S772 Endoscopic Mucosal Resection for Colorectal Dysplasia in Inflammatory Bowel Disease

Abstract

Introduction: Inflammatory bowel disease (IBD) is associated with an increased risk of colorectal cancer due to underlying chronic inflammation. Frequent surveillance colonoscopy is essential for early identification and removal of dysplastic lesions in this group. Many dysplastic lesions in IBD require advanced endoscopic resection due to size, morphology, or indistinct margins. The effectiveness of endoscopic mucosal resection (EMR) for laterally spreading lesions is well documented in the general population; however, EMR continues to be challenging in patients with IBD. We aim to characterize outcomes of EMR for IBD-associated lesions compared to controls. Methods: Retrospective data was collected from 2016 to 2021. CPT code 45390 was used to identify colonoscopies with EMR. ICD-10 codes K50-52 were used to identify patients with IBD. Adenomatous and dysplastic lesions of all sizes that were removed with EMR and underwent endoscopic surveillance were included. Data was recorded and stored in a REDCap database and analyzed using a mixed effects model for logistic regression with SAS Statistical Software. Results: A total of 23 lesions in 14 patients with IBD and 187 lesions in 141 patients without IBD were included. Gender, ethnicity, age, lesion size and location, procedure date and duration, and number of lesions per patient were similar among both groups (Table). Piecemeal resection (as opposed to en bloc resection) was more common in the IBD group (91.3% vs. 42.1%, P< .001). Surgical resection was required for 2 IBD-associated lesions and 7 controls (8.7% vs. 3.9%, P=.27). Recurrence was detected and confirmed by histology in 7 IBD-associated lesions and 39 controls. In a mixed effects logistic regression model (Figure), IBD and lesion size were independently associated with recurrence (OR=3.08, 95% CI 1.04-9.13, P=.04; OR=1.06 per mm increase, 95% CI 1.02-1.09, P=.001). For each year that elapsed from the study start date to the procedure date, odds of recurrence decreased by 28% (OR=0.72, 95% CI 0.58-0.90, P=.004). In this model, 38.4% of IBD-associated lesions (95% CI 18.6-58.1) compared to 19.1% of controls (95% CI 13.8-24.5) recurred after EMR. Conclusion: Recurrence of dysplastic lesions after EMR is more common in patients with IBD compared to controls. Most IBD-associated lesions are resected piecemeal. EMR in patients with IBD should be performed in expert centers with close endoscopic surveillance.Figure 1.: Mixed effects logistic regression model estimating odds of recurrence after EMR of adenomatous and dysplastic colorectal lesions. Diagnosis of IBD, lesion size, year of procedure, and degree of dysplasia were incorporated in the model. OR = 3.08 with a diagnosis of IBD (95% CI 1.04-9.13, p=.04). OR = 1.06 with each millimeter increase in lesion diameter (95% CI 1.02-1.09, p=.001). OR = 0.72 with each year from the study start date that elapsed prior to EMR (95% CI 0.58-0.90, P=.004). Table 1. - Characteristics of adenomatous and dysplastic colorectal lesions resected with EMR in patients with IBD compared to controls Total n = 210 Control n = 187 IBD n = 23 P-value Lesion size (mm): mean (SD) 21.7 (12.2) 21.5 (12.4) 22.9 (11.3) .62 Degree of dysplasia: n (%) .81 adenoma/low-grade dysplasia 191 (91.0) 170 (90.9) 21 (91.3) high-grade dysplasia 14 (6.7) 12 (6.4) 2 (8.7) invasive carcinoma 5 (2.4) 5 (2.7) 0 (0.0) Location: n (%) .36 ascending colon 69 (35.8) 61 (35.3) 8 (40.0) transverse colon 44 (22.8) 40 (23.1) 4 (20.0) cecum 29 (15.0) 28 (16.2) 1 (5.0) sigmoid colon 19 (9.8) 18 (10.4) 1 (5.0) descending colon 14 (7.3) 12 (6.9) 2 (10.0) rectum 11 (5.7) 9 (5.2) 2 (10.0) ileocecal valve 7 (3.6) 5 (2.9) 2 (10.0) Procedure time (min): mean (SD) 74.1 (29.9) 73.3 (30.4) 80.6 (25.5) .27 En-bloc vs piecemeal resection: n (%) < .001 en-bloc 108 (52.4) 106 (57.9) 2 (8.7) piecemeal 98 (47.6) 77 (42.1) 21 (91.3) Complications: n (%) 11 (5.2) 9 (4.8) 2 (8.7) .34