Effectiveness of endoscopic resection of colonic lesions > 20mm

Abstract

Background and aims Endoscopic removal of large colonic lesions (>20mm) is technically more demanding and should be attempted by experienced endoscopists. Furthermore, larger lesions have a higher risk of invasive carcinoma. In cases suspicious for submucosal invasion, en bloc resection is mandatory to ensure a "curative" endoscopic resection. En bloc resection of larger lesions with endoscopic mucosal resection (EMR) or polipectomy is often unfeasible. An alternative method is endoscopic submucosal dissection (ESD), which is associated with longer learning curves, longer procedure times and higher complication rates. Our aim is to evaluate the effectiveness and safety of endoscopic resection techniques for the treatment of colonic lesions >20mm. Methods A retrospective analysis was conducted in a tertiary center. All colonic lesions >20mm removed between January 2013 and December 2017 were selected from a colonoscopy database. Patients with previous colorectal malignancy or surgery were excluded. Data was collected from clinical records and endoscopic and histopathological reports. Results We included 370 colorectal lesions >20mm removed endoscopically: 97 (26%) pedunculated (Ip) and 273 (74%) were non-pedunculated. Median lesion size was 25 mm (range 20-130 mm). Most lesions (n=239, 65%) were in the left colon and 343 (93%) were conventional adenomas. Most non-pedunculated lesions were removed by EMR (n=259, 95%) and 14 (5%) by ESD. Complication rate was 3% (9 cases of delayed bleeding, all managed conservatively/endoscopically, and 1 case of delayed perforation after ESD, with need of surgery). Endoscopic resection was effective in 98% (n=95) of standard polypectomies, 88% (n=228) of EMRs and 64% (n=9) of ESDs. There were 11 cases (3%) on incomplete resection: 1 (1%) after standard polipectomy, 9 (3%) EMR procedures and 1 (7%) ESD procedure. Of the completely removed lesions, 38 (10%) had submucosal invasion and 27 (7%) had high risk criteria. Of the high-risk malignant lesions, in 5 (1%) the only criterium was indeterminate margins: 1 fragmented standard polypectomy, 3 fragmented EMRs and 1 inadvertently fragmented ESD. Recurrence rate for fragmented EMR was 23% (n=36). There was a tendency for higher recurrence rate when ablation of the mucosectomy defect was performed (29 vs 18%, p=0.09). Ablation of the mucosectomy defect was also associated with higher risk of late bleeding (6 vs 1%, p=0.04). Discussion/Conclusion In our series, EMR effectively treated 88% of non-pedunculated colorectal lesions (and 91% of those with complete resection) with acceptable complication and recurrence rates; only 3 lesions would have benefited from ESD (ie, 1 in 86). Of the 273 flat or sessile lesions removed, only 4 (1%) were low-risk malignant polyps and would mandate en bloc resection for endoscopic cure. EMR is a simpler, faster and safer procedure that appears appropriate for the vast majority of colorectal lesions. Careful lesion inspection and characterization is mandatory to select those which benefit from en bloc resection with ESD.