Proposal of an improved prognostic classification for pT3 renal cell carcinoma.

Purpose: The prognostic accuracy of the current TNM 2002 staging system for locally advanced renal cell carcinoma has been questioned. To contribute to the development of a more accurate classification for this stage of disease we assessed the correlation between patterns of invasion in the pT3 category and outcomes in a large multi-institutional series.
Materials and Methods: Pathological data and clinical followup on 513 pT3 renal cell carcinoma cases treated with radical nephrectomy between 1983 and 2005 at 3 Italian academic centers were retrospectively reviewed. Cause specific survival rates were calculated with the Kaplan-Meier method and multivariate analysis was performed using the Cox proportional hazards regression model.
Results: Estimated overall 5-year cause specific survival was 50.1% at a median followup of 61.5 months in survivors. The current TNM classification was not a significant outcome prognosticator. Patients with a tumor invading only the perirenal or sinus fat were at lowest risk for death from the disease. Patients at intermediate risk had tumors with invasion of the venous system alone. Simultaneous perirenal fat and sinus fat invasion or perirenal fat and vascular invasion as well as adrenal gland involvement characterized high risk tumors. Low risk tumors could be further divided into 2 groups with different outcomes based on a size cutoff of 7 cm. Our classification was a significant predictor of survival on multivariate analysis as well as M stage, N stage, Fuhrman grade and tumor size.
Conclusions: We confirm that the prognostic usefulness of the current 2002 TNM system for pT3 renal cell carcinoma is limited. We have identified 4 groups of tumors with distinct patterns of invasion and significantly different survival probabilities in this category. Large prospective series are needed to validate these findings.