Our colonoscopy cohort included 331 consecutive patients who received endoscopic treatment for colorectal polyps larger than 20 mm from May 2005 to November 2011 at Yonsei University College of Medicine, Seoul, Korea. Patients were included if the following polyp criteria were met: 1) sessile polyp (defined as a lesion in which the base is attached to the colon wall) or flat polyp (defined as a lesion with a thickness less than half of the maximum width); 2) equal to or greater than 20 mm in size; and 3) adenoma, carcinoma in situ, or intramucosal cancer indicated in the final pathological report following endoscopic resection.16 We excluded patients with colorectal tumors with stalks, colorectal cancers, carcinoids, a non-neoplastic histology, or patients without a follow-up colonoscopy. Among the 331 consecutive patients, a total of 127 patients were excluded for the following reasons: pedunculated type (n=28), colorectal cancer (n=73), carcinoid (n=5), non-neoplastic histology (n=4), and no record of follow-up colonoscopy (n=17). Finally, 204 patients with sessile and flat-type colorectal polyps larger than 20 mm were included in the study (Fig. 1). Clinical, endoscopic, and demographic data were extracted from computerized clinical information system and reviewed retrospectively. The study was approved by the Institutional Review Board of our hospital. All study procedures were conducted in accordance with the International Conference on Harmonisation Good Clinical Practices and the Declaration of Helsinki and its amendments.

Colonoscopies were performed after bowel preparation with 4 L polyethylene glycol solution (Colyte; Taejun, Seoul, Korea or Colyte-F or Colonlyte; Dreampharma, Seoul, Korea), by using a single-channel high definition colonoscope (CF Q240L, CF Q240I, CF H260AI, CF Q260AI, or PCF Q260AI; Olympus Optical Co., Tokyo, Japan) and a high-frequency generator with an automatically controlled system (VIO300; ERBE Elektromedizin GmbH, Tubingen, Germany). A translucent cap (D-201-13404 or D-201-14304; Olympus, Optical Co., Ltd., Tokyo, Japan) was mounted on the tip of the colonoscope to control the depth of the mucosal incision and to maintain a satisfactory view during the submucosal dissection.17 Procedures were performed with a patient in a left lateral position under monitored anesthesia sedation. All procedures were performed by 4 experienced gastroenterologists. The endoscopic treatments, including EMR or endoscopic submucosal dissection (ESD), were decided according to the characteristics of the lesion and the preference of operators. EMR was performed using a snare after injecting a solution (normal saline, indigo carmine 0.04%, with or without 1:100000 dilution of epinephrine) into the submucosa.18 Sequential submucosal injection and snare polypectomies were performed for tumor remnants.11 For EMR or snare polypectomies, RotaSnare^® (Oval shape, 15/25/35 mm sized; manufactured by Medi-Globe GmbH) was used. ESD was performed with a flex knife (Flex Knife; KD-630L, Olympus, Optical Co., Ltd.) or a dual knife (Dual Knife; KD-650L, Olympus, Optical Co., Ltd.). After spraying and injecting 0.4% indigo carmine dye solution into the submucosa beneath the lesion, a circumferential incision was made with a flex knife set to approximately 2 mm in length. After an additional injection beneath the lesion, the submucosal layer was dissected directly using the flex knife set to approximately 1 mm in length or the dual knife19 to lift it away from the muscularis propria sufficiently. If a residual lesion was detected endoscopically, additional EMR was performed. Resections were defined as follows; en bloc resection was a single piece resection and piecemeal resection was a multiple fragment resection.20 We did not routinely apply argon plasma coagulation or hot biopsy when the post-procedure site was clean. We applied hot biopsy only when the remnant lesion was suspected endoscopically.

Specimens were collected using a basket or by aspiration into the suction channel. For the retrieval of entire sessile polyps located in the right colon, the colonoscope was withdrawn and reinserted as many times as necessary. All specimens were cut into 2-2.5 mm slices and examined microscopically for differentiation, lymphatic invasion, vascular involvement, depth of invasion, the lateral (mucosal) resection margin, and the basal (submucosal) resection margin. The extension of tumor cells to the resected margin was evaluated and graded as complete resection when the lateral and basal resection margins were free of tumor (en bloc resection is essential), incomplete resection when the tumor extended into the lateral or basal margins, or not evaluable when the margins were not evaluable as a result of the artificial effects of coagulation necrosis or multipiece resection.21 We defined negative margin involvement as instances where both lateral mucosal margins of a specimen were free of neoplastic cells. All tumors had negative basal margins. Furthermore, 2 grades of dysplasia (high and low) were recognized. The terms "arcinoma in situ" and "intramucosal adenocarcinoma" were both described as high-grade dysplasia.22

Right-sided colon tumors were defined as those arising from the cecum to the transverse colon. Left-sided colon tumors were defined as those arising from the splenic flexure down to and including the rectosigmoid junction. Rectal tumors were defined as those arising distal to the rectosigmoid junction down to the anus (excluding squamous cell carcinoma).2324

Postprocedural bleeding (PPB) was divided into immediate PPB (IPPB) and delayed (DPPB). IPPB was defined as bleeding observed immediately after polypectomy and required hemostatic procedures because the bleeding continued for over 60 s. DPPB was defined as bleeding that occurred at the polypectomy site within 30 days of the procedure and required hospitalization or treatment.25 Other delayed adverse events of each patients were collected at the out-paitent clinics.

Patients were followed-up with colonoscopy to evaluate recurrence and residual tumor presence. The interval of surveillance colonoscopy was decided according to the previous guidelines26 and recommendations of our institution. Prior to 2009, however, surveillance colonoscopy was usually recommended after a short interval of less than 12 months. After 2009, surveillance colonoscopy was recommended 12 months after initial treatment, even in cases of piecemeal resection or ESD. If a polyp was detected in follow-up examinations, it was resected if possible. Local recurrence was defined as the presence of adenomatous or carcinomatous tissue on follow-up, confirmed by histology, at the site of prior endoscopic treatment.27 When the local recurrence was suspected, endoscopical mucosal resection was initially tried. If endoscopical treatment was not sufficient, additional surgical treatment was applied.

Descriptive statistics are provided for the binary and continuous variables using the incidence frequency (%) and mean±standard deviation [or median (and range) values]. The chi-square test and Fisher's exact test were used to compare binary variables, and the two-sample t-test was used to compare continuous variables. Multivariate logistic regression analysis, adjusted by confounding factors, was performed to discern the contribution of variable factors to inter-group differences in local recurrence (no recurrence vs. recurrence). Two-sided p-values were calculated and significance was accepted at the 5% level. All of the statistical analyses were performed with Predictive Analytics SoftWare for Windows, version 18.0.0 (SPSS Inc., Chicago, IL, USA).

The baseline clinical, endoscopic, and pathological characteristics are summarized in Table 1. Among the 204 patients, there were 128 (62.7%) men and the mean age was 65.1 years. The mean follow-up duration was 44.2±29.5 months. There were 144 flat (70.6%) and 60 sessile polyps (29.4%). The median tumor size was 25 mm. Among the 204 patients, 194 (95.1%) patients were treated with EMR and 10 patients (4.9%) were treated with ESD. The en bloc resection rate was 62.3%, and 77 patients (37.7%) received piecemeal resections. A negative resection margin was observed in 167 cases (81.9%), and 37 cases (18.1%) had a positive lateral margin. Low-grade dysplasia was observed in 100 cases (49%), and high-grade dysplasia was observed in 104 cases (51%). IPPB occurred in 65 patients (31.9%); 59 cases of these occurred during EMR and 6 occurred during ESD. All bleeding was successfully treated using endoscopic hemostasis. There was no case of DPPB. Perforations developed in 4 patients (2.0%); all of these patients were in the EMR group and were successfully treated with clipping.

The median surveillance colonoscopy interval was 9.4 months (1-66 months). A short interval surveillance colonoscopy, with a median of 6.3 months and range of 1-11 months, occurred in 110 patients (53.9%), and 94 patients (46.1%) received a long interval surveillance colonoscopy with a median of 13.6 months and range of 12-66 months. Piecemeal resection patients tended to receive short interval colonoscopies compared to en bloc resection patients (66.2% vs. 46.5%, respectively; p=0.006). There were no significant differences in the colonoscopy interval for patients with different treatment methods (p=0.112), margin statuses (p=0.150), or histologies (p=0.068).

During the surveillance colonoscopy, 14 local recurrences (6.9%) were detected. The mean follow-up duration was 47.6± 25.3 months in 14 patients with recurrence and 43.9±29.9 months in 190 patients without recurrence (p=0.606). Further details of the patients with local recurrence are given in Table 2. Of the local recurrence patients, 7 were located on the right side of the colon. There were 4 patients with sessile-type and 10 patients with flat-type polyps. Three patients developed local recurrences even after an initial en bloc resection with negative resection margin. Mean time to recurrence was 28.4 months (22.7 months with range of 4-54 months in short surveillance interval group vs. 42.8 months with range of 12-82 months in long surveillance interval group, p=0.240). Mean size of recurrent lesion was 14.6 mm (11.7 mm in short surveillance interval group vs. 21.8 mm in long surveillance interval group, p=0.054). There were 11 patients who were successfully treated with a second endoscopic resection, and 3 patients received surgery.

The potential risk factors for local recurrence are shown in Table 3. In addition, we also assessed the effect of the surveillance colonoscopy interval on local recurrence. Using multivariate analysis, a polyp size greater than 40 mm was shown to be a significant risk factor for the detection of local recurrence at surveillance colonoscopy [40-50 mm vs. 20-30 mm, odds ratio (OR) 14.22 with 95% confidence interval (CI) 2.10-96.17; p=0.006; >50 mm vs. 20-30 mm, OR 24.25 with 95% CI 3.32-176.88; p=0.002]. However, location, tumor morphology, the resection method, resection status, margin status, and tumor grade were not associated with local recurrence. Furthermore, the surveillance colonoscopy interval had no effect on the rate of local recurrence.