COST AND COMPARATIVE EFFECTIVENESS OF A GENOMIC CLASSIFIER FOR PROSTATE CANCER TREATMENT DECISIONS

Purpose:    To evaluate the cost effectiveness of genomic classifier (GC) risk-based treatment decisions for prostate cancer patients post radical prostatectomy (RP).   Methods:    We present an individualized decision analysis model to quantify the change in clinical outcomes and costs of post RP treatment uptake based on three scenarios: usual care based on percentages of adjuvant and salvage therapy reported in the literature, 100% adjuvant therapy utilization, and GC risk-based decisions. We first developed a Markov model for prostate cancer progression after RP based upon average, population-level probabilities.  We then evaluated outcomes in this model when individual GC risk probabilities (5 year probabilities of distant metastasis) were used for each individual within a simulated cohort.  We used a cohort simulation approach with a cohort of 139 patients with GC risk probabilities from Thomas Jefferson University.  For each comparison, we used the bootstrap method of sampling with replacement to choose each patient to pass through the Markov model. GC risk-based treatment decisions were based on a study assessing urologist treatment recommendations following RP in the adjuvant and salvage settings. Transition probabilities, costs, and utilities were taken from secondary sources.   Results:    Usual care resulted in 63.1% PSA recurrence-free at 5 years and an average over the 10 year post-RP horizon of 9.07 life years (LYs), 7.69 quality-adjusted life years (QALYs), and $21,980 in treatment costs.  GC risk-based treatment resulted in 73.3% recurrence-free at 5 years, 9.11 LYs, 7.61 QALYs, and $27,611 in treatment costs without the cost of the GC test.  Under 100% adjuvant therapy, 82.3% of patients were recurrence-free at 5 years, and patients had an average of 9.13 LYs, 6.85 QALYs, and $36,549 in treatment costs.  As seen in the figure below, the GC risk-based treatment remains pareto-optimal when the GC test is ≤$8,938.  Compared to 100% adjuvant therapy, GC risk-based treatment increases QALYs by 0.76 QALYs and is cost saving for a GC test ≤$8,938.   Conclusions:    GC risk-based treatment provides a balance of improvement in clinical outcomes and LYs over usual care, and increased QALYs and reduced costs compared to providing adjuvant therapy to 100% of patients.  Our model's individualized decision analysis framework can be extended to include other patient-specific factors including treatment preferences and response to therapy.