Introduction

With increased population screening, the incidence of early rectal cancers (T1/T2) in the USA has been increasing [1]. Annually, approximately 140,000 new cases of colorectal cancers are diagnosed and of these 30 %, or 40,000 cases, are rectal cancers [2]. Approximately 10 % of these newly diagnosed rectal cancers are early T1 or T2 rectal cancers with 45 % being T1 (4500 cases) and 55 % being T2 (5500 cases) tumors [3, 4•]. Clinicians taking care of patients with newly diagnosed early rectal cancer will be faced with complex decisions on how to best care for these patients. Specifically, is this cohort of patients best served with a transabdominal or transanal resection from an oncologic standpoint?

A transabdominal resection involves a proctectomy with total mesorectal excision (TME) to adequately stage regional lymph nodes. A proctectomy with TME has a higher risk of complications (30–75 %) compared to a transanal local resection. These complications can be quite debilitating for the patient including defecatory (25–75 %), urinary (30–60 %), and sexual dysfunction (23–69 % in males and 19–62 % in females) [5–8]. The standard resection TME has a proven low local and distant recurrence rate (LR and DR) for early stage rectal cancer. Standard resection TME for T1 disease has a reported 2.7–6.0 % LR rate and a 2.6–7.0 % DR rate [9–12]. Standard resection TME for T2 disease has a reported 6.0–14.0 % LR rate and a 4.0–19 % DR rate [9, 11, 12]. Transanal local excision (TAE) is appealing because of less morbidity but at the cost of potential higher LR rates (T1 LR rates of 8.2–21.0 % and T2 LR rates of 13.0–30 %) [9, 11, 12]. Given these data, our group offers local excision only to well-selected T1 lesions without any high-risk histologic features. All other high-risk T1 or deeper rectal cancer patients are offered a standard resection unless part of a clinical trial. There is a national trend towards organ-preserving surgery with transanal resections for early rectal cancer (T1/T2) but is it justified?

The goals of treatment of early rectal cancer include optimizing oncologic control while minimizing treatment-related complications and maximizing quality of life. Patient selection begins with an understanding of the biologic behavior of early rectal cancers to predict potential lymph node spread. Clinicians are unable to precisely predict with 100 % accuracy the biologic behavior and potential lymph node spread of superficial rectal cancers. In addition, neither endorectal ultrasound (ERUS) nor magnetic resonance imaging (MRI) is 100 % accurate in clinically staging early rectal cancer. Patients may be able to avoid standard transabdominal rectal resections (low anterior resection (LAR)/abdominal resection (APR)) if they are carefully well-selected.

Patient Staging and Selection

Patients diagnosed with rectal adenocarcinoma are first clinically staged to ensure no distant metastatic disease with cross-sectional computed tomography (CT) imaging of the chest, abdomen, and pelvis. Once metastatic disease is ruled out, patient selection for local excision begins with a review of the tumor histology (a surrogate of tumor biologic aggressiveness and risk of lymph node spread), local clinical staging, and assessment of patient-related comorbidities that may influence perioperative risk. When considering local excision, it is critical to have accurate imaging to carefully select lesions limited to the superficial rectal wall without evidence of regional lymph node spread. In addition, the histology of the rectal cancer must be reviewed carefully to determine if high-risk features are present which may predict occult lymph node spread and therefore preclude a transanal resection.

Review of Tumor Histology

Patient selection with thorough review of rectal tumor histology can be useful to avoid over- and under-treatment of tumors. Histological assessment is the best predictor of biologic behavior and possible lymph node metastases. By identifying high-risk features, clinicians can assess risk of recurrence (local or distant) if a transanal resection is performed. In a systemic review, Bosch et al. identified important features in rectal cancer biopsy specimens that are associated with regional lymph node metastases and they include poor histologic differentiation (relative risk 4.8), lymphatic invasion (RR 5.2), submucosal invasion ≥1 mm (RR 5.2), and budding (RR 5.1) defined as the presence of isolated single cells or small cluster of cells scattered in the stroma at the invasive tumor margin [4•]. Additional histologic features associated with lymph node metastasis include perineural invasion (PNI), mucinous features, and signet ring cells [13–15]. Ideally, in the near future, gene expression and molecular markers would be used to identify a tumor’s risk of lymph node metastasis. Currently, there are no validated molecular markers available for routine clinical use, although methylation status of select targeted genes seems to be a promising area to predict lymph node metastatic potential of T1 rectal cancers [16].

The depth and width of submucosal invasion of early rectal cancer has been correlated with potential micro-metastatic lymph node spread. The Haggitt classification has been used to determine the depth of submusocal invasion in a pedunculated colon polyp, and this depth is correlated with the risk of lymph node spread (0—carcinoma confined to the mucosa, 1—head, 2—neck, 3—stalk, and 4—submucosa of underlying colonic wall) [17]. Haggitt class “4” lesions are associated with an increased risk of lymph node metastasis (6.2 %) [18]. For non-pedunculated lesions, the Kikuchi classification is used to divide the submucosa into three parts: sm1 (superficial), sm2 (middle), and sm3 (deep) for T1 cancers. The risk of lymphatic spread of T1 lesions has been correlated to the degree of submucosal invasion with the risk of lymph node spread being 2, 8, and 23 %, respectively, for each level [19]. A submucosal width of invasion of more than 5 mm has also been associated with an increased risk of lymph node metastasis [20–22]. Currently, the National Comprehensive Cancer Network (NCCN) guidelines identifies high-risk features for T1 rectal lesions to include poorly differentiated tumors, positive margins, lymphovascular invasion, or sm3 invasion [23].

Clinical Staging

Paramount to patient selection is accurate clinical staging of the rectal cancer. Local excision can be considered when the tumor has superficial invasion without evidence of lymph node metastases or metastatic disease. Computed tomography (CT) imaging of the liver and lungs is performed to rule out potential sites of metastatic disease; however, it lacks the resolution to accurately stage rectal cancer in the pelvis.

To assess if a patient would benefit from local excision, the lesion must first be assessed with digital rectal examination (DRE) and proctoscopic examination. The objective is to determine the size (cm), mobility (freely mobile, tethered, or fixed), circumference (%), distance from the anal verge (cm), and location within the rectum (anterior, posterior, or lateral). Although flexible endoscopy is typically used to diagnose rectal cancer, its accuracy to determine the precise location in the rectum has been questioned. Piscatelli et al. reviewed the endoscopic, pathologic, and operative reports of 236 patients with colon and rectal cancer and found that in 49 cases (21 %) the endoscopic location was inaccurate. In addition, this study found that in 12 cases, errors in endoscopic localization resulted in a change of operative approach [24]. These errors in localization can be corrected with the use of rigid proctoscopy to evaluate lesions within 20 cm of the anal verge. Of note, complete colonoscopy is necessary to rule out the presence of synchronous colorectal tumors that may occur in 2-4 % of cases [25].

Endorectal ultrasound (ERUS) and pelvic MRI are both superior to CT in accurately staging rectal cancer [26]. Both imaging modalities are used to assess tumor depth of penetration into the rectal wall (T stage) and evaluate for possible mesorectal lymph node involvement (N stage). Neither modality is 100 % accurate for staging rectal cancer. Both have strengths and weaknesses but instead of being mutually exclusive they can be complementary in assessing depth of invasion and possible lymph node involvement.

In general, ERUS has the ability to depict the layers of the bowel wall better than pelvic MRI and is therefore thought to be the imaging modality of choice for early stage rectal cancer. Based on work by Solomon et al., the sensitivity and specificity of tumor depth for ERUS is 97 and 87 %, respectively [27]. The specificity of pelvic MRI for tumor depth is lower at 75 % but accuracy can be enhanced with the use of phased array multichannel coils [26, 28]. Thus, ERUS is the best imaging modality in staging superficial lesions and in discriminating between T1 and T2 lesions, whereas pelvic MRI is the modality of choice for staging T2 or greater rectal tumors [29••]. Pelvic MRI has the additional benefit of improved visualization of pelvic anatomy including the mesorectum and circumferential radial margin (CRM), mesorectal lymph nodes, lymph nodes outside the mesorectum along the lateral pelvic side wall, and the tumor’s relationship to the anal sphincter and pelvic floor.

Both MRI and ERUS lack sensitivity and specificity of mesorectal lymph node metastasis. Lymph node size is not an accurate predictor of lymph node metastasis, as ≥50 % of lymph nodes less than 5 mm have evidence of metastasis [30, 31]. Lymph node morphology including round shape, heterogeneous signal intensity, and irregular borders can be used to increase the sensitivity and specificity of lymph node metastasis [32]. The sensitivity and specificity of pelvic MRI for detecting mesorectal lymph node metastasis is 77 and 71 %, respectively [26].

Patient Factors

Patients at high risk of surgical complications may be better served with less invasive procedures even if the local recurrence rate or risk of lymph node metastasis is higher. Patients with poor performance status may have greater threats to survival, making local excision a much more attractive alternative to TME. Patient education and involvement in decision-making regarding local excision versus transabdominal TME for early stage rectal cancer enables the patient to be part of the decision-making process. In a multidisciplinary fashion, the treating surgeon, medical oncologist, gastroenterologist, and pathologist should carefully weigh the risk and benefits.

Surgical Management

Surgery offers the best chance of cure for patients with early stage rectal cancer. When deciding upon the surgical approach, one must take into account tumor stage, size, location, risk of recurrence, need for neoadjuvant/adjuvant therapy, and patient-related factors. Population-based studies show that local excision is increasingly being used in the United States for the management of early rectal cancers (T1 and T2). This trend from 1989 to 2003 has been quite remarkable. Local excision for T1 tumors has increased from 26.6 to 43.7 % and for T2 tumors from 5.8 to 16.8 % [12]. Is this trend justified or are we putting patients at risk? One must balance the decreased morbidity of local excision techniques with oncologic control and long-term outcomes (Table 1).

Table 1 Criteria for local excision of early rectal cancer [1, 29••]
Full size table

Local Excision

Transanal Excision (TAE)

TAE is the most commonly used local excision technique and is well-suited for tumors within 6–10 cm of the anal verge. The objective is a full-thickness excision of the rectal cancer with a lateral margin of 1 cm and a histologically negative deep margin. We give patients both mechanical and oral antibiotic bowel prep the day before, and intravenous antibiotics are given to cover intestinal flora prior to the procedure. The procedure is performed under general anesthesia or under monitored anesthesia care (MAC) with local or regional (spinal) anesthesia. The patient is positioned in lithotomy for tumors involving the posterior rectal wall; otherwise, the prone jack knife positioned is preferred for other tumor location sites (i.e., anterior wall). The Lone Star retractor greatly facilitates exposure during which fragmentation of the specimen should be avoided. Care must be taken when removing anterior tumors near the posterior vagina to prevent inadvertent injury leading to potential rectovaginal fistula. Orientation of the specimen can be done prior to histologic assessment by pinning the specimen. Patients are typically discharged the same day on a low-residue diet. The patient is then seen in clinic 2 weeks post-resection to assess for complications, review pathology, and determine future strategy and surveillance.

The benefits of TAE are clearly in the reduced incidence of complications, including reductions in anorectal, bladder, and sexual dysfunction. The most frequent postoperative complications are rectal bleeding (6 %), rectal stenosis (5.5 %), urinary retention (1.5 %), and fecal incontinence (0.5 %) [33, 34]. A retrospective review by Ptok et al. of 479 patients with T1 rectal cancers found significantly fewer postoperative complications in patients undergoing local excision via TAE or transanal endoscopic microsurgery (TEM) versus transabdominal resection (8 vs. 25 %, respectively) [35].

The disadvantage of TAE is the potential high rate of recurrence if optimal local excision is not performed. In the Norway Rectal Cancer Study Group, 291 patients with T1 rectal cancer treated with transabdominal resection (n = 256) or TAE (n = 35) were prospectively followed for 6 years. Patients treated with TAE had a significantly higher 5-year local recurrence rate (12 vs. 6 % p = 0.01) and lower 5-year survival rate (70 vs. 80 % p = 0.04). However, it is important to note that 34 % of the TAE patients had either a macroscopically positive specimen margin or a margin <1 mm [36]. Alternatively, the Cancer and Leukemia Group B study prospectively evaluated 59 patients with T1 disease who underwent TAE with a follow-up of 4 years. All specimens had negative margins and no lymphovascular invasion (LVI). The local recurrence was 3.4 %, highlighting the importance TAE technique and patient selection [37].

Transanal Endoscopic Microsurgery (TEM)

TEM was first described by Beuss in 1984 as a minimally invasive technique that facilitates local excision of rectal adenocarcinoma between 4 and 15 cm from the anal verge, higher than can be accessible with the conventional TAE approach [38, 39]. The TEM platform consists of a beveled rectoscope that creates an airtight seal for standard C02 insufflation to 15 mmHg. The patient is positioned so that the tumor is in the dependent position. Using laparoscopic instruments, the specimen can be excised using the same principles employed during TAE. Within the upper rectum, one must be careful to prevent intraperitoneal perforation which has been reported to occur 4.3 % of the time [40]. This may require conversion to a transabdominal procedure if there is any concern about the integrity of the transanal closure or concern about any other visceral injury. Lesions within 6 cm of the anal verge are best removed via TAE due to difficulties maintaining CO2 insufflation. Further limitations include equipment expense and technical expertise. Patients are typically discharge the same or next day. The most frequent postoperative complications are rectal bleeding (27 %), urinary tract infection (21 %), suture line dehiscence (14 %), and fecal incontinence (1 %) [40]. TEM has also been associated with low short-term anorectal dysfunction [41].

When compared to TAE, TEM has been shown to decrease local recurrence rates. In a retrospective review by Moore et al. of patients undergoing TEM (n = 82) vs. TAE (n = 89) for early rectal cancer, local recurrence was 5 % after TEM versus 27 % after TAE (p = <0.01). TEM was much more likely to produce a non-fragmented specimen (94 vs. 65 % p < 0.001) in this series. Complication rates between the two groups were statistically similar [42]. For T1 tumors, TEM has a 5-year local recurrence rate comparable to transabdominal resection (4.1 vs. 0 % p = 0.95). However, local recurrence for T2 tumors was significantly higher (19.5 vs. 9.4 % p = 0.04) [43]. The local recurrence rate and disease-free survival for T2 lesions would favor a transabdominal resection with TME.

Transanal Minimally Invasive Surgery (TAMIS)

Transanal minimally invasive surgery (TAMIS) was designed, and first reported in 2010, as an alternative to the TEM platform that was more cost-effective and allowed for smoother adoption [44•]. Single-incision laparoscopic (SILS) ports, most commonly the GelPOINT®, are utilized to perform the procedure with conventional laparoscopic instrumentation including the use of a 30°, 10-mm high-definition laparoscopic camera [45]. Patients are positioned in dorsal lithotomy, and the lubricated SILS port is introduced into the anal canal for establishment of pneumorectum to 15 mmHg. A full-thickness resection of the neoplasm with 1-cm margins is then performed as previously described. Limitations encountered with TEM also apply to this technique including difficulties obtaining pneumorectum when attempting to remove distal rectal lesions close to the anal verge and potential intraperitoneal perforation when excising upper rectal lesions.

In a systematic review by Martin-Perez et al., the rate of a positive margin after resection of a malignant polyp was 4.4 %. In the same study, complications following TAMIS were reported at 7.4 %. The laparoscopic or open conversion rate following 390 excisions for benign and malignant lesions was reported at 2.3 %. Intraperitoneal perforation occurred in 1 % of cases of which a few cases were managed with primary transanal closure of the defect [46]. The TAMIS platform is currently being modified to accommodate the Da Vinci© robotic system for local resection and transanal TME [47].

Transabdominal Resection

Standard transabdominal resection has the best proven oncologic outcomes and is the gold standard to which all other procedures should be compared. When feasible, sphincter-sparing resections such as a LAR with TME and colonanal anastomosis should be performed. APR remains the treatment of choice for the surgical treatment of low-lying rectal malignancy when sphincter preservation would result in threatened margins.

It is well understood that morbidity is substantially lower following local excision when compared to transabdominal resection methods. However, how do oncologic outcomes compare? In the aforementioned study by You et al., the authors were able to use the National Cancer Database (NCDB) which represents a more accurate representation of clinical practice across the country rather than single-center experiences. The 5-year local recurrence and 5-year distant recurrence rates for T1 lesions resected transanally were 8.2 and 3.6 %, compared to 4.3 and 2.6 % for patients who had a standard TME resection. For patients with T2 lesions, the 5-year local recurrence and 5-year distant recurrence rates for T2 lesions resected transanally were 22.1 and 7.7 %, compared to 15.1 and 5.0 % for transabdominal TME. The overall 5-year survival was no different for T1 lesions excised locally or with standard resection (77.4 vs. 81.7 %; p = 0.09, respectively). However, the 5-year overall survival for T2 lesions resected transanally was statistically lower than transabdominal standard resection (67.6 vs. 76.5 %; p = 0.01, respectively) [12]. Many other single-center studies support these findings but are variable depending upon patient selection criteria for local excision (Table 2) [43, 48, 49]. It is our practice to offer all T2 rectal cancer patients a transabdominal resection with TME unless the patient has a poor performance status, is unwilling to undergo a TME, or the patient is part of a clinical trial.

Table 2 Review of outcomes following local excision vs. transabdominal resection for T1 rectal adenocarcinoma
Full size table

Local Excision Following Neoadjuvant Therapy

Neoadjuvant therapy prior to local excision has been proposed for select T1/T2 early rectal cancers due to evidence of complete tumor regression in 10–30 % of patients [50, 51]. The American College of Surgeons Oncology Group (ACOSOG) Z6041 phase II trial aimed to evaluate short-term outcomes after neoadjuvant long-course chemoradiation (CRT) therapy for T2N0 rectal cancer. Among the 77 patients completing the protocol, pathologic complete response (pCR) rates were seen in 44 % and tumor down staging occurred in 64 % of the patients [52]. The estimated 3-year disease-free survival for the group was 88.2 %, with a local recurrence rate of 4 % and distant recurrence rate of 6 % [53••]. This study demonstrates a low local failure rate in T2 patients after neoadjuvant CRT and local excision. This approach may offer an organ-preserving alternative in carefully selected T2N0 patients who refuse or who are too high risk for standard resection.

Local recurrence highly correlates with the pathologic response following neoadjuvant treatment. A review of multiple retrospective studies with a median follow-up between 24 and 55 months demonstrate no local recurrence if local resection revealed a pathological complete response (pCR), a local recurrence of 2 % for ypT1 tumors, and 6–20 % local recurrence for ypT2 tumors [54].

It should be noted that chemoradiation-related toxicity is not trivial. In the ACOSOG Z6041 study, 39 % of enrolled patients developed chemoradiation-related grade ≥3 toxicities [52]. Marks et al. retrospectively reviewed short-term outcomes in 43 patients undergoing long-course CRT followed by TEM and 19 patients undergoing TEM alone. Morbidity rates were substantially higher after CRT (33 vs. 5.3 %, p < 0.05), specifically wound healing related to suture line dehiscence (25.6 %) [55]. Ten patients with wound separation required long-term antibiotics, and two patients had major wound separation—one patient required additional surgery for a diverting stoma. Although there is a significant increase in wound complications, most are not major and therefore should not prohibit transanal resection after neoadjuvant radiation treatment.

Surveillance

Surveillance after local excision is crucial in identifying early recurrence so that salvage surgery can potentially be performed. Surveillance guidelines published by the National Comprehensive Cancer Network (NCCN) include the following: (1) complete history and physical exam, including digital rectal exam, every 3–6 months for 2 years; (2) carcinogenic embryonic antigen level every 3–6 months for 2 years; (3) computed tomography of the chest, abdomen, and pelvis annually for 3 years; (4) colonoscopy at 1 year and thereafter depending upon findings; and (5) proctoscopy every 6 months for 5 years. The addition of ERUS and pelvic MRI has been advocated for postoperative surveillance to increase the detection of local regional recurrences. A standardized surveillance schedule does not currently exist for these modalities. In addition to the above recommendations, it is our practice to follow patients with alternating ERUS and pelvic MRI every 6 months [56–58]. Follow-up should continue long-term, especially for patients who have had radiotherapy which has been shown to delay recurrence [59].

Salvage for Recurrence

Patterns of recurrence following TAE for early stage rectal cancer have been reported. Friel et al. reviewed 29 patients with recurrent rectal cancer following TAE. The mean time between local excision and salvage surgery was 26 months. Recurrence involved the rectal wall in 90 % of patients and was purely extra-rectal in 10 % [60]. After TAE for T1 rectal lesions, Garcia-Aguilar et al. reported an 18 % recurrence rate at 54 months. Eighty percent were local recurrences in the pelvis, 10 % were distant recurrences, and 10 % were both local and distal recurrences [9]. Finally, studies have suggested that local recurrence following local excision of early stage rectal cancer may be confined to the mesorectal fascia as opposed to the pelvic sidewall following transabdominal TME, making salvage surgery technically more feasible in this patient cohort [57].

While prospective data is lacking, retrospective studies have shown that outcomes of salvage surgery for recurrence after prior local excision of early rectal cancers are not equivalent to patients initially treated with standard TME resection. Friel et al. noted a 24.1 % (7/29) positive microscopic margin (R1) or unclear margin in patients recurring after initial transanal resection of T1/T2 rectal cancer. Eighty percent of patients who recur after local excision have resectable disease [11]. You et al. report an 80 % R0 resection rate following salvage surgery for recurrence after prior local resection. This required multivisceral resection in 33 %, total pelvic exenteration in 5 %, and metastasectomy in 25 % of patients [61]. Of those who undergo salvage surgery, the 5-year disease-free survival is 50–60 % [60–62].

Stage migration due to discordant preoperative and final pathologic staging may necessitate immediate “salvage” surgery, and this should be part of the preoperative discussion and counseling of the patient. A positive margin following local excision carries a high risk of recurrence and these patients should be offered further treatment. Hahnloser et al performed immediate (<30 days) transabdominal resection (APR = 24, low anterior resection = 28) on 37 patients after local excision due to reasons including a positive margin, LVI, and nodal disease. Similar 10-year overall survival and cancer-free survival were noted for T1N0-1 (n = 37) study patients (62 and 90 %) compared to a standard resection case match-cohort (72 and 84 %, respectively p = 0.04)). They concluded that local excision followed by transabdominal salvage surgery within 30 days does not adversely affect long-term survival. Several other studies have reported similar findings [1].

Conclusion

Appropriate patient selection is the most important factor in determining if transanal resection is the best treatment strategy for patients with early stage rectal cancers. Patient selection involves careful assessment of high-risk histology, clinical staging (ERUS/MRI), and consideration of patient-related factors. We limit transanal resection to only well-selected T1 lesions without high-risk pathologic features (well-differentiated, no lymphovascular invasion, submucosal depth of invasion ≤1 mm) and clinically staged with ERUS as T1. It is our practice to limit transanal resection of T2 lesions to patients with poor performance status, patients who refuse standard TME resection, or patients participating in a clinical trial. Newer techniques such as TEM and TAMIS may decrease local recurrence rates by decreasing fragmentation and increasing R0 resections (Fig. 1).

Fig. 1
figure 1

Early rectal cancer treatment algorithm. ERUS endorectal ultrasound, CEA carcinoembryonic antigen, LVI lymphovascular invasion, Sm submucosal, DRE digital rectal examination

Full size image