P566 Colitis due to immunotherapy with anti-PD-1: a case series

Background Anti-programmed death-1 (PD-1) drugs have been associated with immune-mediated gastrointestinal adverse events in 0.5-9% of the cases. The aim of this study was to investigate the characteristics of colitis occurring in oncological patients treated with anti-PD-1 drugs. Methods This is a prospective cohort study including patients referred for new onset colitis at Azienda Ospedaliera of Padua between 2020 and 2022. All patients were treated with anti-PD-1 drugs at Istituto Oncologico Veneto due to oncological disease [n=2 melanoma, n=1 colorectal cancer, n=1 pleural mesothelioma, n=14 non-small-cell lung cancer (NSCLC)]. Median duration of follow up was 10 months. Results Eighteen patients who developed diarrhea (≥G2, CTCAE) following treatment with anti-PD-1 were included (n=12 males, median age 70 years). The diarrhea began after a mean of seven months of therapy and after two to six months its discontinuation in 3/18 patients. The mean fecal calprotectin values at diarrhea onset was 1063 ug/g. Most of the patients (14/18) had their immunotherapy suspended, among them four required hospitalization. All patients were initially treated with systemic steroids (1 mg/kg). Seven patients had ulcerative colitis (UC)-like colitis (n=5 pancolitis, n=2 left colitis) while the others (60%) had microscopic colitis (MC, n=9 collagenous, n=1 lymphocytic) with unspecific inflammatory mucosal abnormalities in half of the cases. Among patients with UC-like colitis, two were hospitalized and treated with anti-TNFα because of steroid refractoriness; all remaining patients with UC-like colitis are maintaining remission with oral mesalamine or with budesonide non-MMX or low-dose systemic steroid (0.1 mg/kg). Two patients restarted anti-PD-1 drugs with no colitis relapse in the following three months, with mesalamine as maintenance therapy. Among patients who developed a MC, one required Vedolizumab as maintenance therapy, while 8/10 achieved partial or complete remission with oral budesonide non-MMX or low-dose systemic steroid. Three patients restarted anti-PD-1 drugs with no colitis relapse in the following three months, taking budesonide non-MMX as maintenance therapy. In our cohort of study all patients with MC were previously treated with anti-PD-1 drugs for NSCLC, while those with UC-like colitis were treated for NSCLC (n=4), melanoma (n=2) and colorectal cancer (n=1). Conclusion We observed a significant proportion of microscopic colitis in patients with NSCLC, while UC-like colitis was equally distributed in the other forms of neoplasia. Men are at higher risk to develop anti-PD-1 induced MC in NSCLC, rather than women. Anti-PD-1 induced microscopic colitis presents more frequently endoscopic signs of inflammation.