Background: Increased vascularity is a hallmark of renal cell carcinomas (RCC), particularly clear cell RCC. The vascular endothelial growth factor (VEGF) pathway, implicated in tumor angiogenesis, is dysregulated in RCC. The phase 3 trial ECOG-ACRIN E2805 enrolled 1,943 patients (pts) with resected high-risk RCC (pT1b high grade to pT4 any grade or N any). Pts were randomized to adjuvant sunitinib, sorafenib, or placebo. Our aim was to determine the prognostic and predictive role of microvessel density (MVD), VEGF receptors, and ligands in nephrectomy specimens. Methods: We obtainedpre-treatment primary RCC tissue from 822 pts and built tissue microarrays using 3 cores from each sample. Using quantitative immunofluorescence we measured tumor MVD (area of CD34-expressing cells) and intensity of the VEGF/VEGF-R family (VEGF-R1, R2, R3 and VEGF-A, B, C, D) in tumor cells. We tested for association with disease-free survival (DFS) and overall survival (OS) by the stratified log-rank test. Associations with treatment arm and clinicopathologic variables were determined. Results: High MVD (above the median) was associated with prolonged OS for the entire cohort (p = 0.021, HR 0.63) and for pts treated in the placebo group (p = 0.014). The association between high MVD and OS was weaker in patients treated with sunitinib or sorafenib (p = 0.060). High VEGFD expression overall was associated with shorter OS (p = 0.027) but not for placebo (p = 0.16). Yet high MVD was not associated with improved DFS (p = 1.00). High MVD correlated with above-median age ( > 56) (p = 0.032), Fuhrman grade I/II (p < 0.001), clear cell histology (p < 0.001), and absence of necrosis (p < 0.001) but not with gender, sarcomatoid features, lymphovascular invasion, or tumor size. In multivariable analysis, MVD remained independently associated with improved OS for the entire cohort (p = 0.013). Conclusions: High MVD in nephrectomy specimens of high-risk RCC pts is associated with improved OS, regardless of treatment arm. MVD is thus an independent prognostic, rather than predictive, biomarker. Further studies should assess whether incorporating MVD into clinical models will predict outcome in resected high-risk RCC pts and if MVD can be used for pt selection for adjuvant therapy.