Combining low-dose regorafenib with pembrolizumab for patients with MSI-H colorectal cancer: REGPEM-CRC-01- HCRN GI23-643.

Abstract

TPS260 Background: Pembrolizumab is one of the standard front-line immune checkpoint inhibitors (ICI) therapies for patients with MSI-H CRC. However, approximately 40% of patients treated with pembrolizumab experience early disease progression (KEYNOTE 177). Therefore, there is still an unmet need to enhance the efficacy of ICI in MSI-H CRC. Anti-VEGF therapies synergize with ICI in many tumor types. MSI-H CRC is characterized by elevated levels of serum pro-angiogenic factors, including VEGF-A (Hansen et al. Colorectal Dis. 2011). Moreover, exploratory analyses from CALBG-80405 and PARADIGM trials showed that patients with MSI-H CRC derived greater benefit from anti-VEGF therapy compared with anti-EGFR therapy, irrespective of tumor sidedness. Similarly, NSABP C-08 suggested potential activity of anti-VEGF therapy in the adjuvant setting for patients with MSI-H colon cancers. Regorafenib, a potent VEGF inhibitor, also has preclinical tumor immune modulatory effects. In this trial, we hypothesize that combining low-dose regorafenib with pembrolizumab will result in synergistic activity and lead to deep and durable responses for patients with MSI-H CRC. Methods: In the lead arm of this prospective randomized study (HCRN-GI23-643), 22 patients with treatment-naïve MSI-H CRC will be enrolled through the Hoosier Cancer Research Network. Patients may have received one cycle of pembrolizumab or up to 3 cycles of chemotherapy prior to MMR-D/MSI-H determination. Treatment consists of regorafenib 60 mg daily combined with pembrolizumab 200 mg IV during cycle 1, followed by regorafenib 90 mg in subsequent cycles with pembrolizumab every 3 weeks. The primary endpoint is ORR measured using RECIST 1.1. criteria. A formal one-sided hypothesis test will be conducted for futility, assuming that we will reject the null hypothesis of a target ORR only if we have strong evidence. We assume a null hypothesis that ORR is 0.60, which would reflect significant clinical improvement over the current standard of ORR = 0.43 from KEYNOTE 177. The alternative hypothesis is that ORR is less than 0.60. For the lead-in phase of the study, Type I error will be controlled at approximately 0.05. The test statistic will be the number of responders among the 22 patients, modeled using a binomial distribution. Clinical trial information: NCT06006923 .