9509 Background: High-risk uveal melanomas (UM) recur distantly in over 50% by 3 years (median distant metastatic-free survival (DMFS) 32 months). No effective treatment currently exists to reduce the risk of metastatic disease and patients (pts) with metastasis have a very poor prognosis with few therapeutic options. The combination of Nivolumab/Ipilimumab (Nivo/Ipi) has shown limited efficacy in pts with metastatic UM. However, preclinical data suggest differential biology between local vs metastatic UM. We performed a prospective multi-center phase II study of adjuvant Nivo/Ipi in pts high-risk UM patients to compare DMFS to a cohort of UM with the same tumor characteristics. Methods: This investigator-initiated study enrolled pts with high-risk UM, defined as a predicted 3-year distant relapse-free survival (DMFS) of approximately 50% (based on gene expression profiling (GEP) (DecisionDx-UM) class 2 with tumor largest basal diameter (LBD) of 12mm or higher). Pts received Nivo 240mg iv every 2 weeks and Ipi 1mg/kg every 6 weeks for up to 48 weeks. The target sample size was 50 evaluable pts. The primary endpoint was 3-year DMFS. Secondary endpoints were median DMFS, overall survival (OS), and toxicity (AE). In addition, a propensity score method (double robust estimate in R adjusted curves) is used to estimate the average DMFS and compare the nonrandomized treatment (tx) and the control patients. 119 pts from the Cooperative Ocular Oncology Group (COOG) database were defined as high risk, matched for tumor characteristics (GEP class 2, LBD≧12mm) and were used as a control cohort. Results: Fifty-two pts were enrolled and 50 were treated from 12/18/2018-9/29/2021. As of 11/6/2023, with a median follow-up of 36 months, the 3-year DMFS rate vs COOG control, was 69.1% vs 45.1% (two-sided p = 0.012) without any adjustment; and 70.4% vs 43.4% (two-sided p = 0.018) with the propensity score method. The median DMFS was not reached (NR) in the tx cohort vs 33.8 months. There was a reduction in the risk of distant recurrence with Nivo/Ipi vs control cohort (RR 0.52, 95%CI, 0.309-0.888). Median OS has not been reached with only 7 deaths from UM in the tx cohort so far. Grade ≥3 any-cause AE occurred in 27 pts (54%). Grade ≥3 treatment-related (TR) AE occurred in 48%. 22 pts (44%) discontinued the treatment due to toxicity. One pt died due to immune-related myocarditis. Conclusions: This phase II study of adjuvant Nivo/Ipi resulted in clinically meaningful and statistically significant improvement in 3-year DMFS of 70.4% vs 43.4% in control for pts with high-risk UM from the COOG database matched for tumor characteristics. Further investigation of this regimen as adjuvant therapy of high-risk UM is warranted. Clinical trial information: NCT03528408 .