• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

AI in Medicine


Maria Poptsova,
HSE University

 Location: HSE University Cultural Centre, Z201

 Time: October 26, 15:35

Speakers and titles

Michael Nathenzon

National telemedicine agency

PhthisisBioMed medical service for automated analysis of chest radiographs/fluorograms. Artificial medical intelligence. Ultimate capabilities.

In the course of the three-year (starting from 2020) Moscow experiment
The organisers and developers have developed methodologies for the
of the use of weak IMI and successfully introduced it into the systems of
health care systems at the regional level. The experience of developing
medical IMI on the example of IMI-service "PhthisisBioMed" and the results of its application in real clinical conditions.
the results of its application in real clinical practice.
This IMI-service has shown its quality and reliability, which is confirmed by technological monitoring.
is confirmed by technological monitoring. Medical IMI-service
"PhthisisBioMed is registered as a medical device and
allows through the Medical Decision Support System to accelerate the process of obtaining medical treatment.
and cheapen the process of obtaining a medical opinion.
The analysis of external and internal limitations of weak
of the weak artificial medical intelligence, and solutions to overcome them through the creation of a Strong Medical Intelligence are given.
solutions to overcome them through the creation of Strong AI are given.
Maria Poptsova

HSE University

A prognostic system for predicting adverse outcomes in cardiac patients.

The presentation will describe the development of a predictive system based on clinical data and patient biomarkers to predict adverse outcomes of acute coronary syndrome therapy over a horizon of up to 10 years. It will be shown that the GRACE risk scale makes a significant contribution to prediction and has prognostic significance in the Russian population, biomarkers have predictive power, and only using GRACE and PCSK9 it is possible to build a model that is superior to a model based on clinical features.
Maxim Sharaev


BIoMedically-informed AI joint UAE-Russian laboratory (BIMAI-lab): interpretable AI for MRI and precision medicine.

Recently, a joint lab was established between Skoltech Applied AI Center and University of Sharjah (UAE). This BIoMedically-informed AI laboratory (BIMAI-lab) aims at applying novel AI&ML methods to complex clinical data-intensive tasks, including medical decision support systems creation and biomarkers search. Two examples of current lab projects will be presented and discussed during the report. The first is dedicated to creating interpretable ML models for early detection of dementia (Alzheimer Disease and Mild Cognitive Impairment) based on structural 3D MRI data, which is an important clinical task being already explored with Moscow PKB #1 Hospital. The second project has just started jointly with the Center of Excellence for Precision Medicine of University of Sharjah (UAE), and aims at developing new AI methods and mathematical models for spatiotemporal multi-OMICs of histopathology (including single-cell analysis) in different types of cancer.
Diagnostic models. Experience of development and implementation.

We will tell you about the development of models that determine diagnosis from text data contained in electronic medical records, from problem formulation to implementation. These models are integrated into EMIAS and are used daily by thousands of Moscow doctors.
Arjuna Scagnetto

ASUGI (Italy)

Prediction of Atrial Fibrillation with CNN and how to benefit from merging different source of information.