P272 Immune checkpoint inhibitor-related colitis assessment and prognosis: can inflammatory bowel disease scoring point the way?

Background

Immune checkpoint inhibitors (ICI) have revolutionised oncological care but cause immune-related adverse events (irAE). Inflammation of the gastrointestinal tract (enterocolitis) is the most frequent cause of significant morbidity and cessation of ICI. We sought to determine whether Common Terminology Criteria for Adverse Events (CTCAE) classification of colitis, inflammatory bowel disease (IBD) biomarkers, endoscopic scores and histology scores correlate with disease outcome.

Methods

A retrospective study was performed on patients receiving ICI for metastatic melanoma, non-small cell lung cancer and urothelial cancer from 2012 to 2018 at two UK centres by reviewing the electronic patient records and oncology prescribing databases to identify patients who had irAE colitis. irAE colitis was defined as diarrhoea requiring steroid/infliximab therapy for resolution and/or with endoscopic/histological confirmation. Demographics, clinical data, endoscopic findings, histopathology reports and treatment outcomes were obtained. Patients were divided into three categories: (1) Mild-moderate disease where diarrhoea settled rapidly following a course of steroids ≤60 days; (2) Refractory or moderate–severe colitis requiring steroids >60 days; and (3) Severe colitis requiring infliximab ‘rescue therapy’. Endoscopies were re-analysed with image scoring using Mayo and ulcerative colitis Endoscopic Index of Severity (UCEIS) scores by two independent endoscopists. Histology was scored using the Nancy Index by two expert GI histopathologists.

Results

1074 patients were analysed. Twelve per cent (134) developed irAE colitis. Median patient age was 66, 59% were male. Age, sex, smoking status and prior IBD or autoimmune disease were not associated with risk of developing irAE colitis. CTCAE grading did not correlate with steroid duration (p = 0.92) or infliximab requirement (p = 0.18) (Figure 1). CRP, albumin and haemoglobin didn’t correlate with severity of irAE colitis (Figure 2).

22% of patients received rescue therapy with infliximab either as single or multiple doses, with 3 undergoing a colectomy (2.3%; two had ipilimumab and one had combination). The UCEIS (p = 0.008) and Mayo (p = 0.016) scores correlated with disease severity and infliximab requirement. Patients with higher Nancy scores (3/4) were more likely to require infliximab (p = 0.03) (Figure 3).