S720 Endoscopic Severity Score of Immune-Mediated Colitis Is More Effective in Guiding Medical Treatment Than Clinical Severity Grade

Introduction: Endoscopic scoring systems have not been established for immune-mediated colitis (IMC). Previous studies have shown benefits from early endoscopic evaluation, but the value of endoscopy compared to clinical assessment remains uncertain. This study aims to establish an endoscopic scoring system for IMC and explore its utility in predicting the need for selective immunosuppressive therapy (SIT, infliximab or vedolizumab) compared to clinical symptoms severity. Methods: This is a retrospective study from 14 centers worldwide including 674 patients with IMC who underwent endoscopic valuation. Ten total endoscopic features were recorded based on endoscopic reports and assigned one point each (erythema, edema, loss of vasculature, friability, erosions, exudate, any ulcers, large ulcers, deep ulcers, ≥ 2 ulcers). The scoring system was devised by measuring the specificity of a selected score cutoff in predicting the need for SIT based on clinical consensus from the study group. IBM SPSS Statistics 26 was used to calculate specificities, Pearson correlations, and generate ROC curves (Figure). Results: We divided the cohort to include a training set and a validation set. In the training set, an endoscopy score (ES) cut-off ≥4 has a specificity of 82.8% across all colitis grades and 96.4% among grade 1 colitis alone to predict SIT use. A cut-off ≥5 showed a specificity of 87.6% and 98.2% respectively. These specificities were comparable to those of the validation sets. In contrast, clinical colitis and diarrhea grading based on Common Terminology Criteria for Adverse Events (CTCAE) was poorly associated with future SIT use (specificities of 27.4% and 12.3% respectively). Moreover, this new scoring system with a cutoff of 4-5 had a numerically higher specificity to a Mayo Endoscopic Score (MES) of 3 when ulcer was a mandatory factor (85%-88.2% vs 74.6%). Early endoscopic evaluation in disease course was associated with early SIT use (p< 0.001, r=0.4084). (Table) Conclusion: This is the largest, multi-center study to devise an endoscopic scoring system highlighting the important value of endoscopy in guiding the management of IMC for the first time. The results demonstrated that an ES cutoff ≥4 can achieve a higher specificity in predicting SIT use than clinical symptom grading alone. This study supports early and thorough endoscopic evaluation for IMC and paves the way for future external validation of the described scoring system.Figure 1.: ROC curves Table 1. - Specificity of prediction for selective immunosuppressant therapy use (infliximab and/or vedolizumab) by using endoscopy score cutoff 4 and 5 All colitis grades Colitis CTCAE grade 1 only Training set(N=337) Validation set (CI)(N=337) Training set(N=84) Validation set (CI)(N=66) Specificity(score cutoff 4) 82.8% 74.5% (68.5%, 80.3%) 96.4% 91.1% (82.2%, 97.9%) Specificity(score cutoff 5) 87.6% 83.0% (77.8%, 88.3%) 98.2% 93.3% (85.1%, 100%) Specificity of prediction for selective immunosuppressant therapy use (infliximab and/or vedolizumab) by using CTCAE grade of diarrhea and colitis CTCAE grade Patients Specificity Colitis CTCAE grade 1 vs 2-5 All patients (N=666) 27.4% Diarrhea CTCAE grade 1 vs 2-5 All patients (n=619) 12.3% Colitis grade 1 only N=136 18.1% Colitis grade 2-5 only, N=477 10.3% CTCAE: Common Terminology Criteria for Adverse Events.