Assessment of the VENUSS and GRANT Models for Individual Prediction of Cancer-specific Survival in Surgically Treated Nonmetastatic Papillary Renal Cell Carcinoma

BACKGROUND: Guidelines recommend VENUSS and GRANT models for the prediction of cancer control outcomes after nephrectomy for nonmetastatic papillary renal cell carcinoma (pRCC). OBJECTIVE: To test the ability of VENUSS and GRANT models to predict 5-yr cancer-specific survival in a North American population. DESIGN, SETTING, AND PARTICIPANTS: For this retrospective study, we identified 4184 patients with unilateral surgically treated nonmetastatic pRCC in the Surveillance, Epidemiology, and End Results database (2004–2019). OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: The original VENUSS and GRANT risk categories were applied to predict 5-yr cancer-specific survival. A cross-validation method was used to test the accuracy and calibration of the models and to conduct decision curve analyses for the study cohort. RESULTS AND LIMITATIONS: The VENUSS and GRANT categories represented independent predictors of cancer-specific mortality. On cross-validation, the accuracy of the VENUSS and GRANT risk categories was 0.73 and 0.65, respectively. Both models showed good calibration and performed better than random predictions in decision curve analysis. Limitations include the retrospective nature of the study and the absence of a central pathological review. CONCLUSION: VENUSS risk categories fulfilled prognostic model criteria for predicting cancer-specific survival 5 yr after surgery in North American patients with nonmetastatic pRCC as recommended by guidelines. Conversely, GRANT risk categories did not. Thus, VENUSS risk categories represent an important tool for counseling, follow-up planning, and patient selection for appropriate adjuvant trials in pRCC. PATIENT SUMMARY: We tested the ability of two validated methods (VENUSS and GRANT) to predict death due to papillary kidney cancer in a North American population. The VENUSS risk categories showed good performance in predicting 5-year cancer-specific survival.