Vision4Doctor
MedTech DeepTech
System for recognition and processing of surgical operations based on artificial intelligence.
Preparation of robotic automatic surgery.
Problems
2
After the operation, the tired surgeon has to fill out the protocol of the operation with a description of what was done.
According to statistics, about 3,000 items forgotten in patients are found annually in Russia.
In case of an unsuccessful operation, insurance companies and investigators are forced to analyze the past operation without having all the data.
It is impossible to understand the doctor's deviation from the "ideal technological map", no records of errors are kept. Successful or unsuccessful experience of the doctor remains only with him. Important reactions to non-standard situations are lost.
It is impossible to streamly control the quality and professionalism of a doctor.
Solutions
3
The system sees, analyzes and recognizes everything that happens in the operating room. The protocol of the operation is formed during the operation and is ready immediately after its completion.
The system knows the number of tools and material "at the input" and will report if there is a discrepancy "at the output".
Insurance companies and investigators see the full picture of the past operation.
The system sees, recognizes, analyzes and clearly records all intraoperative events, collects the experience of doctors and forms a database for AI training.
The actions of doctors can be analyzed, work with errors, teach young students and interns.
Collection and synchronization of data from the operating room
Overview
camera
Head camera
surgeon 2
Head camera
surgeon 1
Endoscope
Instrumental
table
Monitor
patient
Surgical
microscope
Camera in
operating room lamp
4
Laparoscopy
Ultrasound
Angiograph
Formation of a multimedia library
5
Recording
video
Synchronization
video
Treatment
video
Recognition
video
Markup
video on
stages and events
Definition of intraoperative events
Determination of the surgeon's actions
Definition of pathologies
Recording
video
Synchronization
video
Treatment
video
Recognition
video
Markup
video on
stages and events
Manual
process
Auto
process
Data shaping
for AI
We are building a database for training artificial intelligence.
In the next 10 years, we expect to replace doctors with robots, and the data obtained today will form the basis of first robotic semi-automated, and later robotic automatic surgery.
Outlying territories, Mars, will receive high-quality surgery. The success of the operation will no longer depend on the professionalism of the surgeon.
We have already started preparing.
6
Look beyond the horizon
7
2008�2007
Smartphone
2012�2004
Social media
2019�2009
Taxi
2023�2015
Ai
?�2025
Robotic automated surgery
Predict and change the future
Accept the existing habitual
Subsequence
8
Selling data to surgical robot developers
Operating room equipment
Recording of operations
Synchronization and markup
Storage
The First Evangelists
9
Moscow Regional Research Clinical Institute named after M.F. Vladimirsky (State Budgetary Healthcare Institution of the Moscow Region MONIKI named after M.F. Vladimirsky) + Clinics of the Moscow Region
34 clinics
RoadMap -
for first year
10
| Functional | Benefits for clinics |
Platform for storing and annotating video operations | 1. Secure storage for endoscopic surgery videos (with data security standards) 2. Simple annotation tools for doctors: marking key stages of surgery, adding comments 3. Possibility of sharing for training and analysis of complex cases 4. Formation of a directory of selected entities, such as instruments, organs, colors, textures, etc. | 1. Centralized location for data storage 2. Simplify internal training and analysis |
Error Analysis and Recommendations Tool | 1. Automatic video analysis to identify key moments of the operation (e.g. start, end, critical stages) 2. Simple report with statistics (e.g. duration of operation stages) | 1. Possibility of quick analysis of standard procedures 2. Reduction of the number of errors |
Automatic video tagging of key stages | 1. A simple algorithm that automatically marks the stages of an operation (e.g., cutting, working with instruments, finishing) 2. Generating a report on the stages and the time it took to complete them | 1. Saving time on manual video analysis 2. Basic analysis of the efficiency of operations |
System for improving the quality of medical protocols | 1. Identification of discrepancies in the surgeon's actions with standard protocols 2. Automatic creation of individual recommendations based on the collected data. | 1. Simplifying the transition to standards 2. Improving the quality of treatment |
Estimated implementation time
2-3 months: Video storage with annotations
6-9 months: Automatic analysis of errors or stages (here the MVP is finished and there are sales)
6-12 months: A tool associated with high-quality recommendations.
Media
11
Scientists have taught a robot to stitch patients by showing it videos of surgeries.
2020
Stitching a patient up after surgery is an important but monotonous task for doctors, often requiring them to repeat the same simple movements over and over hundreds of times. But thanks to a joint effort between Intel and the University of California, Berkeley, surgeons of the future could hand off the heavy lifting to robots — like a macro, but for automated stitching.
https://www.engadget.com/intel-uc-berkeley-motion2vec-ai-robot-surgery-203003829.html
Controlled Autonomous Robotic Soft Tissue Surgery
2016
The operating room may one day be run by robots, with surgeons controlling their actions. Shademan et al. developed the “Smart Tissue Autonomous Robot,” or STAR, which includes suture instruments, as well as fluorescence and 3D imaging, force measurement, and submillimeter positioning. With all these components, the authors were able to use STAR for soft tissue surgery, a challenging task for a robot given the deformation and mobility of tissue. Surgeons tested STAR against manual surgery, laparoscopy, and robotic surgery for intestinal anastomosis in pigs, and found that the controlled autonomous surgery offered by the STAR system was superior.
https://www.science.org/doi/10.1126/scitranslmed.aad9398#con1
This AI Learned to Perform Surgeries by Watching Videos, and Now It's Ready to Operate on Humans
2024
A new smart robot developed by researchers at Johns Hopkins University and Stanford University learned by watching videos of surgeries. It can now perform procedures with the skill level of a human doctor.
Team
Vadim Potekhin
CEO, founder
Visionary serial tech entrepreneur with expertise in digital transformation across Proptech, Fintech and Medtech.
9241115@gmail.com
+79859241115
Alexander Terentyev
CTO, co-founder
Experienced CTO and development team lead with more than 10 years of experience in complex IT ML
projects.
Leonid Belskikh
Curator, co-founder
Head of the Surgical Center for the Coordination of Organ Donation and Transplantology, GBUZ MO MONIKI named after A.I. M.F. Vladimirsky.
12