Endoscopic mucosal resection (EMR) has become the standard technique for resection of large sessile and flat colonic polyps. We aimed to assess the clinical outcome of colonic EMR of polyps ≥2 cm at the University Hospital of Leicester NHS Trust, Leicester, UK and to assess changes over a 9-year period. Data was collected for all sessile colonic polyps ≥20 mm removed by EMR between 2006 and 2014 by three endoscopists (Peter Wurm, John DeCaestecker, Richard Robinson). Patient demographics, resection technique, completeness of initial resection, recurrence rate at first surveillance endoscopy, polyp eradication at second surveillance endoscopy, and complication rates were analysed. A total of 569 lesions in 564 patients were assessed for EMR, among which there were 424 completed EMRs (bowel cancer screening programme [BCSP]: 138, symptomatic: 286) by three operators. Of the 145 patients who did not complete EMR, 70 were not attempted and instead were referred for surgical resection (cancer: 36, technical difficulty: 34). In a further 32 patients, EMR was attempted but abandoned; all were referred for surgery (cancer: 19, benign polyp: 13). Finally, 43 patients had no intervention (declined: 13, inflammatory polyps: 22, moved away: 8). The median age was 68.7 years (interquartile range [IQR]: 24–70), there were 226 males (53%), and 198 females (47%). The median polyp size was 30 mm (IQR: 25–40). The site of polyp was 27% in the right colon, 5% in the transverse colon, and 68% in the left colon (rectum: 58%, sigmoid: 4%, descending: 6%). Piecemeal EMR was carried out in 381 patients (90%) and en bloc in 43 (10%). Complete snare resection was the primary objective in each case. Minor residual lesions not amenable to snare excision were subjected to argon plasma coagulation or hot biopsies. Argon plasma coagulation therapy was not used empirically to prophylactically treat the margin in cases in which complete excision was considered to have been achieved. Of those who have undergone surveillance so far, recurrence was found in 56 out of 328 patients (17%, 95% confidence interval [CI]: 7.2–21%) at initial surveillance colonoscopy (median: 7 months, IQR: 4–8) and was endoscopically treated in 53 out of 56 (94.6%, 95% CI: 89.1–100.8%); 3 of 56 (5.4%) were referred for surgical resection (cancer: 2, benign polyp: 1). The two patients with cancer had tubulovillous adenoma with high grade dysplasia and underwent surveillance within 3 months post-EMR as recommended by the British Society of Gastroenterology (BSG). There was no metastasis found on the computed tomography (CT) staging or positron emission tomography (PET) CT scan. Complete eradication at second surveillance endoscopy (median 16 months, IQR: 8–18) occurred in 266 out of 289 patients (92.6%, 95% CI: 86.9–94.3%) with recurrence in 23 (8%, 95% CI: 5.8–18.0%), but in 22 of 23 this was endoscopically resected. One patient was referred for surgical resection (benign polyp). The overall complication rate was 34 out of 424 patients (8%), with immediate perforation in 1 out of 424 (0.2%) post-caecal EMR who required conservative medical treatment; post 14 out of 424 (3.3%) had polypectomy pain syndrome and required admission for overnight conservative medical treatment. Immediate bleeding during the procedure occurred in 17 out of 424 patients (4%) and was treated with soft tips coagulation or coagulation grasper. Delayed bleeding in 2 out of 424 patients (0.5%) required endoscopic therapy to achieve haemostasis. There were no patients with immediate or delayed bleeding post polypectomy requiring a blood transfusion. There were no procedure-related deaths. For each 3-year period (2006–8, 2009–11, 2012–14), there was a consistent reduction in the number of polyps not treated endoscopically that required surgery (overall decrease of 18%), and recurrence rate at first endoscopy surveillance (overall decrease of 16.3%). There were increases in the number of EMRs performed annually (overall increase of 26%), mean polyp size resected (overall increase of 7 mm), Level 3 and 4 polypectomies (overall increase of 11%), and complete eradication rate at first surveillance endoscopy (overall increase of 16.3%).
CONCLUSION
This is a large single-centre series of EMR of 424 sessile colonic polyps ≥2 cm performed by three operators over a 9-year period; 17% had recurrence at initial surveillance, most were managed endoscopically, with an eradication rate at second surveillance endoscopy of over 92%. Examination of time trends over this period showed progressive reduction in recurrence and a trend for larger, more complex polyps to be resected endoscopically, with a corresponding drop in surgical management, demonstrating an improvement in outcome with time. There were limitations to our study. This is a single-centre retrospective study with inherent bias, not limited by specific exclusion criteria. While the results for 16-month post-EMR for all the patients followed-up to date were encouraging, it would be ideal to have follow-up data for 3–5 years and to compare it with our National Cancer Database to further assess the long-term outcome of EMR. Discussion from the audience included the optimal endoscopic management to treat recurrence which has no formal guidelines from international gastrointestinal endoscopy societies. Routine closure with endo-clips post-EMR or prophylactic treatment for visible non-bleeding vessels post-EMR have not been shown to be beneficial. This would be based on a case-by-case basis depending on whether patients were required to restart anti-coagulation. All patients with high-risk polyps, including those with a villous component, need to have a stringent timing for endoscopic follow-up.
