Endoscopic mucosal resection of large colorectal adenomas: Only for expert centers?

Colorectal cancer represents one of the leading malignancies worldwide.¹ Early endoscopic detection and removal of its precursor lesions may prevent colon cancer-related death.² In order to provide timely and effective endotherapy, identification of lesions suitable for endoscopic removal, choice of proper resection technique and complete removal are all important factors of polypectomy success. Snare polypectomy, in flat lesions mostly with the help of prior submucosal saline injection (saline assisted polypectomy, also called endoscopic mucosal resection (EMR)) is currently considered the standard of care. Larger adenomas that may not be removed by ‘en bloc’ snare polypectomy require piecemeal EMR, while en-bloc resection of larger lesions by means of endoscopic submucosal dissection (ESD) is still considered experimental in the colorectum, due to its high complexity, lower success rates in the West, and also its higher rate of complications.^3–5 EMR may leave small remnants behind, plus it was recently shown that biopsies taken from the margins of even smaller polyps (<2 cm) after snare removal reveal adenomatous tissue in 10% of instances overall and in 17% of 1–2 cm polyps.⁶

Employing EMR, including re-treatment of adenoma remnants partially in combination with argon plasma coagulation (APC), the late recurrence rate may be as low as 1.5% as per literature summaries, with complication rates being <4%⁷; however, our current knowledge is based on retrospective case series and may therefore have inherent potential bias, because the number and quality of follow-up colonoscopies is mostly not known. Furthermore, recent prospective data show substantial recurrence rates of large colorectal adenomas: A study from two expert centers in Germany revealed that early remnants and late recurrences were 32% and 16%, respectively. Notably, biopsy evidence of residual or recurrent adenomas was found in 7% of polypectomy scars that were macroscopically normal.⁸ Another recent large series from Australia included 1000 successful colorectal EMRs: 80% had an early follow-up with a 16% rate of residual/recurrent adenomas, even if mostly minimal; and of the 670 cases with a normal first control colonoscopy, only 426 had a second follow-up within the study, revealing a 4% recurrence rate.⁹ The price for the high overall effectiveness of colorectal EMR (98% in the Australian study) is, therefore, the consequent performance of several follow-up colonoscopies within the first 2–3 years. The importance of follow-up, perhaps even as a quality indicator of colonoscopy with polypectomy, cannot be stressed enough. This is underlined by a recent large trial in Norway, which shows that colorectal cancer mortality is even increased after removal of advanced adenomas, but there the guidelines included the first follow-up after 10 years and compliance was not reported.¹⁰

Moreover, as some reports mainly come from specialized centers, these may not reflect the situation in units treating a substantially lower number of cases per year. In the December issue of United European Gastroenterology Journal, Barendse et al.¹¹ describe their findings following EMR of large rectal adenomas performed by endoscopists with lower case-volumes. In 19% of their patients, they found residual adenomatous tissue during sigmoidoscopy, 3 months after the initial EMR treatment. Furthermore, recurrence at the 6-, 9-, 12- and 24-month follow-up time period was eventually diagnosed in 25% of the cases. Both the rate of residual adenoma tissue at 3 months after EMR and the percentage of late recurrence differ from some, but not all, of the previous reports outlined above. A meta-analysis performed by the same group reveals there was remnant adenomatous tissue in 11.2% of cases and recurrence during further follow up in 1.5%.⁷ On the other hand even in expert centers, higher rates of remnant/residual adenomas may occur, as in the recent Australian study mentioned above.⁹ In this paper, increased lesion size, use of APC and intra-procedural bleeding were significant risk factors. Another study, by a single endoscopist at a tertiary center, reveals residual tissue following piece-meal EMR in 38% of cases, with polyp size being a significant risk factor for recurrence.¹² Notably, however, most patients in that study were treated for adenomas in the right colon, which potentially accounts for this phenomenon.

Nevertheless, comparison of those previous reports to the data of the current study reveals that despite limited experience, good clinical results may be achieved; even though later recurrences were somewhat higher than in the reports from expert centers. Unfortunately, data on the individual doctor’s prior training and yearly caseload were not provided. Nevertheless, as 64 patients were treated in 15 hospitals, the mean number of procedures per endoscopist during the study period may have been as little as four. The authors share this concern, by stating that some of the participating doctors “may have been early in their learning curve…”. Still, it would have been of interest to assess the relationship between an individual doctor’s experience and the clinical results.

All studies on EMR in the colon and rectum demonstrate that recurrence may happen; therefore, meticulous follow-up including retreatment is required, particularly as remnants are usually found to be unifocal and diminutive.⁹ Those may usually be ablated easily in almost all cases, leading to low late-recurrence rates, during further follow up.^7,9 The comparably high recurrence rates during further follow-up sessions, as found by Barendse et al.,¹¹ may be explained by insufficient imaging, poor bowel preparation or localization behind folds, as suggested by the authors. In addition, lower experience in recognizing and treating remnants or recurrences at the polypectomy scar may have also played a role. All these factors require careful inspection by the endoscopist, perhaps in some cases also the use of a distant cap, to examine the area behind folds. Lack of those measures in the current study may, therefore, be attributable to insufficient experience, as well.

Unfortunately, information on pre-therapeutic evaluation of lesions is missing in the majority of cases included in the current study: The Kudo pit pattern and Paris classification of the resected adenoma were not routinely assessed, or not described, in 88% of the lesions; however, such information may help to discriminate adenomas from cancers, and moreover, may predict the submucosal invasion depth.¹³ Cancers with limited submucosal invasion (sm1) may be cured by endotherapy, as lymphatic spread is rare in the sm1 situation; however, deeper submucosal invasion is accompanied by an increased risk for lymphatic metastasis. Therefore, those patients should not undergo endotherapy in the first place. Even though the accuracy of pre-interventional assessment may not be as high as desirable, particularly due to low overall specificity, it needs to be done and documented in both daily practice and studies alike. Barendse et al.¹¹ document 5% invasive cancers, though it remains unclear whether those have been sm1 tumors, were endoscopically-resected completely (R0) at the lateral margins and were histologically R0 at the basal margin. One these patients refused to undergo additional therapy and then neither displayed local recurrence nor metastasis. So did another two patients, who underwent subsequent total mesorectal excision (TME) after neo-adjuvant therapy. Both total mesorectal excision (TME) specimens showed no residual cancer.

Perforations and bleeding represent the main complications of EMR in the rectum. Perforations mark an infrequent event, ranging from none to 2% reported by Barendse et al.¹¹ Bleeding, on the other hand, may inevitably occur during the procedure, can usually be treated endoscopically; and therefore, may not be considered a complication per se. Thus, the definition of bleeding should be standardized and probably only applied to cases which require re-intervetion or other measures after colonoscopy, for bleeding. Delayed bleeding within hours or days following EMR have been previously reported at 8%, and may usually be managed by endoscopic clipping. Prophylactic placement of clips may prevent delayed bleeding, yet may not be technically feasible after wide field EMR. Whether the meticulous coagulation of vessels may substantially decrease delayed bleeding remains questionable. In the Barendse et al.¹¹ study, delayed bleeding occurred in 13 of 64 patients; however, seven of those received anticoagulant or antiplatelet therapy.

Conclusion

The current study revealed that EMR for larger rectum lesions may be done successfully and safely in units outside of tertiary referral centers. The lack of systematic, pre-therapeutic assessment and its documentation, the fairly high rates of adenomatous remnants and recurrences during further follow-up periods do, however, underline the need for structured training programs for medical doctors. Those need to address the proper choice of lesion suitable for endotherapy; meticulous resection techniques in order to eradicate the entire lesion, even in anatomically-challenging sites, and follow up under optimal conditions. Last but not least, further studies are required to address the choice of resection technique (EMR versus ESD) and use of additional ablation by APC in order to further minimize recurrence rates.