Efficacy of immune checkpoint inhibitors for metastatic colorectal cancer with microsatellite instability in second or latter line using synthetic control arms: A non-randomised evaluation

Purpose Immune checkpoint inhibitors (ICIs) appeared active in single-arm trials for patients with chemoresistant metastatic colorectal cancer (mCRC) harboring microsatellite instability (MSI). Given the paucity of randomised controlled trials (RCTs) in this setting, we evaluated the effect size of ICIs using intra-patients comparison and ARCAD database as historical controls. Patients and methods Individual-patient data from NIPICOL and CheckMate 142 phase II trials that evaluated a combination of ICIs for MSI mCRC patients (N = 176) and from five non-ICI mCRC historical RCTs in second-line or latter (N = 4026) were analyzed. Firstly, promising of ICIs was identified using intra-patient comparison based on growth modulation index (GMI) defined the ratio of progression-free survivals (PFS) on ICIs and previous line of therapy. Survival outcomes of ICIs-treated patients were then compared with those matched non-ICIs treated from ARCAD database historical RCTs. Results Among ICIs-treated patients, median PFS on ICIs was 32.66 (range 0.10-74.25) versus 4.07 months (range 0.7-49.87) on prior therapy, resulting on median GMI of 4.97 (range 0.07-59.51; hazard-ratio (HR)= 0.16 (95 %CI=0.11-0.22, P < 0.001)). Compared to matched non-ICI patients, in third-line, median overall survival (OS) was not reached with ICIs versus 3.52 months with placebo (HR=0.20, 95 %CI=0.10-0.41, P < 0.001), and 6.51 months with active drugs (HR=0.30, 95 %CI=0.15-0.60, P = 0.001). In second-line, median OS was not reached with ICIs versus 11.7 months with chemotherapy+placebo (HR=0.12, 95 %CI=0.07-0.22, P < 0.001), and 16.3 months with chemotherapy+targeted therapy (HR=0.10, 95 %CI=0.05-0.19, P < 0.001). Conclusion ICIs demonstrates high effect size for MSI mCRC patients in second-line and later. This work might be useful as an example of methodology to avoid RCTs when benefit from experimental therapy is likely to be high.