognitive processing speed. Critically, fibrinogen associated with tauopathy independent of amyloid buildup, and these associations persisted robustly in amyloid-positive individuals while remaining independent of amyloid burden, demonstrating relevance across the neurodegeneration spectrum.

### Conclusion
Our findings offer a paradigm-shifting answer to why aging confers powerful dementia risk by positioning thromboinflammation as an upstream initiator of neuronal vulnerability that challenges the prevailing temporal hierarchy of neurodegeneration. This work reframes vascular disease from comorbidity to causality, introducing a vascular-centered model of brain aging that bypasses classical proteinopathy models. The identified molecular mediators provide an upstream, druggable axis with immediate implications for biomarker development, clinical trial stratification, and brain healthspan-preserving interventions across the aging spectrum. "		2025 AMA Research Challenge – Member Premier AccessJonah Kellerfree
83
128762Neighborhood Socioeconomic Status and Racial Disparities in ADRD in Breast Cancer Survivors### Background
Racial disparities in Alzheimer’s disease and related dementias (ADRD) are well-documented, yet the contribution of neighborhood socioeconomic context to these disparities among breast cancer survivors remains unclear. The Index of Concentration at the Extremes (ICE) captures neighborhood-level deprivation and affluence, potentially mediating racial differences in ADRD diagnosis. Understanding the role of neighborhood-level factors may explain mechanisms behind these disparities.

### Objective
To assess whether neighborhood-level socioeconomic conditions, measured by ICE, mediate the racial disparity in ADRD diagnosis among older female breast cancer survivors.

### Methods
We analyzed SEER-Medicare linked data from 17,741 women aged 65 and older diagnosed with breast cancer (2007–2009) with up to 10 years of follow-up. ADRD diagnosis was identified through diagnostic codes. We conducted causal mediation analysis to estimate the average causal mediation effect (ACME), total effect, and proportion mediated by ICE on the association between race (Black vs. White) and ADRD diagnosis, adjusting for age, comorbidities using the Charlson Comorbidity Index (CCI), treatment type, and neighborhood -level poverty indicator.

### Results
Race was significantly associated with ADRD diagnosis (total effect estimate = 0.052, 95% CI 0.028–0.077, p < 0.001). ICE significantly mediated this association, with an ACME of 0.049 (95% CI 0.035–0.063, p < 0.001), explaining approximately 93% of the total effect. This strongly suggests that neighborhood-level deprivation/advantage (ICE) explains most of the racial disparity in ADRD diagnoses in the dataset, after adjusting for age, neighborhood-level poverty indicator, treatment type, and CCI. The mediation findings remained robust in sensitivity analyses adjusting for additional covariates.

### Conclusion
Neighborhood-level deprivation, as captured by ICE, substantially mediates racial disparities in ADRD diagnosis among breast cancer survivors, highlighting the critical role of social
determinants and structural inequities on cognitive health outcomes. These findings suggest that beyond biological factors, such as hormone-modulating therapy-related cognitive effects, social context profoundly shapes racial disparities in cognitive health outcomes. This study shows the importance of neighborhood socioeconomic context in understanding racial disparities in ADRD diagnosis among breast cancer survivors. Further analyses incorporating additional covariates and stratified mediation will clarify the pathways influencing these disparities. Targeted interventions addressing neighborhood disadvantage may be essential for advancing equity in dementia prevention and care.		2025 AMA Research Challenge – Member Premier AccessNjasi Ojifree
84
128761Preadipocytes as Key Players in Skin Immune Defense Against Staphylococcus aureus Infection### Background
Preadipocytes play a central role in skin immune defense. In response to Staphylococcus aureus infection, they rapidly proliferate, leading to dermal fat expansion. When adipogenesis is impaired, susceptibility to infection increases. During differentiation, preadipocytes and adipocytes produce cathelicidin (CRAMP in mice, CAMP in humans), a vital antimicrobial peptide whose expression is further upregulated by S. aureus exposure. CRAMP accumulates through both neutrophil recruitment and de novo synthesis by preadipocytes, strengthening the skin’s antimicrobial response.

### Methods
C57BL/6 mice were intradermally injected with S. aureus USA300. Preadipocytes were isolated from digested dorsal skin, cultured, and induced to differentiate. Oil Red O staining confirmed adipogenesis. Skin sections underwent H&E and immunofluorescence staining for Pref-1, CRAMP, and Ly6G. Gene expression (Camp, Adipoq, inflammatory markers) was measured by qPCR; protein levels were assessed via Western blotting. CRAMP function was evaluated using neutralizing
antibodies in vitro and in vivo. Bacterial burden was determined through CFU counts. Data were analyzed using unpaired t-tests or ANOVA (p < 0.05).

### Results
In vivo, infection triggered visible skin lesions and dermal fat layer expansion within 24–72 hours. Histology and immunofluorescence confirmed increased Pref-1⁺ preadipocytes in infected dermis. Isolated preadipocytes differentiated into lipid-rich adipocytes in vitro. qPCR showed significant upregulation of Camp and Adipoq in infected tissue; Western blotting confirmed increased CRAMP protein levels. CRAMP localized to both Pref-1⁺ cells and Ly6G⁺ neutrophils, indicating dual
sources. Neutralizing CRAMP resulted in larger lesions and higher bacterial burden (p < 0.01), while in vitro inhibition reduced bactericidal activity. Mice with stronger preadipocyte responses exhibited lower CFUs, linking preadipocyte activation to bacterial clearance.

### Conclusion
Preadipocytes play a critical role in the skin’s immune defense against S. aureus. Beyond serving as adipocyte precursors, they actively contribute to antimicrobial defense through CRAMP production and support of neutrophil recruitment. Their activation and differentiation are essential for limiting infection and promoting bacterial clearance. These findings position preadipocytes as immunologically active cells and potential therapeutic targets to enhance host resistance to bacterial skin infections.
2025 AMA Research Challenge – Member Premier AccessNiketa Dixitfree
85
128760Prevalence and risk factors for severe food insecurity and poor food consumption during a drought emergency in Ethiopia### Background
Frequent drought has heightened nutritional concerns in Ethiopia. This study assesses the prevalence and risk factors of severe food insecurity and poor food consumption in Productive Safety Net Programme households in drought-prone East Hararghe, Ethiopia.

### Methods
Data was from the USAID-funded Resilience Food Security Activity baseline survey in East Hararghe, Ethiopia. Severe food insecurity (n=4628; multivariate n=4335) was defined as Food Insecurity Experience Scale (>7) and poor food consumption (n=4554; multivariate n=4268) was defined as Food Consumption Score (<21). Logistic regression identified adjusted odds ratio and 95% confidence interval of risk and protective factors.

### Results
Severe food insecurity prevalence was 77.79% and poor food consumption was 69.74%. Risk factors for severe food insecurity included women/girls aged 15-19 (1.79; 1.36-2.34), current pregnancy (1.52; 1.17-1.96), history of pregnancy (3.46; 2.76-4.33), cash-earning work (1.35; 1.12-1.61), daily-per-capita food consumption <1.61USD (2.98; 1.91-4.66), crop-planting (1.67; 1.31-2.13), and handwashing facilities (3.83; 1.92-7.63); protective factors included two or more children-under-5 (0.72; 0.60-0.88), raising livestock/crops to sell (0.50; 0.42-0.60), and raising oxen (0.34; 0.26-0.45). Risk factors for poor food consumption included woman/girl (1.44; 1.15-1.81) and household-head no education (1.46; 1.18-1.79), daily-per-capita food consumption <1.61USD (4.01; 2.58-6.21), and financial services (2.10; 1.69-2.59); protective factors included women/girls aged 15-19 (0.59; 0.46-0.76) and 30-49 (0.76; 0.63-0.91), two or more children-under-5 (0.77; 0.64-0.91), current pregnancy (0.57; 0.47-0.70), history of pregnancy (0.70; 0.55-0.89), crop-planting (0.57; 0.44-0.75), raising livestock/crops to sell (0.40; 0.34-0.48) and raising oxen (0.68; 0.52-0.90).

### Conclusion
With high severe food insecurity and poor food consumption prevalence, we identified vulnerable households in drought emergencies as those with pregnant/lactating women, mothers, adolescent girls/women, no education and/or livestock. Notably, households where the WRA respondent was pregnant, had been pregnant, was 15-19 years old, and/or planted crops were more likely to experience severe food insecurity, yet less likely to have poor food consumption. Our findings highlight a conceptual distinction, that women/girls’ pregnancy and/or age status may influence household perception and/or definitions of food security despite adequate consumption. The strongest predictor of severe food insecurity and poor food consumption was daily-per-capita food consumption costs <1.61 USD, and the strongest protectors were households that raised livestock/crops to sell and/or raised oxen. Although crop production appeared insufficient to maintain food security, this may improve food consumption score by increased diversity of nutrient intake. The dual burden of food insecurity and poor consumption threatens current and future generations, and data-driven action can help progress towards the goal of zero hunger in Ethiopia.		2025 AMA Research Challenge – Member Premier AccessNoah Bakerfree
86
128759Testing the Efficacy of Cabozantinib (XL-184) to Treat Glioblastoma (GBM)### Background
Glioblastoma (GBM) is a highly aggressive malignancy with limited therapeutic advancements despite extensive research. Receptor tyrosine kinase (RTK) amplification, a common feature in glioblastoma (GBM), drives tumor growth/recurrence, therapeutic resistance, and invasion, making RTKs attractive therapeutic targets. Currently, Bevacizumab, the only FDA-approved RTK inhibitor in GBM, improves progression-free survival in patients through vascular normalization. This study evaluates the efficacy of cabozantinib (XL184), a multi-kinase inhibitor targeting various RTKs such as VEGFR2, MET, and AXL, in GBM cancer stem cells (CSCs) and patient-derived xenograft (PDX) models, with a focus on uncovering resistance mechanisms and identifying opportunities for combination therapies.

### Methods
Nine genomically diverse GBM CSCs were treated with XL184 to assess drug sensitivity. Reverse Phase Protein Array (RPPA) Proteomics at 24h and 72h was used to evaluate changes in key signaling pathways along with a focused analysis of RTK signaling changes. CSC-derived PDX from two proneural (HF2587, HF3016) and one classical (HF2927) GBM were treated with XL184 monotherapy and in combination with TMZ. Tumor growth and survival were monitored. Immunohistochemical (IHC) analysis of CD31 expression to assess angiogenesis was performed with high-resolution microscopy.

### Results
GBM CSCs exhibited a wide range of sensitivity to XL184 (IC50: 2μM - 34μM). Proteomics showed XL184-mediated inhibition of phospho-VEGFR2, AKT, and ERK, while sub-lethal XL184 doses upregulated phospho-MET and STAT3, suggesting adaptive resistance mechanisms. In vivo, HF2587 (p=0.00022, n>18) and HF2927 (p=0.1847, n>12) xenografts responded to XL184 monotherapy, and in combination with TMZ, with reduced tumor growth and improved survival while HF3016 (p=0.061, n>16) was resistant to monotherapy. HF2587 and HF2927 demonstrated a reduction in vessel density, measured by CD31 expression, compared to normal brain. The non-responsive HF3016 exhibited increased angiogenesis following therapy, a potential indicator of treatment resistance.

### Conclusion
Insights into CSC signaling and tumor-specific dependencies provide a foundation for biomarker-driven, personalized therapeutic approaches in GBM. The vasculature results, assessed through CD31 immunohistochemistry, demonstrate the modulation of angiogenesis and tumor blood supply by XL184, which is crucial for understanding its therapeutic effects on GBM and the tumor microenvironment. These findings highlight the heterogeneity of XL184 response in GBM and the potential for combination strategies to overcome resistance. Moreover, this work builds on limitations of existing RTK therapies (bevacizumab), offering multi-targeted inhibition with microenvironmental modulation. Additionally, it contributes valuable preclinical evidence supporting a rational framework for designing more targeted, effective interventions in a highly treatment-refractory cancer.		2025 AMA Research Challenge – Member Premier AccessMonish Moyalfree
87
128758Cardiac Phenotype and Mitochondrial Effects of Antiretroviral Therapy and PFAS Exposures### Background
The majority of people living with HIV globally are women and girls. Although combined antiretroviral therapy (cART) can drastically improve prognosis and prevent vertical transmission, its off-target metabolic and systemic effects in pregnant women and women of childbearing age remain understudied. Concurrent exposure to ubiquitous environmental pollutants, such as per and polyfluoroalkyl substances (PFAS), may interact with cART to impact maternal health, given their shared potential for mitochondrial and cardiotoxicity. Our broader project investigates the combined structural, functional, and molecular impact of cART and a legacy PFAS compound, perfluorooctanesulfonic acid (PFOS), on cardiometabolic outcomes in a rat model. Here, we present findings from non-pregnant animals, which will later be compared to pregnant counterparts.

### Methods
Female Sprague-Dawley rats were treated with cART (abacavir/dolutegravir/lamivudine; Triumeq at 300/25/150 mg/kg/day) or vehicle (aqueous methylcellulose/Tween 80). Animals also received either PFOS-contaminated water, directly mimicking concentrations measured in a North Carolina community, or vehicle (reverse osmosis-filtered water). They received echocardiography and electrocardiography immediately prior to tissue harvest at study day (SD) 27 or SD 49. Tissue and plasma samples underwent PFAS quantification, pathological evaluation, qPCR, and mitochondrial function analysis.

### Results
The non-pregnant cohort demonstrated perturbations in several structural and functional parameters by exposure group. Echocardiography revealed shortened ejection times at SD 26 in the cART group compared to controls, indicating early systolic dysfunction. Rats in the PFOS + cART group showed slower isovolumetric relaxation times and increased ejection fractions at SD 26. Structural changes included increased interventricular septal thickness, posterior wall thickness, and left ventricular (LV) mass in the cART group. At SD 48, persistent hypertrophy was demonstrated by increased LV mass and heart weight:tibia length ratios for the cART-exposed rats. Surprisingly, cART alone did not impact mitochondrial copy number. However, both the PFOS and PFOS + cART groups showed reduced mitochondrial copy numbers and increased deletions, indicating alterations to the mitochondrial genome.

### Conclusion
These findings suggest that cART exposure, particularly in combination with PFOS, alters cardiac structure and function, with early systolic impairment followed by compensatory adaptations. Despite partial recovery over time, persistent hypertrophy suggests cardiac remodeling. Ongoing analysis of pregnant animals will determine whether the added physiologic strain of pregnancy sensitizes the cardiovascular system of the animals. This work highlights a need to evaluate the cardiovascular safety of cART in the context of common environmental exposures, particularly in women of reproductive age.		2025 AMA Research Challenge – Member Premier AccessMelissa Marchesefree
88
128757Development, Implementation and Evaluation of a Novel Neuropalliative Care Education Curriculum### Background
Neurologic illnesses often involve high symptom burden, complex prognostication, and emotionally charged decision-making, making palliative care (PC) skills essential for neurologists. Despite national recommendations, formal training in neuropalliative care remains limited in neurology residencies and fellowships. A prior needs assessment revealed that only 4.5% of U.S. neurocritical care (NCC) fellowship programs had faculty with formal PC training, and many fellows felt underprepared in communication, prognostication, and discussions surrounding life-sustaining treatments. The objective of this study was to develop a novel, scalable palliative care curriculum for neurocritical care fellows that is being expanded across neurology residency and subspecialty fellowship training programs. We aimed to assess feasibility, learner engagement, and changes in self-reported preparedness in serious illness communication using pre/post surveys, performance-based simulation assessments, and qualitative interviews.

### Methods
We piloted a novel, interdisciplinary neuropalliative care curriculum for NCC fellows over three academic years (2022–2025), combining asynchronous didactics from the Education in Palliative and End-of-Life Care for Neurology (EPEC-N) curriculum with in-person, simulation-based training using VitalTalk techniques. Simulation scenarios were collaboratively developed and vetted by neurocritical care and palliative care faculty to reflect realistic, frequently encountered clinical experiences. Building on the success of this pilot, we are expanding the curriculum to neurology residency and fellowship training across subspecialties. Using Kern’s Six-Step Curriculum Development framework and situated learning theory, we created subspecialty-specific simulation scenarios (e.g., stroke, epilepsy, movement disorders), integrated into an academic neurology curriculum. A prospective, mixed-methods study was used to evaluate the program’s feasibility, effectiveness, and impact.

### Results
Preliminary results from the pilot program demonstrate high learner engagement and improved self-reported preparedness in serious illness communication. Fellows reported increased confidence in delivering bad news, discussing prognosis and code status, and managing emotional conversations with families. Program directors identified key competencies, including prognostication (95%), delivering bad news (85%), and discussions on withdrawal of care (80%), as educational priorities. Evaluation tools included pre/post surveys, simulation performance assessments, and qualitative interviews. Data collection and analysis are ongoing.

### Conclusion
This curriculum represents a novel approach to neuropalliative care education by integrating simulation-based training into neurology and subspecialty education for interdisciplinary teams. Grounded in evidence-based communication frameworks and tailored to subspecialty practice, the program addresses a critical training gap in neurology education. Preliminary outcomes support its feasibility and impact on learner confidence and skill development. Future directions include national expansion and integration across neurology subspecialties to establish a sustainable, standardized model for neuropalliative care education.
2025 AMA Research Challenge – Member Premier AccessKarishma Poplifree
89
128756Assessment of Proximal Postural Control Biomarkers During Spontaneous Play in Infants Using Surface EMG### Background
As the first applied motor task to develop, assessment of head and trunk control in infants may provide valuable insight into potential developmental delays. While electromyography (EMG) is the gold standard for assessing neuromuscular function, its use in infants is limited due to their size, difficulty establishing normative values, and lack of standard protocols. Existing research in infants largely focuses on discrete tasks in advanced postures, with few studying spontaneous movements or proximal muscle activity. This study aimed to assess the feasibility of using surface EMG during prone play and to identify potential movement biomarkers associated with delayed postural control in young infants.

### Methods
This pilot study involved five infants aged 3-6 months (mean age = 4.61 months). Subjects played in the prone position while wearing wireless surface EMG sensors (Delsys) based on SENIAM guidelines. Bilateral recordings were taken from muscles involved in proximal control: upper trapezius, lower trapezius, and triceps. Functional motor performance was assessed using the Alberta Infant Motor Scale (AIMS) percentiles. Raw EMG signals were processed in MATLAB®. The metrics analyzed were mean RMS amplitudes (mV), bilateral symmetry index ratios (> or < 1 indicating asymmetry), and onset of activation (s). Association between EMG metrics and AIMS percentile was analyzed using Spearman correlations.

### Results
EMG data was successfully collected for all five infants. There was no significant correlation between AIMS percentile and EMG metrics collected, but interesting trends were observed. Infants in lower percentiles showed greater upper (r= -0.7) and lower trapezius (r= -0.7) activation and overall unbalanced recruitment between muscle groups compared to infants in higher percentiles. Average muscle onset times across all subjects were: lower trapezius (1.3s), upper trapezius (1.6s), and triceps (3.9s). Symmetry indices varied across participants, with the lower trapezius showing the most consistent symmetry (0.88), compared to upper trapezius (1.22) and triceps (1.56). Lower AIMS percentiles were generally associated with greater asymmetry.

### Conclusion
Surface EMG is feasible during prone play. We identified movement biomarkers that may offer early insight into atypical postural control strategies in at-risk infants. Infants with lower motor performance demonstrated more proximal motor unit recruitment, indicating less stability. These infants may be more likely to have asymmetrical recruitment than higher performing infants. Larger sampled studies are needed to further develop the baseline of expected muscle activation patterns. Increased knowledge in this understudied area could aid in identifying potential developmental delays at an earlier age.		2025 AMA Research Challenge – Member Premier AccessKatherine Triplettfree
90
128755Policy Addendums to Resolve South Korea’s Emergency Room Admission CrisisSouth Korea’s healthcare system is internationally recognized for its accessibility and efficiency; however, beneath this success lies an emergency care crisis marked by widespread hospital refusals of critical patients—a phenomenon known as “ER ping-pong.” Patients, particularly in rural regions, are often transferred between multiple facilities before receiving care, with fatal consequences. Recognizing the absence of national data and clear policy guidance, this study sought to quantify provincial emergency admission refusal rates and evaluate structural reforms to improve equitable access to emergency care.

A PRISMA-guided systematic review was conducted using Covidence to assess 46 peer-reviewed studies on emergency access policies across OECD nations. This was combined with quantitative analysis of South Korea’s 2023–2024 national emergency transfer data, made available through the Ministry of Health and Welfare (MOHW) and the Korean Society of Emergency Medicine (KSEM). Provincial refusal rates were estimated and compared with policy frameworks from the United States (Emergency Medical Treatment and Labor Act, EMTALA) and Germany (Sozialgesetzbuch V and Leitstelle dispatch systems).

Findings reveal that refusal rates range from 30% in Seoul to nearly 98% in North Gyeongsang, with rural provinces experiencing rates more than double those of urban centers. Most refusals cited limited ICU capacity or specialist coverage as primary causes. Comparative analysis demonstrated that regions with integrated referral databases, legal stabilization mandates, and regionalized workforce incentives experience significantly lower refusal rates.

This research proposes a multi-tiered reform model for South Korea, including a Standardized Emergency Referral and Triage Protocol (SERTP), expansion of tele-emergency consultation hubs, and a national integrated emergency operations platform, to reduce refusals by an estimated 60% within a decade. These policy addendums provide an actionable blueprint for improving emergency equity and sustainability within South Korea’s universal health system.		2025 AMA Research Challenge – Member Premier AccessDestiny Kanningfree
91
128754Differentiating the Clinical & Laboratory Presentations of Lyme Arthritis, Septic Arthritis, & Toxic Synovitis### Background
Lyme arthritis (LA) is a late-stage manifestation of Lyme disease characterized by joint swelling and pain. LA’s clinical presentation mimics other pathologies such as septic arthritis (SA) and toxic synovitis (TS). Despite their similar presentations, the necessary treatments vary. LA is treated medically, septic arthritis surgically, and TS resolves with conservative therapy. The similar clinical presentations of these pathologies can lead to misdiagnosis, subjecting patients to unnecessary interventions, and prolonged hospitalizations.
The state of West Virginia has suffered an exponential increase in Lyme disease (LD) over the last decade. The CDC defines high-incidence states as those having an annual incidence of ≥10 confirmed LD cases per 100,000 for three reporting years. Over the span of this study, West Virginia averaged 40.3 cases per 100,000.
The purpose of this study was to conduct a retrospective review of pediatric patients that presented to our health system with concern for joint infection/inflammation to identify physical examination and laboratory findings that may differentiate between LA, SA, and TS.

### Methods
Pediatric patients were selected for the study based on ICD-9/10 codes suggestive of an infectious/inflammatory/Lyme-related process. The initial group was filtered down to patients who had undergone testing for an infectious process. This group was then manually reviewed, and each patient was assigned to a diagnostic category. Relevant patient information was extracted from their charts. Continuous data were summarized with simple numerical summaries and categorical data as counts and proportions.

### Results
506 patients were analyzed in this study. Subgroup analysis focused on LA, SA, and TS. Patients with LA had the highest proportion of joint swelling, followed by SA. Pain with short arc motion was most prevalent in the SA group and they were least likely to bear weight. LA favored the knee, TS the hip and SA the knee and hip however other joints were also involved. SA presented with the highest CRP, ESR, WBC, %neutrophils, and cell count values. TS had the lowest elevation in CRP, ESR, %neutrophils, and cell count. TS had the highest %lymphocyte and LA the highest platelet count.

### Conclusion
As Lyme disease continues to spread, clinicians will increasingly encounter patients with LA. Clinicians must differentiate between LA and similar pathologies to ensure appropriate treatment and avoid unnecessary interventions. Key differentiating variables, such as anatomic location, CRP, and ESR, can aid in diagnosis, though there are limitations. Clinicians can use this
information to guide treatment decisions for these patients.
2025 AMA Research Challenge – Member Premier AccessJonathan Yaarifree
92
128752Characterizing the impact of PSMA PET/CT in the detection and treatment of prostate cancer in the VCU Health System### Background
Prostate-specific membrane antigen positron emission tomography/computerized tomography (PSMA PET/CT) has demonstrated high sensitivity for prostate cancer (PCa) metastases, changing the manner in which PCa is managed. This study aims to report PSMA PET/CT data alongside multiparametric mRI workup and pathological reports from ultrasound/MRI biopsy (fBx) as a gold standard in patients cared for in the VCU Health System.

### Methods
A retrospective analysis of 765 records of patients who received fBx from January 2022 through December 2024 was conducted. Inclusion criteria involved patients who underwent PSMA PET/CT in initial planning (newly diagnosed cases). Exclusion criteria involved patients who received prior prostate treatment, including surgery or radiation. Data recorded from chart review included age, PSA, Prostate Imaging Reporting & Data System (PI-RADS) scores, fBx Gleason Score (GS), and prostate volume. Additionally, the presence of metastasis and how this altered their treatment plan as evidenced by PSMA PET/CT was also distinguished. There were 177 patients with complete data available for analysis. Statistical analysis was performed with GraphPad.

### Results
PSMA PET/CT found 36 patients to be positive for metastatic disease, while 141 were negative. Of those that presented with evidence of metastasis, 27 patients changed their initial treatment plan based on their PSMA PET/CT (75%), compared to 69 patients that presented without metastatic lesions (48.9%). Meanwhile, 9 patients with metastatic involvement as evidenced by PSMA PET/CT did not change their initial treatment regimen (25%), while 72 patients without evidence of metastasis did not (51.1%) The average age of patients included in this study was 68.9 (± 8.61) years. Other descriptive statistics showed an average fBx GS of 7.84 (± 0.97), average PI-RADS of 4.75 (± 0.56), and average prostate volume of 44.65 mL (± 23.49). Lastly, average PSA at time of diagnosis for those with metastasis was 97.13 ng/mL (± 224.53) compared to 23.002 ng/mL (± 52.67) for those without metastasis (p < 0.0406).

### Conclusion
PSMA PET/CT is a valuable imaging modality that not only helps to visualize possible metastatic lesions within patients with prostate cancer, but also alters clinical management in a large proportion of patients in our hospital. Further studies and investigation into the broader utilization of PSMA PET/CT nationwide would prove beneficial.		2025 AMA Research Challenge – Member Premier AccessAmmar Jawadfree
93
128751Sleep-Related Breathing Disorders in Surgically Corrected Single Ventricle Congenital Heart Disease Patients### Background
There is limited data on sleep-related breathing disorders (SRBDs) in pediatric patients with single ventricle congenital heart disease (SVCHD) who have undergone surgical correction, but available studies indicate it is
commonly present. The impact of SRBDs in these patients remains poorly understood. This study investigates the prevalence of SRBDs in patients who have undergone the Fontan procedure.

### Methods
We conducted a retrospective chart review of 62 pediatric patients with SVCHD who had undergone the Fontan procedure and had a polysomnographic study (PSG) before or after the Fontan procedure between 2000 and 2024. Data was abstracted from patients’ most recent PSGs and included demographic characteristics,
apnea-hypopnea indices (oAHI and CAI), end-tidal CO2 (ETCO2) and oxygen saturation metrics.

### Results
The cohort’s mean age at the time of their first PSG was 9.1 ± 5.1 years. The cohort’s mean age at the time of the Fontan procedure was 3.6 ± 0.8 years. 66% of the cohort were male (41/62). 37/62 patients (60%) had OSA. 14 subjects exclusively had PSGs performed only before Fontan; 42 subjects exclusively had PSGs only
after Fontan. 10% of patients (6/62) had PSGs both before and after Fontan. The pre-Fontan cohort had 4 mild OSA, 2 moderate OSA and 1 severe OSA. The post-Fontan cohort had 19 mild OSA, 6 moderate OSA and 4 severe OSA. There was improvement in AHI in 2/6 patients who had pre and post Fontan PSG. 6/62 patients
had CSA (CAI ≥ 5/hr.) and none met criteria for hypoventilation (ETCO2 >50 mmHg for >25% of total sleep time).

The average oAHI pre-Fontan was 3.27 ± 5.1 and the average oAHI post-Fontan was 4.7 ± 8.4. The average oxygen saturation (SpO2) pre-Fontan was 83% and the average SpO2 post-Fontan was 92%. The average SpO2 nadir pre-Fontan was 73% and the average SpO2 nadir post-Fontan was 84%. SpO2 nadir of < 90% was seen in 19 patients prior to the Fontan procedure and in 38 patients after the Fontan procedure.

### Conclusion
There was a high prevalence of SRBDs in pediatric patients with SVCHD patients’ post-surgical correction. Most patients demonstrated mild OSA. Our data suggest that early screening for SRBDs in this population is essential.		2025 AMA Research Challenge – Member Premier AccessJacob Pesachovfree
94
128750Procoagulant platelets and neutrophil extracellular traps restrict fetal growth in a murine model of hypertensive diseases of pregnancy### Background
Hypertensive diseases of pregnancy (HDP) like preeclampsia lead to fetal growth restriction (FGR), where elevated thromboxane A2 (TXA2) levels trigger maternal hypertension, recurrent placental ischemia/reperfusion injury, and fibrosis leading to placental insufficiency and FGR. We recently demonstrated a mechanistic link between procoagulant platelets and neutrophils in the innate immune response to ischemia/reperfusion injury following ischemic stroke - neutrophil extracellular trap (NET) formation. To respond to infection or tissue damage, a subset of neutrophils form NETs. These NETs release lattices of decondensed chromatin decorated with antimicrobial factors into the extracellular space. In ischemic stroke, procoagulant platelets induce NET formation leading to pathogenic inflammation which extends infarct size and compromises neurologic recovery after stroke. In the preclinical stroke model, we showed that Cyclophilin-D (CypD) induces procoagulant platelet formation which induced NETs leading to pathogenic inflammation. However, whether procoagulant platelets and NETs mediate FGR in HDP-related placental insufficiency remains unknown. We hypothesized that platelet-specific Cyclophilin D knockout (CypDplt-/-) mice exhibit decreased FGR and placental NET formation compared to wild-type (WT) mice in a preclinical model of HDP.

### Methods
Pregnant CypDplt-/- and WT mice received a TXA2 analog (U46619; 4 μg/μL at 0.5 μL/hr continuous infusion) or vehicle control (0.5 μL/hr) via micro-osmotic pumps implanted on embryonic day (E) 12.5. At E19, we measured fetal/placental weights and collected pup plasma. We assessed placental NETs and neutrophil levels via Western blotting of placental homogenates for citrullinated histone H3 (citH3) and myeloperoxidase (MPO), respectively. We assessed pup plasma MPO-DNA complex levels as a surrogate for plasma NET levels using an in-house MPO-DNA ELISA.

### Results
Compared to TXA2-treated WT pups, TXA2-treated CypDplt-/- pups showed decreased body weight at E19 (p<0.05), decreased percentage FGR (18% versus 71%), reduced placental MPO (p<0.05), with a trend towards decreased placental citH3 (p=0.051) via western blotting, suggesting decreased placental neutrophils and NETs, respectively, and decreased plasma MPO-DNA complex levels, a surrogate for NET formation (p<0.05). Finally, using a novel NET-inhibitory peptide which we discovered, we demonstrated that NET inhibition in this pre-clinical model decreased FGR in response to TXA2 treatment, a finding correlated with decreased placental (p<0.05) and plasma NET levels (p<0.05), respectively.

### Conclusion
CypDplt-/- TXA2-treated pups show decreased FGR and decreased placental NETs in a preclinical HDP model suggesting that procoagulant platelets trigger platelet activation and NET formation in HDP. Further, NET inhibition may represent a novel therapeutic target to decrease fetal growth restriction in human HDP.



2025 AMA Research Challenge – Member Premier AccessJacob Taylorfree
95
128749Factors Affecting 90-Day Readmission After Acute Appendicitis Within a Large Healthcare System### Background
Acute appendicitis is the most common cause of emergency general surgery in the United States and affects 250,000 adults every year. Readmission after appendectomy is common and represents an opportunity to improve care. We aimed to identify factors associated with 90-day readmission in patients with acute appendicitis to inform patient risk stratification.

### Methods
We conducted a retrospective chart review of the emergency general surgery registry for patients with acute appendicitis between March 1, 2023, and August 31, 2024, at three hospitals within our healthcare system: one academic (H1) and two community (H2 and H3). Our primary outcome was 90-day hospital readmission. We used descriptive statistics and multivariable logistic regression to assess factors related to readmission, performed a sensitivity analysis excluding early readmissions (<7 days), and stratified models by hospital to evaluate site-level variation.

### Results
522 patients with acute appendicitis were included; the overall 90-day readmission rate was 13.0%. Readmitted patients were on average older (56 vs 45 years, p<0.001), had higher Charlson Comorbidity Index scores (CCI) (2.1 vs. 1.2, p<0.001), and more often presented with perforated appendicitis (51% vs. 24%, p<0.001). In multivariable regression, operative management was independently associated with reduced odds of readmission (aOR 0.26, 95% CI 0.12–0.53). CCI (aOR 1.25, 95% CI 0.97–1.59) and diabetes (aOR 1.89, 95% CI 0.91–3.93) increased risk but were not statistically significant. Age, perforation status, and discharge disposition were not significant predictors of 90-day readmission. In sensitivity analysis (n=461), CCI (aOR 1.32, 95% CI 1.01–1.72) and ASA score (aOR 2.23, 95% CI 1.03–4.87) were significant predictors, while operative management remained protective (aOR 0.12, 95% CI 0.06–0.25). Older age was associated reduced readmission (aOR 0.95, 95% CI 0.91–0.99), and diabetes was not significant. Hospital-stratified analysis showed operative management was consistently protective across sites (H1 aOR 0.23, 95% CI 0.09–0.64; H2 aOR 0.04, 95% CI 0.009–0.21; H3 aOR 0.05, 95% CI 0.002–1.21).

### Conclusion
Operative management of acute appendicitis was independently protective against 90-day readmission. Higher CCI and ASA scores increased readmission risk, particularly in patients readmitted beyond 7 days post discharge. These findings highlight the importance of individualized risk stratification, particularly in older comorbid adults, and may inform discharge planning and follow-up strategies.		2025 AMA Research Challenge – Member Premier AccessIain Kehoefree
96
128748Mild flexion of the femoral component sagittal alignment in TKA may improve long-term clinical outcomes and component survival: A systematic review### Background
Total Knee Arthroplasty (TKA) is a beneficial procedure in orthopedics, but malpositioning of the femoral TKA component can cause dissatisfaction or failure. Most existing research focuses on coronal and rotational positioning, making positioning in the sagittal plane relatively unexplored. The objective of this systematic review is to determine whether variations in femoral component sagittal alignment (flexion or extension positioning) influence patient-reported outcomes, range of motion, anterior knee pain, component loosening, and prosthesis survival.

### Methods
This systematic review was performed according to the guidelines of the Preferred
Reporting Items for Systematic Review and Meta-Analyses (PRISMA) protocols (PROSPERO ID number: CRD420251044455). A comprehensive search of PubMed, Cochrane Library databases, and SPORTDiscus was performed between January 2015 and April 2025. Keywords included Total Knee Arthroplasty, Revision Total Knee Arthroplasty, Femoral Component, Sagittal Alignment, Femoral Flexion Angle, Extension Alignment, Posterior Condylar Offset, and Patient-Reported Outcome Measures. After screening 2137 initial results, 10 studies were deemed to meet inclusion and exclusion criteria, with 9 being prospective/retrospective cohort studies and 1 Randomized Control Trial (RCT). Study characteristics and outcomes were extracted, and quality assessment performed for using Cochrane Risk of Bias tool (RCTs) and Newcastle-Ottawa Scale (cohort studies).

### Results
A total of 5,205 TKAs were performed among the 10 studies, each with different sagittal alignment grouping thresholds and functional outcome scales. 5/10 studies identified the optimal sagittal alignment to be approximately 1-4° flexion for OKS, WOMAC, KSFS, or KSS functional scores, while 1 study found slightly improved AKSS with 1–5° extension. 2 studies did not identify optimum femoral angles for functional score outcomes. For ROM, 5/10 studies identified the optimum sagittal angle for ROM to be 1-7° flexion. Only 2 studies explored failure rate, with 1 study reporting the best failure rate (0%) in neutral alignment 0-3° and the other stating no cases of loosening at 1 year for any group. Anterior knee pain was only reported by 1 study with best outcomes at 1.4° flexion.

### Conclusion
Of the 10 studies, mild sagittal flexion (1-4°) appears to appropriately balance functional outcomes, ROM, survival, and pain. However, the heterogeneity of the included studies in terms of grouping measurement as well as functional outcomes precluded our ability to conduct a meta-analysis and generate exact quantitative outcomes as to the optimal sagittal alignment of the femoral component. More RCTs are needed to explore sagittal component positioning with consistent femoral flexion angle grouping criteria and clinical outcome measurements.		2025 AMA Research Challenge – Member Premier AccessAndrew Huffmanfree
97
128747Socioeconomic Impact on Patient-Report Outcome Measures Participation in Arthroplasty Patients### Background
Patient-reported outcomes (PROMs) are standardized questionnaires patients complete to gather information concerning health outcomes, symptoms, and functional status. Prior studies have evaluated completion and/or participation rate across orthopaedic subspecialties. However, few have evaluated both completion and participation outcomes in joint replacement patients specifically. This study aims to determine the association between sociodemographic factors and PROMs completion and participation rates.

### Methods
A retrospective chart review was conducted of all patients who underwent primary total knee and total hip replacements by at a Midwestern, tertiary health system in 2023. PROMs used in this analysis included PROMIS PF, PROMIS PI, and either HOOS or KOOS JR. Demographic variables and PROMs completion were recorded, and the Area Deprivation Index (ADI), a measure of socioeconomic disadvantage, was determined using patients’ address. Patients were considered to participate if they filled out at least one questionnaire before and after surgery and were considered to complete PROMs if they filled out all pre- and postoperative questionnaires. Logistical regressions were performed to identify factors associated with PROMs participation and completion.

### Results
646 patients were included in our analysis, of which 31 (4.8%) completed all PROMs and 529 (82%) participated in PROMs. In univariate analysis, Black patients had significantly lower odds of PROMs participation (56% lower, p<0.001), as were older patients (OR 0.97; p = 0.004) and patients living in zip codes with high national ADI (OR 0.98; p < 0.001), while married patients were more likely to participate (OR 1.89, p=0.015). In the multivariable model, older age (OR 0.97, p=0.004) and Black race (OR 0.38, p=0.013) were associated with lower participation. Univariate analysis of PROMs completion suggested 69% lower likelihood among Black patients (p=0.056) and with additional preoperative visits (69% lower per day, p<0.001).

### Conclusion
Sufficient sociodemographic representation in PROMs is necessary to determine equitable outcomes amongst all members of society. Our data is consistent with established literature with regards to how ADI and age influence PROM completion. Further exploration regarding the factors behind low PROM completion is needed in order to ensure adequate representation in orthopaedic outcomes in clinical and research environments.		2025 AMA Research Challenge – Member Premier AccessMichelle Yifree
98
128746Test of Executive Function in an Emergency: A Real-World 911 Tool That Differentiates Mild Alzheimer’s Disease from Other Dementias### Background
Executive dysfunction in Alzheimer’s disease (AD)—often linked to early frontal lobe involvement—can impair real-world behavior even during the mildest stages, well before memory loss becomes disabling. Yet many clinicians delay supervision planning until patients reach moderate dementia, overlooking critical safety risks in earlier phases. High misdiagnosis rates in early dementia further delay appropriate intervention, especially when subtle but meaningful executive deficits go undetected. Although tools like the Mini-Mental State Examination and Trails B assess cognition, none directly evaluate whether a patient can live safely without supervision. The Test of Executive Functioning in an Emergency (TEFE), developed at UNT Health, addresses this gap. TEFE is a brief, performance-based tool that assesses whether patients can complete real-world safety tasks—such as dialing 911, stating their address, and contacting a family member. This study examined whether TEFE performance—particularly on the 911 task—can differentiate mild AD from vascular dementia (VD), mixed dementia, and mild cognitive impairment (MCI), while also identifying clinically meaningful patterns of executive dysfunction across diagnoses.

### Methods
We retrospectively reviewed records of 908 patients from a university-affiliated memory clinic. Diagnoses included mild AD, mild VD, mild mixed dementia, mild cognitive impairment (MCI), or normal cognition. TEFE includes six items total: three knowledge-based questions (e.g., “What number do you call in an emergency?”) and three behavioral tasks (e.g., dialing 911). Total TEFE scores were compared across diagnostic groups using Kruskal-Wallis and chi-square tests. We further analyzed failure rates on the 911 knowledge and behavioral items individually and used contingency tables to examine disconnects between knowing and performing the task (e.g., correctly identifying 911 but failing to dial it).

### Results
TEFE performance significantly differed by diagnosis (H = 215.73, p < 1×10−44). AD patients scored significantly lower than MCI or cognitively normal groups. Nearly one-quarter of AD patients failed the 911 knowledge (24.1%) and behavioral (27.0%) tasks. In contrast, 911 failure was rare in non-AD groups—under 6% in MCI and 0% in normal controls. Chi-square tests confirmed AD patients were significantly more likely to fail 911 tasks than VD or mixed dementia (p < 0.001). Only AD patients exhibited a disconnect of identifying 911 but failing to dial it.

### Conclusion
TEFE effectively detects executive dysfunction and emergency readiness deficits in mild AD, identifying safety risks not captured by traditional cognitive screens. By revealing real-world impairments before more obvious cognitive decline, TEFE offers actionable insight that help guide earlier interventions in dementia care.		2025 AMA Research Challenge – Member Premier AccessHoward Kimfree
99
128745Dorsal Nerve of the Clitoris and Dorsal Nerve of the Penis Size: Implications for Reconstruction### Background
The dorsal nerve of the clitoris (DNC) and dorsal nerve of the penis (DNP) are vital to sexual sensation, erectile function, and reproduction; however, these nerves are frequently subject to injury from tumor resections, aesthetic procedures, and female genital mutilation (FGM). Concerning FGM alone, the World Health Organization notes that approximately 6.5 million females in Europe and North America, and over 230 million females worldwide, are living with FGM. Nerve coaptation or neurotization may be warranted to restore function after dorsal nerve injury. Accordingly, close nerve size-matching is necessary to minimize the risk of fascicle loss and neuroma formation. However, little is known about the diameter of the dorsal nerves. Therefore, this study aims to describe the diameter of the DNC and DNP through gross anatomical assessment.

### Methods
A total of 101 donor bodies (51 female, 50 male) were dissected to visualize the paired dorsal nerves as they traveled inferior to the pubic bone. Dorsal nerves were photographed, and images were screened for quality. A total of 199 dorsal nerves met criteria for measurement. Nerve diameter was measured via ImageJ software. Sexual dimorphism and asymmetry were assessed by unpaired and paired t-tests, respectively. Correlation between nerve diameter and age was evaluated via Pearson’s test.

### Results
Nerve diameters were normally distributed in both females and males. The DNC diameter averaged 2.9±0.6mm (mean ± SD) (min=1.5mm; max=4.9mm; range=3.4mm) and the DNP averaged 3.6±0.6mm (min=1.6mm; max=5.1mm; range=3.5mm). Significant differences were found between sexes (t=7.647; p<.00001); however, within-sex, no significant side-to-side difference was identified (females: t=0.6278; p=0.5332) (males: t=1.206; p=0.2337). Age was not correlated with nerve size (r=0.17; p=0.248). Although generally symmetrical, bilateral asymmetry of up to 1.7mm in females and 1.5mm in males was noted.

### Conclusion
This report details DNC and DNP size from the largest sample population to date. Surgeons should anticipate dorsal nerve size differences between sexes, wide size variation within-sex, and possible bilateral asymmetry. These results, assessed alongside other reports, suggest that commonly grafted nerves are generally more size-compatible with the DNC as compared to the relatively larger DNP. Regarding unilateral graft procedures, such as an ilioinguinal nerve coaptation, asymmetry may influence the side upon which the procedure is performed. The results of this study will improve genital reconstruction procedures and guide optimal nerve graft selection for dorsal nerve repair.		2025 AMA Research Challenge – Member Premier AccessVeronica Gibbonsfree
100
128744Non-Thermal Pulsed Electric Field Ablation of Blood Clots Reduces Fibrosis### Background
Deep vein thrombosis (DVT) is a major healthcare burden with significant morbidity and mortality. Anticoagulation therapy is the current treatment of choice. However, this approach can sometimes lead to only partial resolution, allowing for thrombus organization. This can result in the development of post-thrombotic syndrome (PTS), a sequela of chronic DVT that includes symptoms such as limb pain, cramping, edema, stasis dermatitis, or even venous ulceration. There is currently no effective treatment for PTS, and as a consequence, it can become a lifelong problem that significantly impairs a patient’s quality of life. Here, we examine whether the local delivery of non-thermal irreversible electroporation (IRE) can impede thrombus organization and fibrosis by achieving cellular ablation and allowing time for natural thrombus resolution to occur.

### Methods
In vitro and ex vivo methods were utilized to investigate the thermal and ablative effects of IRE. A microfluidics device was seeded with neutrophils and endothelial cells to determine voltage thresholds for selective, safe ablation. Rheological analysis was conducted to evaluate the impact of IRE on blood coagulation. Porcine blood clots, generated in a syringe model, were used to assess red blood cell and fibrin architecture following IRE treatment. These models guided the optimization of IRE parameters for subsequent in vivo studies. Calibrated pulsed electric fields were then delivered locally using a tweezer-style electrode system to target acute DVT lesions induced in the femoral veins of rat and swine models. Animals were monitored for three or seven days post-treatment. Histological, immunohistochemical, and molecular analyses were performed to assess thrombus organization, vessel integrity, luminal cellularity, cytokine expression, and inflammatory or fibrotic markers following IRE-treatment.

### Results
IRE-treated blood demonstrated preserved clotting kinetics and minimal hemolysis, supporting the safety and efficacy of blood cell ablation using IRE. Histologic analysis of IRE-treated animal femoral vein segments revealed significantly reduced cellular proliferation, collagen deposition, and angiogenesis. IRE-treated tissues also exhibited lower expression of pro-fibrotic, pro-thrombogenic, and pro-coagulant markers compared to controls. Across both models, vessel wall thickness and luminal architecture remained intact, confirming the non-destructive nature of IRE, with no evidence of structural wall damage or aneurysm formation.

### Conclusion
These findings suggest that IRE may serve as a promising adjunctive therapy to anticoagulation by preventing the organization of acute thrombi into chronic DVT. This could ultimately reduce the incidence of post-thrombotic syndrome and improve long-term patient outcomes.		2025 AMA Research Challenge – Member Premier AccessGia Sainifree