Clinical Outcomes of Elective Early Discontinuation of Immunotherapy Based on Objective Response in Microsatellite Instability-High Metastatic Colorectal Cancer

Abstract

Patients with microsatellite-high (MSI-H) metastatic colorectal cancers (CRC) may experience long-lasting benefit from immune checkpoint inhibitors (ICI) upon stopping therapy. However, optimal timing and patient selection criteria for early treatment withdrawal remain undefined. In this single-center retrospective study, we characterized the clinical response and associated survival outcomes of patients who received elective early versus late treatment discontinuation. We retrospectively analyzed patients with MSI-H metastatic CRC treated with ICI therapy from May 2015 to April 2024. Early ICI discontinuation was defined as treatment withdrawal before 2 years, and late ICI discontinuation as after 2 years. Response was assessed using Response Evaluation Criteria in Solid Tumors. Progression-free survival (PFS) and overall survival (OS) were estimated using the Kaplan Meier method. Efficacy outcomes between early and late ICI discontinuation groups were compared using a log-rank test. Of 36 patients with MSI-H metastatic CRC, 12 underwent elective early ICI discontinuation and 9 experienced late ICI discontinuation. After a median follow-up of 32 months post-treatment, 91.7% (11/12) in the early discontinuation group remain off therapy without progression. PFS and OS outcomes between the early and late discontinuation groups were similarly favorable ( = .88 and = .85, respectively), despite a 12-month difference in median duration of ICI therapy (13.3 and 25.6 months, respectively). The most common reason for elective early treatment discontinuation was clinical remission (n = 10), defined as a complete response, or a partial response with negative PET and/or ctDNA testing. Early ICI discontinuation guided by response criteria resulted in low rates of recurrence. Survival outcomes between early and late ICI discontinuation groups were comparable, suggesting that treatment duration can be individualized based on clinical response without compromising favorable long-term prognosis.