Experts report AI breakthrough in kidney cancer diagnosis

Scientists at Sechenov University, together with Vimpelcom, have created an artificial intelligence-based system for analyzing dystrophic changes in kidney tissue. The algorithm helps pathologists to accurately determine the degree of malignancy of kidney tumors. This was reported on October 28 at the First Moscow State Medical University "Газете.Ru ".

It is reported that the system analyzes digital images of tumors, identifying cells with visible nucleoli — a sign of the aggressiveness of the disease. Unlike the subjective visual assessment according to the WHO/ISUP classification, AI automatically classifies each cell on the slice.

"This technology was originally created to facilitate the routine work of pathologists, but in the course of the study it led to a fundamental discovery," the publication says.

Trained on more than 200,000 cells, the algorithm revealed a key prognostic criterion: if there are more than 11% of cells with pronounced nucleoli in the tissue, the average survival rate drops to 2.2 years versus 6 years with a low index, the publication clarifies. Based on the analysis of 50,000 cells from 144 samples, scientists have identified four stable morphological patterns with different prognosis, which do not always correspond to the official gradation.

The authors propose to supplement the international classification with two objective criteria: the threshold for the significance of the proportion of cells with nucleoli and taking into account dystrophic changes as a marker of an unfavorable prognosis.

The model is already integrated into digital analysis software and is ready for use in clinical practice, providing doctors with quantitative metrics instead of subjective assessments.

Machine potential: businesses will be able to request the necessary data for AI training

How will this affect the domestic economy

Earlier, on October 24, German Gref, Chairman of the Board of Sberbank, said during the IX conference "More than Learning" that artificial intelligence makes it necessary to review existing educational technologies. During the session, Gref and neurophysiologist Alexander Kaplan and learning expert Nick Shackleton-Jones discussed fundamental changes in human thinking in the age of AI.

All important news is on the Izvestia channel in the MAX messenger.