Prediction of Recurrence-Free Survival in Patients With Renal Cell Carcinoma and Tumor Thrombosis of the Renal and Inferior Vena Cava of Levels I–ii Using an Extended Cox Model and Machine Learning Methods

The aim of this study was to compare the predictive accuracy of the classical Cox regression and the neural network model DeepSurv in predicting recurrence-free survival in patients with renal cell carcinoma (RCC) and tumor thrombosis of levels I–II after radical surgical treatment. The retrospective study included 100 patients who underwent nephrectomy with thrombectomy. Model evaluation was based on the concordance index (C-index) and interpreted using SHAP analysis. The DeepSurv model demonstrated higher accuracy on the test sample compared to the classical Cox model and also identified the most significant risk predictors, including tumor size, ISUP grade, and level of thrombosis. The obtained data confirm the feasibility of using machine learning methods for personalized prognosis and optimization of monitoring strategies in patients with RCC.