Overall survival (OS) by response to first-line (1L) induction treatment with atezolizumab (atezo) plus platinum/gemcitabine (plt/gem) vs placebo plus plt/gem in patients (pts) with metastatic urothelial carcinoma (mUC): Updated data from the IMvigor130 OS final analysis

4503 Background: For pts without disease progression during 1L plt-based chemotherapy (chemo), maintenance immunotherapy is a new mUC treatment option. IMvigor130 was a global, randomized Phase III study evaluating 1L atezo + plt/gem (Arm A) vs atezo monotherapy (Arm B) and placebo + plt/gem (Arm C) in patients with mUC (Galsky Lancet 2020). The final analysis showed that improved OS in Arm A vs Arm C of the intention-to-treat (ITT) population did not reach statistical significance (Galsky ASCO GU 2023). Here we report a post hoc analysis examining OS outcomes by response during atezo/placebo + chemo “induction” based on the final OS analysis. Methods: For Arms A and C, per protocol, investigators pre-specified the type of chemo pts received (gem + either cisplatin [cis] or carboplatin [carbo]), which was also a randomization stratification factor. Pts without progressive disease (PD) were allowed to continue atezo or placebo after 4 to 6 cycles of plt/gem. This post hoc analysis evaluated post-induction (Wk 18) OS in pts who completed 4 to 6 cycles of chemo followed by ≥1 dose of atezo or placebo and who had a best response of at least stable disease (SD) without PD at any time (up to and including Wk 18 tumor assessment). Post-progression OS was evaluated in pts who had PD at any time (up to and including Wk 18 tumor assessment). OS analyses by type of chemo were also performed. Multivariable Cox proportional hazards models were used, with HRs adjusted for known prognostic factors (and response for non-PD pts) and stratified by enrollment stage. Results: The time from last pt randomized to the data cutoff (Aug 31, 2022) was 49 mo (overall ITT population). OS improvements favoring Arm A vs C appeared greater for pts treated with cis than with carbo. In the cis subgroup, OS rates at 36 mo were 47% (Arm A) and 34% (Arm C) for pts who achieved at least SD during induction. Additional efficacy data are shown. Conclusions: In this post hoc analysis, the initial response to induction therapy did not seem to impact OS outcomes. Consistent with prior analyses, these data suggest that cis-treated pts may derive a greater benefit from the addition of atezo than carbo-treated pts. Clinical trial information: NCT02807636 . [Table: see text]