Clinical and genomic predictors of outcomes in patients (pts) with advanced urothelial carcinoma (aUC) treated with enfortumab vedotin and pembrolizumab (EVP): UNITE study analysis.

4560 Background: EVP is the recommended 1 st line (1L) therapy for pts with aUC and is moving into the perioperative setting. There is limited understanding of clinical factors and biomarkers predictive of EVP outcomes. Methods: We analyzed pts with aUC treated with EVP in the multi-site retrospective UNITE study. χ 2 test and logistic regression were used to assess observed response rate (ORR) and disease control rate (DCR). Log-rank test and Cox proportional hazard models were used to assess progression-free survival (PFS) and overall survival (OS) from EVP start. Multivariable analysis (MVA) included age, sex, race, ECOG performance status (PS), neutrophil-to-lymphocyte ratio (NLR), albumin, hemoglobin (Hgb), primary tumor site, histology, treatment line, and metastatic sites as covariates. Results: A total of 521 pts with aUC from 16 sites received EVP. Median age was 71, 75% were male, 81% were Caucasian, 76% had ECOG PS 0-1, and 72% were treated with 1L EVP. 74% had lower tract primary tumor, 65% pure urothelial histology, 18% liver metastases (mets), 43% non-liver visceral mets, and 28% lymph node (LN)-only mets. At 8.1 mos median follow-up from EVP start, the overall cohort had ORR 55%, DCR 82% 16% complete response (CR), 39% partial response (PR), 27% stable disease (SD), median PFS (mPFS) 6.5 mos, and median OS (mOS) 19.8 mos. For 1L EVP pts, at 9.8 mos median follow-up, ORR was 59%, DCR 83% (20% CR, 39% PR, 24% SD), mPFS 10.1 mos, and mOS 23.3 mos. In both the overall cohort and 1L EVP pts, clinical factors associated with longer PFS and OS included ECOG PS 0-1, Caucasian race, albumin ≥ 3.5 g/dL, Hgb ≥ 10g/dL, NLR < 5, and absence of visceral mets. CR was associated with ECOG PS 0-1 (OR 3.6, 95% CI 1.4-12.5, p = 0.02), Hgb ≥ 10g/dL (OR 2.3, 95% CI 1.2-4.8, p = 0.02), NLR < 5 (OR 2.7, 95% CI 1.5-5.3, p = 0.002), LN-only mets (HR 3.8, 95% CI 1.7-9.8, p = 0.003), immunotherapy-naive status (OR 4.1, 95% CI 1.4-17, p = 0.02), and 1L EVP for aUC (OR 3.3, 95% CI 1.7-7.3, p = 0.001). Pts with CR had longer mPFS than pts with PR or SD (26.5 vs. 14.5 vs. 7.5 mos, p < 0.001) and longer mOS (NR vs. 33.4 vs. 14.5 mos, p < 0.001). Among 315 pts with molecular profiling, the most common alterations were in TERTp (55%), TP53 (45%), FGFR3 (22%), ARID1A (19%), KMT2D (18%), PIK3CA (18%), and CDKN2A/B (17%). In MVA, KMT2D alterations were associated with lower CR rate (6% vs. 19%, p = 0.05), shorter mPFS (5.0 vs. 8.0 mos, p = 0.04), and shorter mOS (12.0 vs. 20.0 mos, p = 0.04). Very high tumor mutational burden (≥ 15 mut/Mb) was associated with higher DCR (97% vs. 79%, p = 0.04) but no significant difference in PFS or OS. Conclusions: In this large real-world cohort, ECOG PS, metastatic site, treatment line, and prior immunotherapy exposure were associated with EVP outcomes. KMT2D -altered aUC may represent a high-risk molecular subset with worse outcomes. Further validation in prospective cohorts is needed.