Management of large polyps in a colorectal cancer screening program with fecal immunochemical test: a community- and population-based observational study

Background and study aims The aim of this study was to analyze presentation, management, and outcomes of large polyps (LPs; ≥ 20 mm) detected in a colorectal cancer (CRC) screening program using a quantitative fecal immunochemical test (FIT). Patients and methods This was a retrospective community- and population-based observational study of all LPs detected in patients aged 50 to 74 years between 2015 and 2019 during FIT-positive colonoscopies within the screening program organized in Alsace (France). Results Among 13,633 FIT-positive colonoscopies, 1256 LPs (8.5 % malignant and 51.8 % nonpedunculated) were detected by 102 community gastroenterologists in 1164 patients (one in 12 colonoscopies). The sensitivity of optical diagnosis of malignancy was 54 % for nonpedunculated and 27 % for pedunculated T1 CRCs. The endoscopic resection rate was 82.7 % (95 % confidence interval [CI] 80.3–84.9) for benign LPs (70.2 % [95 % CI 66.4–74.1]) nonpedunculated, 95.2 % [95 % CI 93.4–97.1] pedunculated), varying from 0 to 100 % depending on the endoscopist. It was correlated with cecal intubation (Pearson r = 0.49, P < 0.01) and adenoma detection rates ( r = 0.25, P = 0.01). Most endoscopists did not refer patients to more experienced endoscopists, and as a result, 60 % to 90 % of 183 surgeries for benign LPs were unwarranted. Endoscopic resection was curative for 4.3 % (95 % CI 0.9–12.0) of nonpedunculated and 37.8 % (95 % CI 22.5–55.2) of pedunculated T1 CRCs. Overall, 22 endoscopic submucosal dissections had to be performed to avoid one surgery. Conclusions Compared with current recommendations, there is tremendous room for improvement in community endoscopy practices in the diagnosis and management of LPs. Detection and polypectomy competencies are correlated and highly variable among endoscopists. Endoscopic resection is curative for 83 % of benign LPs and 16 % of T1 CRCs.