P298 Checkpoint inhibitor colitis: insights from bench and bedside

Aims Immune checkpoint blockade (ICB) is the mainstay of treatment for metastatic melanoma and lung adenocarcinoma, and their use is growing in cancer. Immune checkpoint blockade induced colitis (ICB colitis) presents a management challenge and its mechanisms remain poorly elucidated. Methods We performed next generation single-cell RNA sequencing of immune cells from patients given ICB. We validated findings using confocal microscopy, drawing comparisons bioinformatically with ulcerative colitis. In parallel, we conducted a review of 1,074 patients given checkpoint inhibitors across two tertiary centres between 2011–2018 to discover patterns in incidence and predictors of clinical outcome. Results Using single-cell RNA sequencing, we discovered excessive local CD8 T cell proliferation was a key feature of ICB colitis, and we were able to visualise higher numbers of replicating CD8 T cells in gut tissue sections of patients with colitis than those without ICB colitis. The degree of replication was greater than seen in ulcerative colitis by bioinformatic analysis. From our clinical review, age, gender and smoking status did not alter the risk of developing colitis, whereas type of immunotherapy did (incidence 9% in PD-1 Monotherapy vs 32% Combination Therapy). Having prior IBD did not guarantee the development of ICB Colitis. Systemic markers of inflammation (C-Reactive Protein, Albumin) did not predict outcome, whereas local markers of inflammation (endoscopic UCEIS scoring, histological Nancy Index) did. Conclusions We putatively link novel insights from bench science to clinical trends. ICB colitis may be driven more by a gut localised inflammation response in comparison to ulcerative colitis. As we also demonstrate, this may explain why endoscopic and histological scoring may have better prognostic value than systemic measures of inflammation.