Timing-dependent skin toxicity phenotypes during enfortumab vedotin plus pembrolizumab therapy for advanced urothelial carcinoma

BACKGROUND: Enfortumab vedotin plus pembrolizumab (EV/Pem) has improved the survival of patients with advanced urothelial carcinoma; however, many patients suffer treatment-related adverse events, particularly skin toxicity. This study aimed to clarify the clinical impact of skin-toxicity timing on treatment outcomes. METHODS: We retrospectively analyzed patients with advanced urothelial carcinoma treated with EV/Pem. Patients were categorized into three groups according to skin-toxicity timing: no rash, early-onset rash (< 1 month), and late-onset rash (≥ 1 month). Baseline clinical and laboratory characteristics were compared among groups and incorporated into multivariable Cox proportional hazards analyses. Kaplan-Meier analyses with pairwise log-rank tests were performed to evaluate progression-free survival (PFS) and overall survival (OS). The 1-month cutoff was selected for clinical interpretability. RESULTS: The three-group analysis included 210 patients: no rash (n = 87), early-onset rash (n = 84), and late-onset rash (n = 39). Late-onset rash was associated with significantly prolonged PFS and OS in conventional analyses. In multivariable analysis, late-onset rash was significantly associated with prolonged PFS (HR 0.34, 95% CI 0.15-0.74, p = 0.007) and OS (HR 0.25, 95% CI 0.07-0.81, p = 0.02). Because late-onset rash could only be assigned after continued observation, these associations should be interpreted as exploratory and potentially affected by immortal time bias. In the 3-month landmark analysis, PFS remained significantly longer in the late-onset than early-onset rash group, whereas OS did not differ significantly. CONCLUSIONS: Skin toxicity during EV/Pem therapy appears to have timing-dependent prognostic implications. Late-onset rash showed favorable outcomes in conventional analyses, but these findings should be considered hypothesis-generating.