OBJECTIVES: The potential cost effectiveness of pembrolizumab suggested from earlier studies were subject to immature survival data from early trial results. We aimed to re-evaluate the long-term cost-effectiveness of pembrolizumab vs. taxane-based chemotherapy in advanced urothelial carcinoma as a second-line treatment.

METHODS: We built a partitioned survival model with a 20-year time horizon with a 3^rd-party United States payer perspective. Overall survival (OS) and progression-free survival (PFS) were extrapolated from the 5-year follow-up KEYNOTE-045 trial. Costs, utilities, and adverse events were collected from the published literature. Outcomes included quality-adjusted life-years (QALYs) and incremental cost-effectiveness ratios (ICERs). We conducted deterministic and probabilistic sensitivity analyses. Additionally, we ran a scenario analysis comparing outcomes when using dynamic time-to-death utility (TTD) values vs. static health state utility values.

RESULTS: Base-case results demonstrated pembrolizumab increased costs to the system but provided additional health benefits vs. taxane-based chemotherapy for an ICER of $137,666 per QALY. The one-way deterministic sensitivity analysis suggested a high uncertainty of ICER due to OS (95% CI: more costly, less effective to $110,542). A scenario analysis showed a modest difference in incremental QALYs between the dynamic TTD estimation (0.63 incremental QALYs) vs. the static health-state utility estimation (0.61 incremental QALYs).

CONCLUSIONS: Results demonstrate that pembrolizumab may be cost-effective using a $150,000 per QALY threshold compared to chemotherapy in advanced urothelial carcinoma using long-term follow-up survival evidence. We found no difference in the conclusion between dynamic and static utility estimation scenarios.