A neural network capable of recognizing diseases at an early stage, including cancer, can be developed and implemented in medicine in five years, TASS was told by Ivan Tolmachev, Candidate of Medical Sciences, Head of the Scientific and Technological Center "Digital Medicine and Cyberphysics" of SibSMU.
According to him, the preparation of a digital dataset will take three years, model development will take six months, and registration with Roszdravnadzor will take about a year and a half. In the future, such systems will be able to predict outbreaks of diseases by analyzing micro-changes and the dynamics of pathologies based on historical data.
Tolmachev noted that the development of neural networks largely depends on the quality of digital archives. In Russia, there are no structured and unified storage facilities for medical data for the last 10-20 years, and information is fragmented between organizations, which slows down the process of creating and implementing such technologies.