S3062 Clinico-Histologic Correlation of Small Bowel Ulcers: Differentiating NSAIDs, Crohn’s, CMUSE and CNSU

Introduction: Small bowel (SB) ulcers and strictures can stem from many underlying pathologies. Crohn’s disease (CD), nonsteroidal anti-inflammatory drug (NSAID) injury, and the emerging genetic disorders of prostaglandin production (cryptogenic multifocal ulcerous stenosing enteritis [CMUSE]) or reuptake (chronic nonspecific ulcers of the small intestine [CNSU]) have significant overlapping clinical and endoscopic features. Device-assisted enteroscopy (DAE) has allowed direct visualization and tissue sampling of ulcers deep in the small bowel. We aim to assess the correlation frequency between the clinical and histologic diagnoses of CD, NSAID, and CMUSE/CNSU SB ulcers. Methods: A retrospective cohort study at The University of Chicago (04/2004–04/2024) included adults with jejunal or ileal SB ulcers/strictures identified during DAE and recorded in ProVation (Minneapolis, MN). Patients with clinical diagnoses of CD, NSAID injury, or CMUSE/CNSU based on history, labs, imaging, and endoscopy were included. Exclusions applied to those with duodenal, terminal ileal, anastomotic ulcers, ulcers from other etiologies, or missing data (n = 38). Two gastroenterologists (DM, CS) and RM assigned diagnoses; a blinded expert pathologist (JH) performed a histologic review. Cohen’s kappa (κ) assessed agreement between clinical and histologic diagnoses. Results: Among 102 patients with SB ulcers/strictures, 41 met the inclusion criteria. Based on clinical consensus, 26 (63.5%) had NSAID injury, 12 (29.5%) had CD, and 3 (7%) were CMUSE/CNSU. Of 26 patients with a clinical diagnosis of NSAID injury, blinded histology review reported 23/26 (88%) as NSAID injury. Of 12 patients with a clinical diagnosis of CD, blinded histology review reported 5/12 (41%) as CD. Of 3 patients with a clinical diagnosis of CMUSE/CNSU, blinded histology review reported 2/3 (67%) as NSAID or CMUSE/CNSU and 1/3 (33%) as NSAID enteropathy. The correlation between a clinical diagnosis and histology was 28/41 (68%), corresponding to a k of 0.45, indicating moderate agreement. Of the 13 discordant cases, 6 (46%) remained uncertain after longitudinal follow-up for a median of 7 years. Of 7 discordant CD cases, 4 (57%) required multiple biologics and surgical resections, with 2/4 (50%) developing short bowel syndrome needing parenteral nutrition (PN). Conclusion: Differentiating the underlying cause of SB ulcers/strictures is challenging due to overlapping clinical and histologic features with only moderate agreement necessitating additional diagnostic tools.