Objective: To determine the cost-effectiveness of using the Oncotype DX Genomic Prostate Score (GPS), a 17-gene expression assay that can be used to inform decisions regarding active surveillance (AS) vs immediate treatment. Methods: We developed a Markov model simulating 20-year outcomes for 65-year-old men with very low-, low-, or favorable intermediate-risk prostate cancer undergoing AS vs immediate treatment using GPS vs no testing. Utilities, costs, and probabilities were extracted from the literature and National Medicare Fee Schedules to determine incremental cost-effectiveness ratios (ICER) from a payer perspective. Results: In the overall cohort, the ICER of GPS-guided therapy was $31,394 per quality-adjusted life-year (QALY). When stratified by risk group, the ICER was $25,343 per QALY in very low-risk, $28,911 per QALY in low-risk, and $39,695 per QALY in favorable intermediate-risk patients. On sensitivity analysis, findings were robust against a willingness-to-pay of $100,000 per QALY to variations in key model parameters, including the cost of annual management of AS, probability of exiting AS to treatment, cost of treatment, and probability of biochemical failure post-treatment. However, the cost-effectiveness was sensitive to small differences in the utility of AS and the utility of no evidence of disease post-treatment states. Conclusion: The use of the GPS was cost-effective in guiding treatment decisions regarding AS vs immediate treatment. The cost-effectiveness was sensitive to small differences in the utilities of the AS and no evidence of disease post-treatment states, highlighting the importance of assessing patient preferences.
ASJC Scopus subject areas