Transanal endoscopic microsurgery (TEM) was introduced in the early 1980s as a minimally invasive procedure designed for local resection of rectal lesions that otherwise would require major abdominal or abdominoperineal resections. Compared to conventional surgery, this method results in significantly reduced morbidity and mortality rates below 1 % [3]. TEM was initially proposed for large adenomas out of reach for transanal excision and unsuitable for colonoscopic removal. Later, the indication has expanded to early rectal carcinomas [4] or for palliation in more advanced stages. However, the role of TEM in rectal cancer is still a subject of much debate. The most common controversies are:

Thus, the indications for TEM-surgery are highly dependent upon a multidisciplinary approach involving endoscopists, pathologists, surgeons, radiologists and oncologists.

Malignant changes in large polyps can be very difficult to diagnose even after several biopsies. Ideally, all tumours considered for TEM should undergo meticulous endorectal ultrasonography (ERUS) with the purpose to investigate invasive growth and local lymph node involvement. Its diagnostic accuracy in the assessment of early T1 carcinomas can be up to 89 %, with a sensitivity and specificity of 92 and 50 %, respectively, in experienced hands [5], and thus superior to digital examination by colorectal specialists, computed tomography and MRI [6, 7]. Despite these results, any patient with early rectal cancer undergoing TEM is recommended to undergo a full-thickness resection, preferably including mesorectal fat allowing adequate pathological examination of the specimen.

Precise preoperative staging is imperative since the procedure does not remove any, or only a few, of the perirectal lymph nodes. This is the main reason for a higher risk of local recurrence after TEM compared to conventional surgery. The recurrence rate for T1 cancer is 0–12 %, for T2 cancer 12–28 % and for T3 cancer 36–79 % [8–10]. In past decades, TEM has usually been indicated in patients with low-risk pT1N0 adenocarcinoma. Low-risk lesions are primarily those with a small-size (<3 cm), well-differentiated histology with the absence of vascular, lymphatic or perineural invasion. When these criteria are met, survival and local recurrence rates achieved by TEM are similar to those with conventional radical surgery. Local recurrence varies between 4.2 and 9.6 % with a 5-year survival rate varying from 79 to 100 % [11–15]. In the T1 high-risk patients, the local recurrence rate was as high as 39 % [16], and in these cases, salvage surgery must be considered.

In none of the studies were the recurrence rate and long-term results significant related to invasion of the submucosal layer (sm). Lymph node involvement of 1–3 % in sm1, 8 % in sm2 and 23 % in sm3 lesion has been reported [17]. This may suggest that the indication of TEM for cure should be reserved for patients with sm1 lesion. However, larger patients’ series are needed for more firm conclusions. The preoperative staging of submucosal invasion is a great challenge to the standards of the equipment. In a study using high-resolution, three-dimensional endorectal ultrasonography, the overall kappa for the concordance between ultrasonographic and histopathologic staging for the degree of submocosal invasion (slight or massive) was 0.81, and no invasive carcinomas remained undetected [18].

No preoperative investigations (ERUS, MR or CT) have revealed a sufficient sensitivity or specificity for lymph node metastasis in patients with early cancer. This has led to the development of endoscopic posterior mesorectal resection with the preservation of the anorectal function and with a low rate of morbidity [19]. In a series of 11 patients with T1 tumour, 4–20 lymph nodes (median 8) were removed in each patient and without significant complications. In two patients, lymph node metastasis was detected. Thus, combining TEM with the posterior endoscopic mesorectal excision might reduce the local recurrence rate after TEM. However, further investigations with larger prospectively evaluated patient series are needed.

Whereas the use of preoperative neoadjuvant RCT is controversial in early T1 cancers considered for TEM, it is mandatory in T2 or larger tumours where a local recurrence rate varying from 29 to 50 % has been reported [20]. In a study with 100 patients undergoing TEM after radiotherapy (54 patients with uT2 and 46 patients with uT3 uN0), complete response or microscopic residual tumour was found in three and 15 patients, respectively [21]. Minor complications occurred in 11 patients and major complications in two patients. The cancer specific survival rate after 90 months follow-up was 89 % and the overall survival rate 72 %. Salvage abdominoperineal surgery was performed in three patients, two of whom were disease-free at 15 and 19 months. Similar results were found in another study, where the patients were randomized to either TEM or laparoscopic resections [22]. Local or distant failure was 10 % after TEM and 12 % for laparoscopic resections. The survival was 95 % for TEM and 83 % for laparoscopic resection after a median follow-up of 56 months, but the difference was not statistically significant. Other studies have shown similar results. However, in all studies, the patients were highly selected, and the treatment modality with preoperative RCT followed by TEM should be reserved for patients with a known higher risk at conventional surgery. Another and unsolved problem is the place for postoperative RCT in patients, who at histology turned out to have a higher tumour stage than T1 at the final histology examination than judged preoperatively. In general, these patients should be offered salvage surgery, as the oncological outcome after salvage is comparable to primary radical surgery [14, 15, 20, 23]. In a study which included patients with T2 lesions, that following TEM were treated with 5-FU and radiation (54 Gy), local recurrence was observed in 14 %. However, salvage was successful in less than half of the patients with local recurrence [24]. There is no available information on the results of postoperative RCT compared to salvage surgery.