Background: In the RADIANT-3 study, everolimus (EVE) improved progression-free survival (PFS) by 6.4 mo versus placebo (PBO; HR, 0.35; 95% CI, 0.27-0.45; P < 0.001) in patients (pts) with advanced, progressive pancreatic neuroendocrine tumors (pNET). Here we present final overall survival (OS) results of the RADIANT-3 study with updated safety and also report OS adjusted for confounding due to crossover. Methods: A total of 410 pts with advanced, progressive, low-/intermediate-grade pNET were randomized to EVE 10 mg/d (n = 207) or PBO (n = 203), both with best supportive care. Upon disease progression in double-blind phase, crossover from PBO to open-label EVE was allowed at investigator’s discretion. At the end of core phase, ongoing pts from both arms were unblinded and switched to open-label EVE. OS was analyzed by one-sided stratified log-rank test. Rank-preserving structural failure time (RPSFT) analysis was performed to estimate treatment effect corrected for crossover bias. Results: A total of 225 pts received open-label EVE; including 85% (172 of the 203) pts randomized to PBO arm. Median OS (95% CI) was 44.0 (35.6-51.8) mo for pts initially randomized to EVE and 37.7 (29.1-45.8) mo for those randomized to PBO (HR, 0.94; 95% CI, 0.73-1.20; P= 0.30). RPSFT analysis showed a relative survival benefit of 3.27 (95% CI, 0.10-13.93) with survival rates of 82.6% vs 74.9% and 67.7% vs 55.6% at 12 and 24 mo, respectively for EVE vs RPSFT corrected PBO arm. The most frequently reported drug-related adverse events included stomatitis (45%), rash (37%), and diarrhea (26%) in open-label EVE arm. Conclusions: The randomized, placebo-controlled, phase III RADIANT-3 study reports unprecedented median OS of 44 mo with EVE in advanced, progressive pNET. Although statistically not significant, a survival benefit of 6.3 mo with EVE was observed, consistent with PFS benefit reported previously. A stronger OS advantage with EVE after a correction for crossover effect confirms that crossover of majority of PBO pts (85%) likely confounded the survival results. The safety profile of EVE remained consistent with earlier experience. Clinical trial information: NCT00510068