AI Sees More: Russian Students Teach AI to Read ECGs and Assist Surgeons
Artem Osadchy and Yulia Vcherashnyaya, senior students at the N. V. Sklifosovsky Institute of Clinical Medicine, won the fifth All-Russian Conference of Young Scientists on Digital Technologies in Modern Medicine. Their project, GISS AI, helps analyze electrocardiograms (ECGs) and generate detailed reports.
The idea behind the service stems from routine medical practice. ECGs are performed during preventive health screenings, in hospital emergency departments, at feldsher-midwife stations in rural areas, and in ambulances. In other words, the test can be performed in almost any setting. Interpreting the results correctly, however, can be challenging, particularly when clinicians have limited experience or must handle a high volume of patients.
As Yulia Vcherashnyaya explained, GISS AI analyzes 12-lead ECGs and looks for abnormalities in heart rhythm and electrical conduction. The clinician only needs to photograph the paper ECG tracing with a smartphone. A specialized module converts the image into a digital signal, after which a neural network processes the data and a language model generates a detailed report describing the waves, segments and leads. According to the developers, the model was trained on a dataset of 345,000 ECGs and can recognize six types of pathology.
The Service Does Not Diagnose
The students’ invention could benefit primary care physicians, paramedics, cardiologists and clinical physiologists, particularly in settings where clinicians must interpret large numbers of ECGs and cannot always obtain a prompt second opinion from a specialist.
“The service does not diagnose patients in place of a physician. It takes care of the routine work involved in the initial ECG reading, helps clinicians sort through a large volume of tracings more quickly, and immediately flags those that need attention. This is especially important at night, on weekends and during holidays, when patient volumes are high and a specialist is not readily available,” Vcherashnyaya said.
The potential benefit is straightforward: physicians can identify what is happening with a patient’s heart sooner, and patients can receive care sooner.
A Navigator for Surgeons
Another project was presented at the conference by Sechenov University. This one focuses on surgery. Mark Tokarev, an assistant in the Department of Hospital Surgery at the Institute of Clinical Medicine, is developing Laphub, an AI platform for laparoscopic procedures. These operations allow surgeons to work through small incisions using a camera and specialized instruments. The service is designed to become an indispensable assistant for surgeons.
Laphub analyzes video in real time and identifies the patient’s tissues and anatomical structures visible to the camera. It indicates where the safe operating area lies and which areas require extra caution. After surgery, the platform can automatically generate a procedure review, identifying the main stages, recording how long each took, flagging moments of elevated risk and preparing a report.
“Think of it as a navigation system for a surgeon. It doesn’t drive the car for the driver, but it warns about junctions and dangerous stretches ahead. The software sees the operation through the camera and provides real-time guidance on which anatomical structure is in view, where the safe area lies and where extra caution is needed,” Tokarev said.
The technology also has an educational purpose. Reviewing surgical video can help physicians spot details that are difficult to notice during a lengthy procedure. For surgeons in training, it can also show them which stages require particular attention.
From Project to Clinical Practice
Both projects share the same principle: digital tools should solve specific problems for healthcare professionals. GISS AI helps clinicians process ECGs faster and draw attention to potentially significant changes. Laphub is designed to support surgeons during operations.
These AI services could also be relevant to healthcare providers outside Russia. In virtually every country, physicians face heavy patient workloads, and training qualified medical professionals remains a priority. Technologies that help clinicians analyze medical data and train early-career specialists therefore have clear potential for use internationally.
