LAUSR

BACKGROUND AND AIM Immune checkpoint inhibitors (ICIs) have shown promise in treating a variety of cancers. Their increased use coincides with increased incidence of immunotherapy-mediated colitis (IMC), a common adverse effect. Optimal strategy for endoscopic evaluation of IMC (full colonoscopy or flexible sigmoidoscopy) is not well-defined. METHODS Retrospective review of all patients at City of Hope referred to gastroenterology for evaluation of IMC due to gastrointestinal symptoms was performed. Patients with an existing histologic diagnosis of IMC established at an outside hospital or a diagnosis of infectious or chronic colitis were excluded. RESULTS We identified 51 symptomatic patients on ICIs prompting evaluation for IMC with colonoscopy (47/51) or flexible sigmoidoscopy (4/51). All distal rectosigmoid biopsies during flexible sigmoidoscopy demonstrated histologic evidence of IMC. In full colonoscopy, IMC was either present in all segments of colon simultaneously (35/47) or absent from all segments (12/47). No isolated proximal colonic biopsies demonstrated IMC. Endoscopically normal mucosa demonstrated histologic evidence of IMC up to 68.6% of the time. Endoscopically abnormal right, transverse, and left colon had low sensitivity (35.3%, 34.3%, and 41.7% respectively) and high specificity (100.0%, 100.0%, and 91.7% respectively) for histological presence of IMC. CONCLUSIONS Distal colon biopsies in patients on ICI therapy with diarrhea and suspected IMC were sufficient for diagnosing IMC in our cohort. Further, we found histologic evidence of IMC in biopsies taken from normal-appearing mucosa in a number of patients, suggesting that a normal endoscopic appearance does not preclude the presence of IMC and biopsies should be taken from both normal and abnormal-appearing mucosa.