S. Wadhwa1, N. J. West1, P. Philips1, R. C. Martin1, C. R. Scoggins1, K. M. McMasters1, M. E. Egger1 1University Of Louisville, Hiram C Polk Jr, MD Department Of Surgery, Louisville, KY, USA
Introduction: Adjuvant immunotherapy for resected stage IIB/C and III cutaneous melanoma is now approved. However, there is a wide range of prognosis across stage II and III patients who undergo staging with sentinel lymph node biopsy (SLNB). Previous cost-effectiveness studies have focused on particular stages of disease. We sought to evaluate the cost-effectiveness of adjuvant immunotherapy as a function of age and risk of recurrence in a stage-agnostic manner so that the decision-making could be generalized for any stage IIB/C or III melanoma patient staged with SLNB.
Methods: A decision tree analysis was performed modeling the decision to give adjuvant immunotherapy. Across a range of predicted 5-year risks of recurrence, we calculated the proportion of people expected to be alive without recurrence after 5 years in both the adjuvant therapy and observation groups. Costs of treatment were estimated based on treated with either pembrolizumab or nivolumab. Calculations were performed separately for the following ages: <50, 50-64, 65- 69, 69-79, 80-84. The Incremental Cost Effectiveness Ratios (ICER) were calculated based on the increased number of patients alive and the increased costs from adjuvant therapy compared to observation alone in US dollars. Sensitivity analyses were performed across ranges of drug costs and hazard ratios for effectiveness of therapy.
Results: The predicted risk of recurrence was the dominant factor in determining the cost-effectiveness of adjuvant immunotherapy. Calculated ICERs and NNT decreased rapidly with increasing predicted 5 year rates of recurrence (Figure). The ICER reached approximately $3,000,000/life year once the 5 year risk of recurrence reached 20%; the NNT was 24. This ranged from $2,999,477 for ages < 50 to $4,305,399 for ages 80-84. NNT remained below 20 for 5-year risk of recurrence > 25%. The estimates for treatment with nivolumab were similar to pembrolizumab. ICER estimates and NNTs were robust across the sensitivity analyses varying cost and hazard reduction estimates. Assuming a traditional willingness-to-pay threshold of $100,000/life year, no adjuvant immunotherapy is cost-effective with an ICER < $100,000 until the 5-year risk of recurrence approaches 70-80%.
Conclusion: The cost-effectiveness of adjuvant immunotherapy for patients with stage IIB/C-III melanoma can be estimated in a stage-agnostic fashion based on the 5 year risk of recurrence. The high cost of treatment with these agents makes adjuvant therapy cost-ineffective using traditional ICER thresholds, but the NNT are relatively low for adjuvant therapy decision making. Reducing the cost of these immunotherapies would have the greatest impact on cost-effective-based medical decision making.
