Outcomes of patients (pts) with metastatic urothelial cancer (mUC) and poor performance status (PS) receiving anti-PD(L)1 agents.

Abstract

4525 Background: Anti-PD(L)1 immune checkpoint inhibitors (ICI) prolong overall survival (OS) after platinum chemotherapy in mUC. However, clinical outcomes in pts with poor PS at time of ICI initiation are unknown. We hypothesized that ICI initiation in pts with ECOG PS 2-3 would be associated with worse outcomes vs. pts with ECOG PS < 2, and impact death location. Methods: A retrospective cohort study in 8 institutions identified pts with mUC who received ICI. Demographic, clinicopathologic, treatment (tx) patterns, tx response, and outcomes were collected. Primary endpoint: overall response rate (ORR). Secondary endpoints: median (m) OS in pts receiving ICI as 1st and 2nd line (1L, 2L); odds of dying in hospital (vs elsewhere) for pts receiving ICI (vs no tx) within 30 days of death; and estimated drug cost for pts with ICI within 30 days of death based on average wholesale price. Unadjusted logistic regression was used to assess association between ORR and ECOG PS (2-3 vs < 2) and wald test was used to compare mOS between ECOG PS (2-3 vs < 2). Results: 194 consecutive pts (30% women, 41% never smokers, median age at diagnosis 69) treated with ICI for mUC were identified. Median number of total tx lines was 2; all pts received ≥1 ICI line (6 pts received 2 ICI lines); 97, 79, 17 and 7 pts received ICI in 1L, 2L, 3L and 4L, respectively; 26% pts with ICI in 1L and 2L had ECOG PS 2-3. ORR and mOS are shown in table. Among 106 pts who died, 96 had available death location; of those, 8% received ICI within 30 days of death. Starting ICI within 30 days of death (vs no tx) was associated with higher odds of hospital death (OR 6.05, 95%CI 1.3-27.6). Estimated average ICI cost/pt within 30 days of death was $1400.58. Conclusions: Pts with ECOG PS 2-3 at time of ICI initiation had similar ORR vs ECOG PS < 2 but worse mOS. ICI initiation within 30 days from death was associated with higher likelihood of hospital death. ICI may not circumvent the negative prognostic role of poor PS, so biomarker-based pt selection is critical. Limitations include lack of adjustment for selection bias and other confounders at time of ICI initiation; data validation is ongoing. [Table: see text]