Addressing Uneven Treatment Discontinuation Rate in the Chemotherapy Arm of the EV-302 Phase 3 Randomized Clinical Trial: Implications for Outcome Interpretation.

[INTRODUCTION] The EV-302 trial demonstrated a very significant overall survival (OS) benefit for Enfortumab Vedotin plus Pembrolizumab (EVP) relative to standard chemotherapy (CHT) for patients with metastatic urothelial carcinoma. However, questions have been raised regarding the high rate of treatment discontinuation in the CHT arm for reasons unrelated to adverse events or progression (33% vs. 10% with EVP, P < .01), potentially resulting in loss of unaccounted information, or informative censoring, and affecting survival results interpretation.

[MATERIALS AND METHODS] We performed a multistep analysis to assess the impact of differential dropout on trial outcomes. First, Kaplan-Meier (KM) curves were reconstructed from published data to estimate time-to-event outcomes. Second, a reverse KM analysis was conducted to evaluate censoring patterns in the overall population and key subgroups (PD-L1 expression; cisplatin eligibility). Third, simulation models were employed to test whether informative censoring could negatively impact survival benefit by EVP. Finally, we compared the CHT arm of EV-302 to those of other contemporary RCTs through reconstructed survival analyses and risk-of-bias assessments.

[RESULTS] Overall, no significant imbalance in censoring between the treatment arms of EV-302 was found on reverse KM analysis when assessing OS (P = .73); however, a significant difference was noted for progression-free survival (PFS) (P = .002). Simulation analysis revealed that even under extreme assumptions of informative censoring, the OS benefit of EVP remained statistically significant. Comparison with historical RCTs confirmed that the CHT outcomes in EV-302 were not anomalously poor. Risk of bias was low overall, although deviations from intervention and outcome measurement were flagged for EV-302.

[CONCLUSIONS] Despite the high discontinuation rate in the CHT arm, OS benefit with EVP remains robust. These findings support the reliability of EV-302 results and mitigate concerns about informative censoring, thus encouraging the use of EVP in clinical practice.