LAUSR

Background: With checkpoint inhibition becoming first line therapy in several cancer indications, there is a growing number of patients that are refractory to PD1 therapy. It is unclear if these patients should remain on PD1 inhibitors while new therapies are added on top or if PD1 should be discontinued when new therapies are initiated. Here we simulate a randomized controlled study comparing pembrolizumab + a checkpoint inhibitor (P+C) to a checkpoint inhibitor (C) in virtual patients that are refractory to PD1 therapy. Methods: An IO QSP model was developed that includes intra-patient heterogeneity in tumor dynamics (target, non-target, and new metastatic lesions). The model was calibrated to individual lesion data from a large multicentre, open-label phase III study (Keynote-001) in patients with metastatic melanoma. A two year phase III study comparing P+C to C was simulated with 310 virtual patients in each group. Virtual patients were removed from the trial at the time of RECISTv1.1 progression or dropout due to AE or loss to follow-up. Results: Virtual patients in the P+C arm displayed deeper (-11.5% vs 24.5%) and longer response (186.4 days vs 96.5 days) than virtual patients in the CPI arm. The percent of virtual patients displaying low prevalence progression (< 50% of the target lesions with PD) was 86% vs 52% for P+C vs. C. Median progression free survival was also longer in the P+C arm (3.8 vs 2.1 months). Difference in progression free survival were driven mainly by virtual patients that displayed oligoprogression during the pembrolizumab monotherapy run-in period. Conclusions: Simulations presented here show that virtual patients that continued pembrolizumab treatment beyond RECISTv1.1 progression (as part of a combination) did better than virtual patients that discontinue pembrolizumab (and switch to a new therapy). Citation Format: Kannan Thiagarajan, Madhav Channavazzala, Brian Gregory Topp. Does progressive disease justify pembrolizumab discontinuation: a simulation analysis. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4384.