Epithelial lesions in the lower rectum involving the dentate line pose a particular challenge for the endoscopist, and several resection techniques have been suggested for their treatment. We have presented the findings of a bicentral, retrospective study comparing ESD and EMR for the resection of these lesions. To the best of our knowledge, this is the first study to directly compare the two methods. Our data show that ESD can achieve an en bloc resection of the tumors in most cases and is associated with a significantly lower local recurrence rate and a high curative resection rate for locally resectable, low-risk adenocarcinomas, while having a similar safety profile to that of EMR.

The anatomy of the lower rectum and the anal canal is the main reason that interventions in this area are technically difficult. The narrow anal canal with the contracted anal sphincter is followed by the abrupt widening of the rectal ampulla, forming an almost 90° angle. This geometry restricts the view and maneuverability of the endoscope, and some parts of the anorectal junction can only be reached in the retroflex position, which often restricts the function of endoscopic instruments [12]. The presence of the hemorrhoidal veins also increases the risk of intraprocedural and postprocedural bleedings and the complexity of the resection [13]. The direct drainage of these veins to the systemic circulation may also increase the risk of periprocedural bacteremia [3, 8]. Apart from that, the sensory nerve supply of the anal canal is somatic, making procedures in this area painful and thus requiring periprocedural and postprocedural analgesia [3]. Additionally, lesions on the dentate line have been associated with increased risk of submucosal fibrosis, restricting the lifting and visualization of the submucosal plane [14].

EMR is the most commonly used endoscopic resection technique for large colorectal lesions and has been associated with high technical success rates and an extremely favorable safety profile [15]. However, the restricted view and maneuverability of the endoscope in the lower rectum increases the difficulty of this technique even in experienced hands [3]. In a forward view, the almost 90° angle of the rectal wall at the anorectal junction requires an extreme angulation of the endoscope to enable the snare to be pressed against the tissue. The retroflex position offers a better visualization, but the tip of the endoscope is perpendicular to the wall of the lower rectum, which also hampers the positioning of the snare parallel to the lesion. The combination of these factors makes the precise placement of the snare more difficult than in other areas of the colorectum, thus increasing the risk of incomplete resections and injuries of the hemorrhoidal veins. Intraprocedural bleedings from the hemorrhoidal plexus are rarely clinically relevant, but they can quickly reduce the endoscopic view, thus further increasing the risk of incomplete resection. The precision of ESD helps overcome these problems, since the resection margins are carefully marked and followed during the epithelial incision. The reduced overview of the lesion, which is extremely problematic during EMR, is not a restricting factor for ESD because of the close proximity of the endoscope to the tissue required. The initial incision in the anal canal is kept shallow so as not to affect the underlying hemorrhoidal veins, which can then be carefully dissected, coagulated, and transected, thus effectively reducing the risk of intraprocedural bleeding [16]. After the establishment of an initial pocket into the submucosal plane, it is also easier to maintain a stable position of the endoscope and achieve adequate exposure of the submucosal tissue. These technical aspects lead to the increased en bloc and compete resection rates depicted in our data, which are also in accordance with the current literature [12, 17].

Before ESD became popular, TEM was commonly used as an alternative to EMR for the treatment of epithelial lesions in the lower rectum, especially in the suspicion of malignancy. This technique allows for a minimally invasive and precise resection of rectal lesions with the use of a 4cm wide rectoscope and rigid instruments and has been associated with complete resection rates of 83–87% and low recurrence rates [18,19,20,21]. Although mucosectomy and hybrid TEM-ESD have been described, most authors advocate for full-thickness resections when performing TEM [22,23,24,25]. However, TEM is technically challenging in the lower rectum and especially in the anal canal because of the instability of the rectoscope and must often be combined with transanal excision, which is known to increase the risk of incomplete resections [26,27,28]. Additionally, full-thickness resection is an overtreatment since a submucosal dissection is enough for the complete resection of all benign and early malignant lesions without deep submucosal invasion. On the contrary, full-thickness resection has been associated with increased scarring in the mesorectum, which can compromise the dissection plane and the outcomes of completion surgery in case of occult high-risk carcinomas [29, 30]. Finally, the insertion of the TEM rectoscope usually requires muscle relaxation and therefore TEM is mostly performed under general anesthesia [26, 31]. The introduction of ESD offered a new method with all the advantages of TEM regarding complete resections and low recurrence rates, while offering better access to difficult localizations like the anal canal, leaving the mesorectum intact and being performed in the endoscopy suite under sedation, thus leading to ESD gradually replacing TEM for most types of lesions in many centers.

The main advantage of ESD, as demonstrated by our findings, is the significantly lower local recurrence rate in comparison to EMR. Although no other study to our knowledge has directly compared these two techniques for lesions involving the dentate line, the single-arm retrospective studies that are available have shown similar recurrence rates to ours for both techniques, thus confirming our findings [2, 3, 9, 17]. As previously mentioned, ESD has an intrinsically higher en bloc resection rate, which is directly associated with a lower recurrence rate, in all parts of the gastrointestinal tract, but this effect is further enhanced in the lower rectum [32]. Additionally, the en bloc resection rate and recurrence rate of ESD are not directly affected by the size of the lesion, as in the case of EMR, thus further highlighting the advantages of ESD for the treatment of large tumors [2, 12, 33]. New developments in EMR, and particularly the addition of thermal ablation of the resection margins, have significantly reduced its recurrence rate, which nevertheless remains at 16% for lesions in the anorectal junction [15]. It must be pointed out that most recurrencies can be treated endoscopically, but this requires the patient’s commitment to an intensive follow-up program in order to detect the recurrence in time [15]. Additionally, the technical challenges of the anorectal junction increase even more in the presence of fibrotic tissue and the absence of lifting usually found in the site of the previous resection. Similar to previous studies, our data show that most of the patients with recurrent adenomas after EMR required multiple further endoscopic procedures and often transanal surgery to effectively treat these lesions, thus increasing the overall treatment costs and reducing patient satisfaction [2].

A subgroup of patients that will particularly benefit from ESD are those with low-risk early rectal cancer, since the resection can be curative and offers a pristine specimen, allowing for the accurate staging and evaluation of the resection status [4, 34, 35]. Epithelial lesions involving the dentate line are reported to have an increased incidence of malignancy with submucosal invasion, which highlights the importance of en bloc and complete resection of those lesions [36]. In our study, almost all low-risk T1 tumors were able to be completely resected, thus sparing the patient a major, amputating surgical procedure with high morbidity and the need for a permanent ostomy. Even in case of a lateral positive margin, careful marking of the specimen can point out the exact area of the resection defect with potential residual tumor, thus offering the possibility of a further endoscopic resection, provided that the histological tumor characteristics are favorable [35]. This is usually not possible after pmEMR, since the quality of the fragmented resection specimen is insufficient for obtaining an adequate estimation of tumor infiltration or resection status [4].

The main arguments raised against ESD are the long procedure times, the technical difficulties, and the high complication rates [3]. Similar to previous reports, our results revealed that ESD in the lower rectum required double the procedure time for lesions of the same size [7, 37]. However, it should be noted that our cohort included the entire ESD learning curve of both endoscopists primarily performing the procedure, and studies have shown that accumulating experience not only improves the outcomes but also significantly reduces procedure time [5, 38, 39]. There is no doubt that ESD procedures in difficult anatomical areas such as the anorectal junction should be performed by experienced endoscopists, but the same applies for EMR, since overcoming the technical challenges described above requires considerable experience [3]. As for periinterventional morbidity, we found no difference between ESD and EMR in our study. The much-feared colon perforation is clinically less relevant in the lower rectum, since it is supported by the mesorectum and peritoneal entry is nearly impossible [7, 9]. The risk of postprocedural bleeding is not negligible, and various studies have shown that it is higher for rectal lesions, particularly those involving the dentate line [9]. However, this is mostly associated with the anatomy of the lower rectum and not with the procedure itself, which explains the similar bleeding rates found in both groups in our study. The same applies for strictures in the anal canal and the lower rectum.

The main strength of our study is the direct comparison of ESD with EMR, with both techniques being performed in two high-volume centers by the same experienced endoscopists, thus contributing to the homogeneity of the groups regarding indication, definitions, and experience of the contributing teams. The multicentric design and the large number of consequent cases included also contribute to the validity of the results. The main limitation of this study is its retrospective nature. The retrospective study design also explains the discrepancy in the median follow-up periods between the two groups, since ESD was primarily performed in the later part of the study period. However, the median follow-up period of the ESD group, although shorter, is still long enough and exceeds the expected interval for recurrence, as demonstrated in the EMR group. Therefore we believe, that the influence of this discrepancy on our findings is minimal. Nevertheless, a prospective, randomized trial comparing these two techniques is required to verify our conclusions.

Based on the findings of our study we believe that ESD is superior to EMR for the treatment of epithelial lesions in the lower rectum that involve the dentate line, especially regarding local recurrence rates and the possibility of curative resection of low-risk early rectal cancer. In experienced hands, ESD is an effective procedure with a safety profile comparable to that of EMR. The long procedure times, the costs involved, and the technical challenges of ESD certainly do not justify its application for all colorectal lesions, and careful patient selection is required in order to maximize its advantages [40]. On the other hand, our data show that patients with lesions located in the anorectal junction are a subgroup who can profit from ESD and therefore we believe that it should be considered as the treatment of choice for those lesions.