Removing large rectal polyps: when the whole may be greater than the sum of its parts

Rectal lesions containing dysplasia or early neoplasia confined to the mucosa are typically managed with a minimally invasive transanal approach. Two commonly used modalities for endoscopic removal are EMR and transanal endoscopic microsurgery (TEM). Barendse et al recently published a multicenter randomized trial (TREND study) comparing TEM with EMR for the resection of large nonpedunculated rectal polyps. The study consisted of 204 patients with rectal adenomas ≥3 cm randomized to EMR or TEM. Importantly, lesions with suspected submucosal invasion (SMI) based on white-light endoscopy, virtual chromoendoscopy, EUS, or a combination of these, were excluded. Overall, the recurrence rates were 15% after EMR and 11% after TEM without reaching statistical noninferiority for EMR. Although adverse event rates did not differ between the 2 techniques, the incidence of serious adverse events was significantly lower with EMR than with TEM (1% vs 8%; P = .06). On the basis of these findings, the authors postulated that EMR should be considered the primary method of choice because of a tendency toward lower adverse event rates and favorable cost profile in comparison with TEM.

In this issue of Gastrointestinal Endoscopy, Bronzwaer et al report a post-hoc analysis of the patients found to have unexpected rectal cancer detected by histopathologic analysis from the TREND study. The aim was to retrospectively compare the diagnostic assessment and procedural characteristics of lesions with and without unexpected SMI derived from their multicenter trial. In all, 27 patients (15 treated with EMR and 12 with TEM) received diagnoses of unexpected rectal cancer on histopathologic examination. Most of the unexpected cancers were T1 lesions (82%) with well to moderately differentiated histopathologic features (78%), and none had lymphovascular invasion. There were no differences in baseline patient characteristics between patients with unexpected rectal cancers and those with histologically proven rectal adenoma, except that rectal bleeding was more frequently reported in patients with cancer (P = .004). The mean size of the unexpected rectal cancers was 47.0 ± 11.8 mm. Most of the lesions were classified as Paris Is (21/27; 75%) whereas a Kudo pit pattern III-L or IV was observed in 15% or 30%. Overall, the authors did not detect any differences in lesion size, morphology (Paris classification), surface pattern (Kudo classification), EUS staging, or preprocedural histopathologic features of biopsy specimens between the rectal cancer and benign adenoma cohorts. From an EMR procedural standpoint, however, the success rate of submucosal lifting was significantly lower in the unexpected cancers than in the benign adenomas (60% vs 93%; P < .001). In comparison with benign adenomas, more unexpected rectal cancers were incompletely resected with EMR (33% vs 5%; P = .01) and terminated prematurely (60% vs. 8%; P = .001). In addition, more patients with unexpected rectal cancers treated with EMR had to undergo additional completion surgery than were those who had initial TEM (80% vs 42%; odds ratio 5.6%; 95% confidence interval, 1.02-30.9; P = .04). Overall survival was 100% at a mean follow-up time of 4.4 ± 1.2 years. On the basis of this post-hoc analysis of the TREND study, the authors concluded that there were no significant differences in preprocedural diagnostic characteristics distinguishing adenomas from rectal cancers. Conversely, poor submucosal lifting, incomplete resection, or early termination during EMR should raise the suspicion of SMI.

The current study demonstrated that lesion characteristics based on validated classification systems (Paris and Kudo) and EUS are less than ideal in distinguishing between benign and malignant rectal polyps. A few issues deserve consideration when these results are interpreted. For one, the sample size of patients with unexpected rectal cancers was relatively small. More importantly, there was no predetermined protocol for the diagnostic evaluation of rectal polyps, which introduced variability in how lesions were characterized based on the local expertise and availability of advanced imaging techniques (ie, virtual chromoendoscopy). Nevertheless, these findings further emphasize that the applicability of these classification systems outside of expert endoscopy centers may be limited and that structured training in advanced imaging techniques should be implemented to help increase the overall diagnostic accuracy of lesions with SMI. In addition, index histopathologic analysis based on mucosal biopsy specimens did not differ between the unexpected cancers and the benign adenomas in this study, which dispels the utility of preresection biopsies and underscores the risk of understaging with this maneuver alone.

Adequate lesion selection is crucial to determine the optimum therapy and to avoid unnecessary noncurative interventions. This post-hoc analysis of rectal cancers detected unexpectedly in large nonpedunculated rectal polyps highlights some of the current limitations of diagnostic lesion characterization. These findings are not isolated, and mounting evidence has accumulated that even in the best of hands our preprocedural evaluation is far from perfect.^,  In a large prospective cohort study of 2277 lesions, covert SMI was present in 7.6%. The factors associated with covert invasive cancer were rectosigmoid location, increasing size, combined Paris classification with Is component, and surface morphology.

Hence, if diagnostic characteristics may not reliably help discern SMI in large nonpedunculated rectal polyps, should piecemeal EMR be routinely advocated as the first-line method? The elephant in the room is this: how does endoscopic submucosal dissection (ESD) fit into the array of available therapies for large rectal polyps? The main benefit that ESD brings to the table is the ability to provide en-bloc resection. This in turn can allow for accurate histopathologic evaluation and provide cure in cases with favorable histologic criteria, thus avoiding need for additional surgery. We here consider factors that favor ESD with en-bloc resection in the management of selected large nonpedunculated rectal polyps:

1. Resected fragmented specimens with piecemeal EMR prevent accurate histopathologic analysis of depth of invasion and adequacy of resection margins, which increases the risk of understaging a lesion. Indeed, this study by Bronzwaer et al demonstrated that 3 out of 12 (25%) cancers initially treated with EMR were upgraded in tumor stage after completion surgery. Given that reliable pathologic staging is not feasible without en-bloc resection, it is conceivable that the true incidence of colorectal lesions with SMI is underestimated in the EMR literature. This is a key issue with profound clinical implications, which needs to be further studied because it may influence our strategy for the management of these lesions.

2. Piecemeal EMR of potentially curable lesions still subjects the patient to unwarranted surgery. In this study, more patients with rectal cancer who underwent EMR (80%) required completion surgery than did patients who underwent TEM (42%) (P = .04). For lesions with SMI, surgery can still be avoided, and endoscopic resection is considered curative if the depth of invasion is <1000 μm below the muscularis propria (SM1), the histologic features are well to moderately differentiated, and there is absence of lymphovascular involvement and tumor budding. The majority of the lesions in this study were low-risk T1 cancers (78%). It is possible that many of these rectal cancers could have been adequately managed with ESD, although the true proportion cannot be determined from this study because the SMI depth on the pathologic specimen was not specified in all cases.

3. En-bloc resection with ESD is safe and effective for lesions in the rectum. When compared with ESD of lesions in the proximal colon, rectal ESD is technically less challenging, given the ease of access, endoscope maneuverability devoid of paradoxical movement, and a wider operative field with fewer angulations. Given the advantages of ESD over EMR in terms of higher en-bloc, curative, and lower recurrence rates, both the Japan Gastroenterological Endoscopy Society and the European Society of Gastrointestinal Endoscopy recommend ESD for large (≥20-30 mm) rectal lesions, those with suspected SMI, or both.^,  Although comparative data between ESD and TEM are limited, ESD may be preferred over TEM because it is less invasive and has a better cost-effective ratio.

In summary, distinguishing benign adenomas from lesions with SMI remains a common challenge in clinical practice. EMR remains the mainstay resection technique for most colonic lesions. The risk/benefit equations seem to shift in the rectum, where en-bloc removal of large nonpedunculated rectal polyps should be favored over piecemeal resection, particularly in lesions with possible SMI, such as those with poor lifting, larger size, Paris Is component, and nongranular laterally spreading tumor morphology. The ideal approach to the management of any particular colorectal lesion should be one that is cost effective, is associated with a high curative rate, has an acceptable risk profile, minimizes the need for recurrent interventions, and ultimately avoids surgery. It is clear that no single resection technique will meet all of these criteria; therefore, the most suitable strategy for any given case needs to be individualized. EMR, ESD, and TEM are all viable options based on lesion specifics and local expertise; yet, further advancements in endoscopic imaging are needed to improve accurate lesion characterization and allow the selection of the most appropriate management strategy.