Russian Scientists Create AI System to Assess Cancer Malignancy
A new algorithm can pinpoint tumor aggressiveness by analyzing individual cells—offering doctors objective data instead of subjective visual estimates.
Researchers from the Institute of Regenerative Medicine at Sechenov University, together with VimpelCom (Beeline), have developed an AI-powered system that analyzes kidney tissue samples to assess cancer malignancy with high precision. Initially designed to automate routine pathology tasks, the project led to a breakthrough discovery: the algorithm can determine a tumor’s malignancy grade and even predict patient survival outcomes.
Trained on 200,000 cell images, the system examines digital histological slides of tumors and automatically detects and classifies cells with visible nucleoli—a key morphological marker linked to tumor aggressiveness.
Personalized Prognosis and Targeted Therapy
Using AI, the researchers identified four distinct morphological tumor patterns, each associated with a different prognosis. These patterns outperform the traditional WHO/ISUP classification, where tumor grading still relies largely on visual assessment.
Following pilot testing, the model was integrated into clinical diagnostic software to support oncologists in analyzing tumor samples. The system aims to make cancer diagnostics more accurate, data-driven, and tailored—helping doctors choose the right therapy for each individual patient.
