Combined Positive Score and Cost-Effectiveness of Perioperative Pembrolizumab for Head and Neck Cancer
Aidan H. Coyle, David W. Hutton, Marisa R. Buchakjian, Keith A. Casper, David W. Forner, Sarah Hawley, Molly E. Heft Neal, Kelly M. Malloy, Michelle L. Mierzwa, Mark E. Prince, Jennifer L. Shah, Paul L. Sweicicki, Andrew G. Shuman, Chaz L. Stucken, Christine M. Veenstra, Francis P. Worden, Pratyusha YalamanchiImportance
The KEYNOTE-689 trial demonstrated improved event-free survival with the addition of perioperative pembrolizumab to standard of care for patients with locally advanced head and neck squamous cell carcinoma (HNSCC), yet the cost-effectiveness remains unknown.
Objective
To evaluate the cost-effectiveness of the KEYNOTE-689 perioperative pembrolizumab protocol.
Design, Setting, and Participants
A partitioned survival model simulated the KEYNOTE-689 trial cohort with 3 mutually exclusive health states: event-free survival, disease progression, and death. Model inputs included Healthcare Cost and Utilization Project–derived and US Centers for Medicare & Medicaid Services–derived costs for drug acquisition (pembrolizumab and cisplatin), administration, surgery, radiation therapy, supportive care, management of serious adverse events, and disease progression.
Main Outcome and Measures
KEYNOTE-689 Kaplan-Meier survival estimates were digitized and data for the combined positive score (CPS) of 1 to 10 subgroup were derived from the reported CPS of 1 or greater and CPS of 10 or greater subgroups. Quality-adjusted life-year (QALY) values for different health states were sourced from published literature, with the US population preference-weighting algorithm applied to EuroQol 5-Dimension 3-Level health questionnaire data from the CheckMate 141 trial. The primary outcome was the incremental cost-effectiveness ratio (ICER) results stratified by programmed cell death 1 ligand 1 (PD-L1) combined positive score (CPS of 1 to 10 vs greater than 10). A probabilistic sensitivity analysis (PSA) was performed to evaluate parameter uncertainty.
Results
Subgroup analyses based on PD-L1 expression yielded divergent results. In the CPS greater than 10 cohort, perioperative pembrolizumab was cost-effective, producing approximately 1.35 additional QALYs at an incremental cost of $181 900, resulting in an ICER of $134 700. PSA showed a 57% probability that the pembrolizumab strategy was cost-effective at a $150 000 willingness-to-pay (WTP) threshold. In the CPS of 1 to 10 cohort, perioperative pembrolizumab resulted in approximately 0.08 additional QALYs at an incremental cost of $166 500, for an ICER of $2 179 400, with the PSA demonstrating an 11% probability of cost-effectiveness at a $150 000 WTP threshold.
Conclusions and Relevance
In this economic evaluation, addition of perioperative pembrolizumab to standard treatment for locally advanced HNSCC may represent a cost-effective intervention for patients with a PD-L1 CPS greater than 10 if KEYNOTE-689 data are replicable in clinical practice. Perioperative pembrolizumab is not a cost-effective strategy for patients with a CPS less than 10 based on current data. Biomarker stratification may inform the value-based integration of immunotherapy among patients with HNSCC.