Abstract A33: Markov model of VeriStrat© selection criteria for EGFR Inhibitor therapy predicts improved survival for previously treated patients with NSCLC

Abstract

Abstract Background: EGFR inhibitor therapy extends survival as 2nd or 3rd line treatment of patients with previously treated metastatic non-small cell lung cancer (NSCLC), but a randomized trial of 2nd line therapy in unselected patients showed similar benefit versus conventional chemotherapy. EGFR mutation analysis is useful in choosing therapy, but it is applicable to only a minority of patients due to low frequency of mutation in a North American population and ignores the majority of patients who may benefit with wild type status. Alternative selection strategies are needed. VeriStrat is a protein expression profile utilizing matrix-assisted laser desorption/ionization time of flight mass spectrometry to analyze serum. The test classifies patients as VeriStrat Good or VeriStrat Poor based on likelihood of benefitting from EGFR inhibitor therapy. The objective of this research was to model the anticipated survival of patients using a Veristrat Selection Strategy for 2nd line therapy of NSCLC to the alternative treatment strategies of chemotherapy for all (Chemoall), EGFR inhibitor to all (EGFR-all) l, or to performance status selection (PS selection), a common practice, where those with good performance status receive chemotherapy and those with poor PS receive EGFR inhibitor therapy. Methods: We developed a Markov model to estimate the expected survival of patients for 4 different treatment strategies for NSCLC: (1) VeriStrat Selection, (2) chemotherapy to all patients, (3) EGFR inhibitor to all patients, and (4) PS selection. Model inputs, taken from published literature, for the base-case analysis (and ranges) were the following: PS 0-1, 50% (35%–65%); VeriStrat Good, 60% (45%–75%); VeriStrat Good if PS >1, 40% (25%–55%); 1-year survival on EGFR inhibitor if VeriStrat Good, 47% (43–50); 1-year survival on EGFR inhibitor if VeriStrat Poor, 12% (8–16) and 1-year survival on chemotherapy, 29% (23–35). Probabilistic sensitivity analysis was performed using 10,000 2nd order Monte Carlo simulations. Results: In the base case analysis, the median survival for EGFR-all was 8.97 months, Chemo-all was 8.48 months, PS selection was 8.20 months and VeriStrat selection was 10.03 months. In the sensitivity analysis, the VeriStrat selection strategy yielded the longest survival in each of the 10,000 simulations. Conclusion: These results suggest that VeriStrat-based selection of “standard of care” 2nd line therapy of NSCLC improves survival to a clinically significant degree. This strategy, in the absence of new therapies, appears to improve median survivals by 4–6 weeks.