AI Model from Russian Researchers Predicts Long-Term Survival for Diabetes Patients
A neural-network system developed at Lobachevsky University identifies ten key biomarkers that can forecast all-cause mortality risk with high accuracy.
Scientists from the Institute of Biology of Aging and the Artificial Intelligence Research Center at Lobachevsky University have developed an AI model capable of predicting the risk of death from any cause in patients with diabetes.
The model was trained on physiological indicators from more than 550 diabetic patients who were monitored over a 17‑year period. From an extensive dataset, the neural network selected ten key biomarkers that shape a long-term prognosis. The model’s survival‑prediction accuracy reaches 84 percent.
Personalized Risk Maps
According to the study’s lead author, Mikhail Ivanchenko, director of the Institute of Biology of Aging at Lobachevsky University, SHAP‑based interpretation makes it possible to identify connections among dozens of patient‑health parameters. The most powerful predictors of mortality in diabetes were age, disease duration, and the number of complications.
Depending on the dominant risk factor, physicians can tailor individualized treatment strategies.
