Impact of Inflammatory and Fibrotic Lesion Involvement on the Outcomes of Endoscopic Resection in Ulcerative Colitis

초록

Background: The prevalence of ulcerative colitis (UC) continues to rise worldwide, while biologic therapies have reduced colectomy rates, allowing long-term preservation of the inflamed colon. However, cumulative inflammation increases the risk of colitis-associated dysplasia and sporadic adenoma, highlighting the importance of endoscopic resection (ER) as a colon-preserving strategy. Current guidelines recommend ER for endoscopically resectable lesions, but residual inflammation or scarring can obscure margins and increase technical difficulty. This study evaluated the clinical outcomes and safety of ESD and EMR according to lesion involvement with surrounding mucosa in UC patients. Methods: We conducted a single-center retrospective cohort study including 379 ER procedures (ESD, n = 44; EMR, n = 335) performed in 273 UC patients at Samsung Medical Center between January 2012 and September 2025. Both dysplastic and cancerous lesions were included. Lesion involvement was categorized into four groups based on endoscopic findings: non-involved, endoscopic remission, active inflammation, and scarring. The primary outcome was the en bloc resection rate. Secondary outcomes included R0 resection, procedure-related adverse events, local and metachronous recurrence, and post-resection surgery. The median follow-up duration was 37 months (IQR 17–61). Results: En bloc resection was achieved in 331 of 379 cases (87.3%). En bloc resection rates differed by lesion involvement category: non-involved 92.6%, remission 89.2%, inflammation 91.9%, and scarring 79.4% (p = 0.008). Procedure-related adverse events occurred in four cases (1.1%; bleeding 2, perforation 2), all of which were managed endoscopically without surgical intervention. Metachronous recurrence was more frequent in the scarring group compared with the non-involved group (18.3% vs. 7.4%, p = 0.029). In multivariable logistic regression, scarring lesions independently predicted failure of en bloc resection (OR 0.413, 95% CI 0.178–0.955, p = 0.039), and non-polypoid morphology was also associated with lower en bloc resection rates (OR 0.135, 95% CI 0.056–0.321, p < 0.001). Conclusion: Scarring lesions and non-polypoid morphology were associated with significantly lower en bloc resection rates, and scarring was an independent predictor of en bloc resection failure. In contrast, ER could be performed safely in lesions with active inflammation or those in endoscopic remission, with a very low complication rate. These findings suggest that ER remains a feasible and safe treatment option even in the inflamed colon of UC patients, whereas scarring lesions require careful pre-procedural evaluation and meticulous intra-procedural technique due to increased technical difficulty and complication risk.