Comment on “Clinical and endoscopic features of checkpoint inhibitor-induced colitis and their association with selective immunosuppressive therapy use”

To the Editors, Naber et al.1 report a well-characterized 2-center cohort of patients with histologically confirmed immune checkpoint inhibitor (ICI) colitis in which every endoscopy was rescored from the original images by IBD specialists, who were the first to examine endoscopic predictors of selective immunosuppressive therapy (SIT) separately by ICI class.1 We would value knowing the thoughts of the authors Naber et al. on 3 questions. The SIT diagnosis was used as a treatment decision as much as a disease state and was reached along very different timelines in the 2 groups, with a median of 7 days under anti-cytotoxic T-lymphocyte-associated protein 4 (aCTLA-4)–based therapy and 60 days under anti-programmed cell death protein 1/programmed death-ligand 1 (aPD-[L]1) therapy. The use of SIT at any point during follow-up therefore combines primary steroid refractory colitis with later steroid dependent or relapsing disease, and baseline endoscopy might not be expected to predict both equally. Since earlier SIT has been associated with shorter hospitalization and fewer steroid-tapering failures,2 thresholds for escalation may also have shifted across the 2014 to 2024 study period. Would analyzing the time to SIT with death as a competing risk and separating primary refractoriness from later escalation recover a signal that a binary endpoint may absorb?