A radical resection with a total mesorectal excision (TME) is the standard of care for all patients with rectal cancer including T1 cancers where overall survival is greater than 80 % with a local recurrence rate less than 10 % [28, 29]. However, despite advances in minimally invasive technology over the last decade, this is not without morbidity. In many series, there is a 2–3 % risk of death with up to a 40 % risk of complication with proctectomy [30]. There is often a need for a temporary or permanent stoma [31]. There is a significant risk of poor bowel function with a low anastomosis, poor healing with perineal wounds, sexual and bladder dysfunction, and depression. This is in drastic comparison to TEM, which has been shown to have minimal effect on fecal incontinence with an excellent quality of life following surgery [32]. For patients with early-stage tumors, the oncologic goals must be balanced against the effect on quality of life.

As such, surgeons have employed local excision for these distal early cancers for decades. Early reports in the 1990s were quite promising. Data from the CALGB 8984 trial, which evaluated the efficacy of local excision (TAE) in the treatment of T1 and T2 rectal cancers, revealed a local recurrence rate of 8 % with a distant metastasis rate of 5 % in 59 patients undergoing local excision for T1 tumors [33]. This was associated with an overall survival of 84 %. The recurrence rate for T2 tumors was higher at 18 % with a 12 % distant recurrence rate corresponding to a 66 % overall survival rate. However, reports began to surface demonstrating significantly higher rates of recurrence. Garcia-Aguilar published the University of Minnesota group’s experience demonstrating an 18 % recurrence rate in 55 T1 lesions with a 98 % survival with a mean of 54 months of follow-up. The recurrence rate for T2 lesions was 37 % [34]. Data from Memorial Sloan-Kettering was similar with a 17 % 10-year local recurrence rate and 74 % 10-year disease-free survival in 74 patients with T1 cancers. In this study, 50 % of recurrences were local recurrences only, suggesting inadequate resection as the cause of treatment failure [35]. A nationwide cohort study from the National Cancer Database demonstrated a dramatic increase in the use of local excision for T1 rectal cancers from 27 to 43 % between 1989 and 2003 with an associated increase in local recurrence (12.5 % vs. 6.9 % for T1 and 22.1 % vs. 15.1 % for T2, p < 0.05) and decline in overall survival as compared to standard resection for T2 tumors (77.4 % vs. 81.7 % for T1, p = 0.09, and 67.6 % and 76.5 % for T2, p < 0.05) [28]. Findings of higher recurrence and lower overall survival were also demonstrated in several other observational studies comparing traditional TAE to radical resection as well [14, 36]. This led to scrutiny for the increasing rate of local excision for Stage I rectal cancer and thus methods to improve TAE.

Transanal excision fails secondary to inadequate removal of the primary tumor, unrecognized nodal disease, or systemic spread. Luminal recurrences account for the majority of the local recurrences following TAE [37]. Therefore, inadequate removal of the primary tumor and tumor implantation by poor surgical technique likely contributes substantially to the significant recurrence rates seen in the prior studies utilizing standard TAE. It stands to reason that improved surgical technique could equate to improved oncologic outcomes.

As previously mentioned, TES is a superior technique to standard TAE with improved rates of negative margins and less fragmentation. Table 2.2 demonstrates the case control studies to date with reported outcomes following TES (majority with TEM) for T1 rectal cancer. The recurrence rates are seemingly lower than earlier reports with standard TAE although certainly there is risk for publication bias in these series. To date there are no randomized controlled trials comparing standard TAE to TES, only small retrospective studies with the expected limitations. Moore et al. [5] in Vermont compared 28 patients undergoing TEM for malignancy to 89 patients undergoing traditional TAE. The recurrence rate for TEM was only 3 % compared to 26 % after TAE. However, the follow-up for standard TAE was more than double that for TEM (53 ± 44 months vs. 20 ± 16 months, p > 0.05), thereby influencing the results. Christoforidis et al. [7] reported on 37 TEM procedures for malignancy performed by one surgeon compared to 117 TAE performed by 21 different surgeons and found the recurrence to be 12 % vs. 22 % for T1 tumors and 25 % vs. 33 % for T2 (p > 0.05). The follow-up was more similar in these two groups at 60 months (10–125 months) for TEM compared to 45 months (7–133 months) for TAE.

Although not randomized, De Graff et al. [38] reported their experience with TEM as a recruitment center for the Dutch TME trial and referral center for TEM. Compared to 75 patients with T1 tumors that underwent TME, there was a significantly higher rate of recurrence in 80 patients that underwent TEM at 25 % compared to 0 % with similar DFS at 87 and 90 %. There was no mention of pathologic risk stratification for T1 tumors in the Dutch trial, which is known to influence nodal metastasis and thus recurrence. This recurrence rate is the highest of the series in Table 2.2 and highlights the difficulty in interpreting the data when there are such wide variances between groups.

Studies including patients with T2 rectal cancer are outlined in Table 2.3. Local excision alone for T2 tumors is insufficient with recurrence rates up to 40 %, mirroring the known incidence of nodal metastases in these patients. It remains to be seen whether local excision combined with chemoradiation therapy (CRT) will prove to be an acceptable form of treatment. However, the initial results of the American College of Surgeons Oncology Group (ACOSOG) Z6041 trial are promising [39]. The ACOSOG Z6041 trial is a prospective, multicenter, single-arm Phase II trial to assess the efficacy and safety of neoadjuvant CRT and local excision for T2N0 rectal cancer. Ninety patients were accrued and 79 patients completed therapy. There was a 44 % complete pathologic response rate and a 39 % CRT-related toxicity, mainly rectal pain following surgery [39]. The oncologic outcomes were reported at the American Society of Colon and Rectal Surgeons annual meeting in Hollywood, FL. After a mean follow-up of 4.2 years, there was a 3 % local recurrence rate and 7 % distant recurrence rate with an 87 % 3-year DFS [40]. However, given that 44 % of patients had a complete pathologic response to CRT, it remains unclear what clinical effect the addition of local excision had in these patients. A trial of nonoperative treatment in these patients is currently underway by the same group.

Other studies evaluating the feasibility of TES in T2N0 tumors combined with CRT have been reported. The Urbino trial was an Italian randomized controlled trial that randomized 70 patients with T2N0 lesions who underwent neoadjuvant CRT to either TEM or laparoscopic low anterior resection or abdominoperineal resection. Five-year follow-up demonstrated less than 10 % recurrence rates with a 94 % disease-free survival in each group [41]. The CARTS study is an ongoing multicenter feasibility study in 15 Dutch hospitals to evaluate whether chemoradiotherapy followed by TEM will provide effective oncologic outcomes. Data from this trial are not yet available [42].

The Achilles heel of local excision of rectal cancers is inaccurate nodal staging with current technology. Up to 12 % of T1 tumors and 30 % of T2 tumors are associated with lymph node metastases (LNM) in rectal cancer specimens [43]. Preoperative staging for patients with rectal cancer involves either endorectal ultrasound (EUS) or magnetic resonance imaging (MRI) to evaluate the depth of tumor extension through the rectal wall and the presence of lymphadenopathy. Unfortunately, as Salinas et al. [44] demonstrated, preoperative imaging in Stage I cancers is inaccurate. In this study, 109 consecutive patients with preoperative imaging suggestive of T1N0 or T2N0 tumors underwent radical surgery with TME at Massachusetts General Hospital. Twenty-seven patients were found to have T3 disease (25 %) on final pathology, while 11 % of the T1 and 28 % of the T2 lesions were found to have LNMs [44]. Given that this is the exact proportion of LNMs that you would expect in T1 and T2 tumors based on historical data, this implies that nodal staging is inaccurate in these early-stage tumors. This was further demonstrated by Landmann et al. [45] who found the sensitivity of EUS for nodal staging to be linked to the size of the lymph node. Because T1 tumors are typically associated with small LNMs (median lymph node size of 3.3 mm compared to 8.0 mm for T3 tumors), EUS accuracy is less than 50 % in T1 rectal cancers compared to 84 % in T3 tumors [44, 45]. MRI has emerged as an important advance in preoperative staging of late-stage rectal cancer. It is highly accurate in predicting the circumferential radial margin and identifying invasion into surrounding structures. However, its use in patients with early Stage I rectal cancer is of questionable efficacy [46], and there are data to suggest that EUS is more accurate for staging in patients with early-stage rectal cancer [47].

Success with local excision of early-stage cancers relies heavily on patient selection. There is significant variation in oncologic outcomes between groups as evident in Tables 2.2 and 2.3. It is difficult to explain the variances in outcomes between studies. Certainly technique plays a significant role. But patient selection as it pertains to predicting the risk of LNM is just as likely to influence outcomes between studies. In addition to T stage, there are other known factors for predicting LNM, including differentiation, lymphovascular invasion (LVI), SM classification, and tumor budding. Well to moderately differentiated tumors are associated with LNMs in 14 % of T1 or T2 tumors as opposed to 30 % with poorly differentiated tumors [48]. Similarly, the presence of LVI doubles the risk of LNMs from 14 to 33 % [48]. In the absence of any of these high-risk features, Blumberg et al. found the risk of LNMs to be 7 % for T1 lesions [48].

The SM classification further stratifies T1 tumors according to the degree of invasion into the submucosa, which correlates with LNMs. Nascimbeni et al. [49] stratified the degree of invasion into the submucosa into three levels and found significant correlation with nodal disease. SM1 tumors had a 3 % risk of nodal metastases, SM2 8 %, and SM3 23 %. Along with LVI and distance from the anal verge, SM classification was found to independently predict LNM. Finally, the presence of tumor budding is emerging as an increasingly important predictor of nodal metastases [50]. Tumor budding refers to the presence of individual or clusters of tumor cells at the invasive front of the tumor, representing an epithelial to mesenchymal transition (Fig. 2.2). In one study involving 55 patients with T1 rectal cancers undergoing radical resection, tumor budding had a sensitivity of 83.3 %, a specificity of 60.5 %, and a negative predictive value of 0.958 for LNM [51].

Patient selection, focused on these high-risk pathologic features, can improve oncologic outcomes. Ramirez stratified patients with T1 rectal cancer following TEM into low-risk and high-risk categories based on the presence of poor differentiation and LVI. High-risk patients were treated with adjuvant CRT. The recurrence rate (including both local and distant) was 7.4 % in the low-risk group and 12.5 % in the high-risk group for an overall recurrence (local and systemic) rate of 9 % [52].

As evident from the studies outlined above, the decision to perform local excision for early-staged rectal cancer cannot be entered into lightly. Given the recurrence rates for T2 lesions and the known risk of nodal metastases, TES alone is not appropriate for patients with T2 tumors unless the patient is otherwise unfit for an operation or is in a clinical trial. In this circumstance, local excision should be combined with CRT. Whether or not this should be given before or after local excision remains to be seen.

With regard to T1 rectal cancers, the NCCN guidelines list the following requirements for consideration of transanal excision:

Close surveillance following transanal excision of early-stage rectal cancer is imperative as an unrecognized recurrence is often incurable. Doornebosch followed 88 patients with T1 rectal cancer that were treated with TEM [53]. Eighteen patients recurred at a median of 10 months (4–50) following local excision. Two had metastatic disease at the time of presentation, while 16 underwent salvage surgery. An R0 resection was achieved in 15 of 16 patients. Median follow-up after surgery was 20 (2–112) months. Overall survival was 31 % at 3 years; cancer-related survival was 58 %. Stipa followed 86 patients with T1 rectal cancer that underwent TEM [53]. Ten patients recurred (11.6 %) and all but one underwent salvage surgery with an R0 resection. Overall 5-year disease-free survival was 92 % in all 86 patients [53]. Median time to recurrence was 11.5 months (1–62). Finally Weiser et al. reviewed 50 patients who underwent salvage surgery for curative intent following local recurrence after transanal excision (not necessarily TEM) at Memorial Sloan-Kettering. Fifty-five percent of these patients required a multi-visceral resection although this is likely biased by the referral pattern of the institution. Five-year disease-specific survival was 53 % [54].