RESUMO
OBJECTIVES: To evaluate the prognostic role of venous tumor thrombus consistency in patients with renal cell carcinoma. METHODS: A retrospective evaluation of the data of patients with renal cell carcinoma and a tumor thrombosis submitted to surgery from 2000 to 2013 was carried out. Histological slides were revised by two uropathologists, blinded of the clinical outcome, to assess venous tumor thrombus consistency classified as solid venous tumor thrombus consistency or friable venous tumor thrombus consistency. The statistical correlation between venous tumor thrombus consistency and other adverse features was assessed. Then the predictive ability of an integrated prognostic model, generated by Cox regression and random survival forest, was evaluated, with and without the inclusion of venous tumor thrombus consistency, by integrated Brier score, dynamic receiver operating characteristic curves, integrated discrimination improvement index and category-less net reclassification index. RESULTS: The data of 147 patients were analyzed, 79 with a solid venous tumor thrombus consistency and 68 with a friable venous tumor thrombus consistency, followed for a median period of 40.5 months. Venous tumor thrombus consistency was assessed with a high interobserver agreement (145/147 cases). The presence of a friable venous tumor thrombus consistency was associated with some adverse prognostic factors (symptoms, lymphnodal and distant metastasis, larger tumor diameter, higher cephalad thrombosis level, necrosis, microvascular invasion) and to a worse cancer-specific and overall survival at univariate analysis. However, venous tumor thrombus consistency was not predictive of survival, and did not improve the performance of a multivariable model that included a set of informative predictors. CONCLUSION: Venous tumor thrombus consistency does not seem to have an independent prognostic role in patients with renal cell carcinoma.