Endoscopic Mucosal Resection for Colonic Mucosal Neoplasia and Evaluation of Long-Term Recurrence

A Single-Center Experience of 500 Cases

Saeed Ali, MD; Neelam Khetpal, MD; Evgeny Idrisov, MD; Asad Ur Rahman, MD; Sameen Khalid, MD; Yuan Du, MS; Udayakumar Navaneethan, MD; Shyam Varadarajulu, MD; Robert Hawes, MD; Muhammad Khalid Hasan, MD


South Med J. 2021;114(4):199-206. 

In This Article


Patient Characteristics

The study included a total of 480 patients (48% men; mean age 67.5 years; range 36–90) with 500 lesions. Complete endoscopic resection was unsuccessful in 27 lesions (5.4%; Figure 3). Histology was available for all of the lesions. The mean lesion size was 33.9 mm. Successful en-bloc resection was performed in 96 lesions and piecemeal resection in 377 lesions. At SC1, 354 post-EMR scars were examined. The remaining 146 lesions did not undergo SC1 (Figure 3). Hyperplastic polyps were treated as possible sessile serrated adenomas and underwent SC; however, pedunculated polyps were not subjected to SC and were excluded from the analysis.

Figure 3.

Recurrence flow diagram. EMR, endoscopic mucosal resection; SC1, first surveillance colonoscopy; SC2, second surveillance colonoscopy.

Lesion Characteristics

The most common site was ascending colon (161 lesions [32.2%]). Flat lesions (Paris classification 0–IIa), Kudo pit pattern IIIs, and granular surface morphology were the most prevalent. The most common histological subtype was tubulovillous adenoma (201 lesions [40.2%]). High-grade dysplasia was found in 55 lesions (11%). Cancer was found in 36 lesions (7.2%). Cancer was poorly differentiated in 13, well-differentiated in 12, moderately differentiated in 7, and undifferentiated in 4 cases.

Early Recurrence at SC1 following Successful EMR and the Ability to Treat it

A total of 354 lesions were examined at SC1. There was no evidence of RRA in 277 lesions (78.2%). Early recurrence was present in 77 (21.8%; Figure 3). Recurrent adenomas were treated endoscopically in 76 of 77 lesions (98.7%). One patient was referred to surgery because of advanced endoscopic appearance of the lesion.

Risk Factors for Early Recurrence

The risk factors associated with early recurrence at SC1 using univariate analysis are presented in Table 1. Lesion size, lesion location, lesion morphology, and lesion histology were significantly associated with recurrence using univariate analysis.

Independent predictors of recurrence were analyzed using multiple logistic regression analysis (Table 2) and are explained as follows:

  • Increased lesion size: lesion size >40 mm was significantly associated with increased recurrence at SC1 (OR 15.41, 95% CI 4.3–55.3, P < 0.0001; Table 2).

  • Lesion morphology: recurrence of lesions with mixed morphology (OR 2.423, 95% CI 0.978–6.002, P = 0.05) and nongranular morphology (OR 1.09, 95% CI 0.599–1.97, P = 0.26) was not significantly different than for lesions with granular morphology and therefore had no association with increased recurrence at SC1.

Late Recurrence at SC2 After Normal SC1

Of 354 post-EMR scars, 277 (78.2%) had no recurrence at SC1. Because of the nature of our practice, only 41 of 277 cases (15%) came for follow-up and were examined at SC2. Only 4 lesions (9.8%) had late recurrence, and all of them were successfully excised endoscopically.

Persisting RRA at SC2 Following Apparent Successful Endoscopic Management at SC1

A total of 76 lesions (98.7%) had complete endoscopic treatment of recurrence at SC1. Of these, 22 lesions were examined at SC2. RRA was found in 8 lesions (36.3%) and managed endoscopically in 6. Recurrence at SC1 was indicative of a high risk of recurrence at SC2.

Summary of Recurrent Adenoma Results

In total, recurrence occurred in 81 cases, including 77 early and 4 late recurrences. Endoscopic therapy succeeded in treating recurrence in 78 of 81 lesions (96.3%).

Complications during Initial EMR

Intraprocedural Complications. Intraprocedural bleeding occurred in 23 patients and was treated endoscopically by using snare tip soft coagulation in 15 patients, clips in 7 patients, and epinephrine injection in 1 patient. Immediate perforation (target sign) occurred in one patient, which was treated successfully by closing the defect endoscopically with clips.

Postprocedural Complications: Postprocedural (24 hours) complications occurred in 11 patients: delayed bleeding in 8 patients, with 7 requiring hospitalization, 5 of whom required repeat colonoscopy; abdominal pain in 2 patients, with pain resolving after overnight observation; and perforation in 1 patient that was not evident at the conclusion of the procedure. The latter patient subsequently had abdominal pain and distention and required surgery. The patient did well after surgery and was discharged without any further sequelae.

Adverse Events From the Treatment of RRA. There were no perforations or clinically significant bleeding episodes during the treatment of RRA at SC1 and SC2.