Background: FGFRalt are observed in ~20% of pts with mUC. Erda is an oral selective pan-FGFR tyrosine kinase inhibitor granted accelerated approval to treat locally advanced or mUC in adults with susceptible FGFR3/2alt who have progressed after platinum-containing chemo. THOR (NCT03390504), a randomized phase 3 study, assessed whether erda provided a survival advantage vs investigator’s choice of chemo in pts with mUC who progressed after 1 or 2 prior treatments, including an anti-PD-(L)1 agent. Methods: Pts (≥18 y) with unresectable advanced/mUC and select FGFR3/2alt (mutations/fusions), ECOG performance status 0-2, adequate organ function, progression on/after prior systemic therapy (tx) that included an anti-PD-(L)1 agent, and ≤2 prior lines of tx were randomized 1:1 to receive erda (8 mg with pharmacodynamically guided uptitration to 9 mg on day 14) QD or investigator’s choice of chemo (docetaxel or vinflunine) Q3W until disease progression or intolerable toxicity. The primary end point was overall survival (OS). Secondary end points included progression-free survival (PFS), objective response rate (ORR), and safety. Results: 266 pts were randomized: 136 pts assigned to erda, 130 to chemo. In all pts, median age was 67 y; 30% had 1 prior line of tx, 70% had 2 prior lines; 74% had visceral metastases; 90% were PD-L1 low (CPS <10). Median follow-up was 15.9 mo. The primary endpoint of the study was met, with erda significantly increasing OS and reducing the risk of death by 36%; median OS was 1 y (Table). These data met predefined stopping criteria for superiority. Erda also significantly improved median PFS (5.6 vs 2.7 mo) and ORR (46% vs 12%) vs chemo. No new safety signals were seen. Serious treatment-related adverse events (TRAEs) were observed in 13% and 24% of pts with erda and chemo, respectively, and grade (Gr) 3/4 TRAEs were observed in 46% and 46% of pts with erda and chemo, respectively. TRAEs leading to death were reported in 1 and 6 pts with erda and chemo, respectively. More TRAEs leading to dose reduction were observed with erda (66%) vs chemo (21%); 8% and 13% of pts had TRAEs leading to discontinuation of erda and chemo, respectively. Central serous retinopathy occurred in 23 pts (17%) with erda (Gr 1-2, 20 pts). Conclusions: In pts with FGFRalt advanced/mUC after prior treatment with PD-(L)1, erda significantly improved OS, PFS, and ORR vs investigator’s choice of chemo. Erda toxicity was consistent with the known safety profile. These results support the role of erda to treat pts with FGFRalt mUC after PD-(L)1 tx. Clinical trial information: NCT03390504.