1 of 54

MAGIC

Mentee Presentation

Fieldston�May 16th, 2023

1

2 of 54

Agenda

  • Zoe Gorman
    • ZWELL
  • Sophia Thompson
    • Digital Image Processing
  • Mali Gilbert
    • Anaplastic Large Cell Lymphoma

​

2

3 of 54

Zoe Gorman

9th Grade, Fieldston School, Bronx

Katie’s Intro

3

4 of 54

My Mentor

  • Katie Gallagher
    • Katie lives in California
    • Works as a botanist
    • Has a daughter
    • Enjoys science specifically ecology
    • Katie is pregnant

4

One of Katie’s Recent Projects:

5 of 54

My Project - Zwell

  • event planning, organizing, advice giving, restful, educational, and nutritional App.
  • mainly directed for students, but can be used and customized to everyone's needs
  • I have always been interested in creating an app as it is an outlet for creation and expression for things I am passionate about
  • create websites, learn coding, import information and much more

5

6 of 54

My MAGICal Experience

  • Magic has helped me learn and grow as an engineer by receiving feedback and help from my mentor and peers
    • seeing my app come to life, letting others try it out, and experimenting with appy pie
    • persevere through challenges and difficult times, especially when you think it may not always work out
    • Challenges:creating format to make the app appealing and getting certain features to work always, publishing

​

6

7 of 54

My Demo

Below are screenshots of the app at its beginning state from February 1st:

​

7

8 of 54

My Demo

Below is a video of the app at its beginning state from February 7th, before the information had been uploaded in by me

​

8

9 of 54

My demo

Below are screenshots of the app at its current state

9

Four Features:

10 of 54

My demo

Below are screenshots of the app at its current state

10

Events:

11 of 54

My demo

Below are screenshots of the app at its current state

11

Plan:

12 of 54

My demo

Below are screenshots of the app at its current state

12

Rest:

13 of 54

My demo

Below are screenshots of the app at its current state

13

Learn:

14 of 54

Questions?

14

15 of 54

Thank you!

15

16 of 54

Sophia Thompson

9th Grade, Ethical Culture Fieldston School, New York City

Sneha’s Intro

16

17 of 54

My Mentor

  • Sneha Naik
    • Works and studying for a PhD at Columbia University in Biomedical Engineering
    • From London, UK, and moved to New York City one year ago
    • Loves Yoga, enjoys running, likes cooking,

17

18 of 54

My Project

  • Describe your project
    • Medical Image Processing
    • A program, that can analyze different disease pathologies, on medical images in python
    • I wanted to learn how to code in python, and was interested in biomedical engineering, so this project intersects my two interests
    • Topics and technologies learned
      • Learning python from scratch using Codeacademy
      • Numpy, Matplotlib, OpenCV, Kaggle Notebooks
      • Digital Image Processing and how to understand a data set
    • Learned quite a lot of python, and how to read a research paper

​

​

​

18

19 of 54

My Demo: A summary of my project

  • Motivation:
    • This project stemmed from my interest in medicine - which comes from seeing my mom work as a radiologist
    • I wanted to help create a piece of code that can be used to help doctors simplify their analysis process
    • The pieces of code that I created are reusable and alterable for different images, and different tasks
  • Table of Contents:
    • Learning Python
    • Tasks/Data Set
    • Processing Images
    • Conclusions

19

20 of 54

Learning Python from Scratch [Codeacademy]

20

Control Flow:

  • If, else, elif statements
  • Boolean statements
  • In simple english, true false statements

Lists:

  • Indexing/Slicing lists
  • Appending Lists
  • Modifying Lists
  • In simple english, how to create, modify, and access lists (a collection of data)

Loops:

  • For/While/Nested Loops
  • Range
  • In simple english, repetitive actions

Functions:

  • Parameters
  • Calling functions
  • Returning Values
  • In simple english, a reusable set of actions that only runs when you call it

We streamlined through Python to get the most that we could, and the information that would help us the most when creating a piece of image processing code

21 of 54

Image Processing Libraries

21

22 of 54

Our Goal: To count the number of cells in an image

22

Image of Cells Under Microscope

Manually Segmented Cells (white=cell, black=background)

There are 21 cells

23 of 54

Analyzing Images

  • What does this image look like as an array?

23

Shape = (256, 256, 3)

​

​

Value = (0, 255)

​

​

Channel:

0 = red

1 = green

2 = blue

24 of 54

Analyzing Images Part 2

We analyzed the intensity of the red, green and blue channels separately, using histograms to visualize this

​

​

I wrote my own function for this!

24

25 of 54

My Histograms

25

26 of 54

Pre-Processing

  • The histograms for this image are quite spiky - I wanted them to be smooth in order for it to be easier to analyze the important features of the image
  • I did this by applying a blurring filter to the image
  • In plain English - this is the same thing as smudging a charcoal painting with your finger to blend in sharp edges, and smooth the shape
  • The amount of smudging/smoothing depends on the size of your finger/blurring filter

​

26

27 of 54

Post Blurring

27

No Blurring

Blurring Patches of size 5x5

Blurring Patches of size 50x50

28 of 54

Post Blurring Histograms

28

You can see that the middle filter (5x5) has smoothed the histogram without changing the peaks significantly

29 of 54

Thresholding to Segment Cells

29

30 of 54

Code for Plotting

30

A lot of images have been shown side by side

​

This piece of code to the right was what I used to plot images in matplotlib side by side, using subplots

31 of 54

Counting Cells

31

  • In plain English, we are counting the number of white blobs that there are in the mask image
    • (white blobs = cells)
  • I used OpenCV to calculate and draw the contours

32 of 54

Conclusions

  • The number of contours found by my processing method = 38
  • The true number (found in the ground truth image) = 21
  • Reasons for Overestimating Number of Cells:
    • Algorithms are not always correct and or precise in their counting and evaluation of images
    • Here what I found was that my algorithm picked up a lot of dust and thought it was cells
  • Improvements/Future Work:
    • If I were to blur the image a bit more, or have a size limit on the contours the algorithm counts as cells, this could improve the accuracy of the counting method

32

33 of 54

My MAGICal Experience

  • Lessons Learned
    • Debugging is such a huge part of coding and computer science as a whole
    • It’s important to stay patient, and try and work out your bugs - google can be really helpful
  • Favorite Parts:
    • I loved learning something that was so new and different than anything I had ever learnt before, and stretching myself to try out a new section of science
    • Working with Sneha was so amazing, and I was so lucky to work with her, and have a mentor who was able to teach me so much, and learn along with me

​

33

34 of 54

My Kaggle Notebook: A Short Video Demo

34

35 of 54

My Kaggle Notebook: A Video Demo

35

36 of 54

Thank you!

36

37 of 54

Questions?

37

38 of 54

Mali Gilbert

11th Grade, Fieldston School, Bronx

Namrata’s Intro

38

39 of 54

Namrata GR Raut

  • A few bullet points about your mentor
    • Postdoctoral fellow in the Cincinnati Children's Hospital
    • Currently lives in Cincinnati (born in Nepal)
    • She loves riddles!!, ACDC, and MCU Movies
    • Fun fact: We always start or end with a riddle

​

​

39

40 of 54

My Project: Research on Anaplastic Large Cell Lymphoma

​

40

41 of 54

Anaplastic Large Cell Lymphoma (ALCL)

​

​

Lymphoma = a cancer that starts in the lymphatic system

​

First described in 1985

​

Rare type of non-hodgkin's lymphoma

​

“Large-cell neoplasm with anaplastic morphology”

​

Type of T-cell lymphoma

​

​

​

​

​

​

​

41

http://imagebank.hematology.org/image/1765/anaplastic-large-cell-lymphoma--4

​

42 of 54

What Does ALCL Affect?

​

​

​

Invasion of lymph node sinuses

​

Tumor cells are identifiable by their large cytoplasms

(Called anaplastic)

​

Shows morphological heterogeneity

(Cell structures vary)

​

Affected cells can form tumors in lymph nodes, skin, bones, and internal organs

​

​

42

https://www.mypathologyreport.ca/pathology-dictionary/anaplastic/

​

43 of 54

ALK Positive ALCL

​

​

2 types of ALCL: ALK positive and negative

​

ALK=anaplastic lymphoma kinase

protein that is involved in cell growth and division

​

​

ALK positive ALCL → abnormal lymphocytes have a genetic mutation that results in production of too much ALK protein

More common in younger patients

Better prognosis than ALK negative ALCL

​

​

​

43

https://www.hippocraticpost.com/cancer/new-therapy-approach-for-alcl/

https://www.haematologica.org/article/view/5297

44 of 54

44

45 of 54

What is an IRB?

​

​

​

IRB = Institutional review board → “a committee that applies research ethics by reviewing the methods proposed for research to ensure that they are ethical”

  1. Formed questions
  2. Got formal permission from an advisor and executive members
  3. Sent out form
  4. Anonymous and voluntary
  5. Translated responses into graphs

​

​

45

46 of 54

Lymphoma Survey Results

​

​

46

47 of 54

47

48 of 54

Therapies and Treatment

48

  • ALK+ → anthracycline-based treatments (antibiotics)

​

  • Radiotherapy and chemotherapy
    • Debated but prescribed

​

  • Patients at high risk of relapse → high-dose chemotherapy + stem cell transplantation

​

  • Clinical studies have reported excellent results using Brentuximab vedotin
    • New drug
    • Good for relapsed ALCL

​

49 of 54

Factors Involved in ALCL

​

​

Male predominance

​

Caused by mutation in genetic translocations involving ALK

​

Typically affects adults 40 to 65 years but people of all ages are diagnosed

​

People with breast implants are more prone to ALCL (BIA ALCL)

​

​

49

50 of 54

Why are chemists excellent at solving problems?

Because they have all the solutions!

50

51 of 54

Acknowledgements

​

​

Thank you to…

​

Namrata for being a fantastic mentor

MAGIC for such an amazing program

Hollis for setting everything up and always checking in

Everyone who took the time to fill out my survey

Mr. Vega to introducing me to MAGIC

​

​

51

52 of 54

Bibliography

​

​

​

Falini, Brunangelo, and Maria P. Martelli. 2009. “Anaplastic large cell lymphoma: changes in the World Health Organization classification and perspectives for targeted therapy.” Haematologica 94 (7): 897-900.

Hapgood, Greg, and Kerry J. Savage. 2015. “The biology and management of systemic anaplastic large cell lymphoma.” Blood 126, no. 1 (July): 17-24.

Jaffe, Elaine S. 2019. “Diagnosis and Classification of Lymphoma: Impact of Technical Advances.” Semin Hematol. 56, no. 1 (January): 30-36. 10.1053/j.seminhematol.2018.05.007.

Jiang, Manli, Andrew L. Feldman, and N. N. Bennani. 2017. “Lymphoma classification update: T-cell lymphomas, Hodgkin lymphomas, and histiocytic/dendritic cell neoplasms.” Expert Rev Hematol. 10, no. 3 (March): 239-249. 10.1080/17474086.2017.1281122.

Tsuyama, Naoko, Kana Sakamoto, Seiji Sakata, Akito Dobashi, and Kengo Takeuchi. 2017. “Anaplastic large cell lymphoma: pathology, genetics, and clinical aspects.” Journal of clinical and experimental hematopathology 57 (3): 120-142.

​

​

52

53 of 54

Thank you!

53

54 of 54

Questions?

54