Effect of pathological high-risk features on cancer-specific mortality in non-metastatic clear cell renal cell carcinoma: a tool for optimizing patient selection for adjuvant therapy.

PURPOSE: Adjuvant therapies for non-metastatic renal cell carcinoma (nmRCC) are being tested to improve outcomes in patients with high-risk (hR) nmRCC. The objective of the current study is to test the ability of three hR features to identify patients who are at the highest risk of cancer-specific mortality (CSM) after partial or radical nephrectomy.

METHODS: Within the Surveillance Epidemiology and End Results (SEER) database (1988-2013), we identified 23,632 nm "clear cell" RCC partial or radical nephrectomy patients with hR features: Fuhrman grade (FG) 3 or 4 or pathological classifications T3a or T3b or lymph node invasion (LNI), or combination of these. Kaplan-Meier analyses (KM) and multivariable Cox's regression models (CRM) evaluated the effect of hR features on CSM.

RESULTS: Overall 11,568 (48.9%) patients harbored FG3-4, 5575 (23.6%) pT3a/b, 140 (0.6%) LNI, 5366 (22.7%) FG3-4 and pT3a/b, 183 (0.8%) LNI and pT3a/b, 203 (0.9%) LNI and FG3-4 and 597 (2.5%) LNI, FG3-4 and pT3a/b. Median CSM-free survival was 51, 58 and 22 months for LNI and pT3a/b, for LNI and FG3-4 and for LNI, FG3-4 and pT3a/b and was not reached for the other groups. These results remained unchanged in multivariable CRMs, where all hR features represented independent predictors.

CONCLUSIONS: Individuals with combination of LNI with FG3-4 or pT3a/b and patients with all three hR features are at highest risk of CSM. In consequence, these patients may represent ideal candidates for adjuvant therapy either in clinical practice or future prospective trials.