Transanal Endoscopic Microsurgery After Neoadjuvant Chemoradiotherapy for Rectal Cancer

Introduction

Neoadjuvant chemoradiation therapy (CRT) followed by radical rectal surgery has been considered the preferred treatment strategy for locally advanced rectal cancer. ¹ The relatively high morbidity of radical rectal surgery can make these operations contraindicated for some patients with severe comorbidities, as well as unacceptable for some patients who refuse to consider stoma. ² Preoperative CRT has an important effect on improved local disease control by tumor down-staging, downsizing, and decreasing the rate of lymph node metastases. Such tumor regression may alter final surgical strategy by leading to more frequent performance of sphincter-preserving procedures and fewer stoma formations.^3,4

Tumor response to CRT ranges from no response to clinical complete response (cCR [the total absence of endoluminal tumor in the clinical evaluation]) and pathologic complete response (pCR [complete tumor disappearance on pathological examination of the surgical specimen]).^5–7 Tumor down-staging together with pCR after CRT is an important prognostic factor for local recurrence and survival. Several reports had shown improved long-term outcome in patients who underwent radical rectal resection after down-staging or a cCR following preoperative CRT compared with those with minimal response.^8,9 Some reports observed a cCR to CRT without surgical resection, with excellent outcomes in terms of local recurrence and overall survival rates.^10,11

Because a cCR is not an accurate predictor of a pCR,^12,13 a full-thickness transanal local excision (LE) is offered to patients with a cCR to determine the rectal wall pathologic response, avoiding radical surgery if there was a rectal wall pCR (ypT0).^10,13 For that matter, transanal LE is often considered as being a powerful diagnostic tool that provides adequate and complete pathological information on the ypT status, the tumor regression grade, and other pathological features that can be useful for the surgeon in arriving at the final management decision.

Transanal endoscopic microsurgery (TEM) is a minimally invasive endoscopic technique that improves the quality of LE for rectal lesions. TEM enables excellent access and visualization of the rectal lesions and allows precise, full-thickness excisions of those lesions. ¹⁴ TEM is usually associated with few requirements for postoperative analgesics, few complications, short duration of hospitalization, and good functional results.^15,16 It is also considered as being superior to traditional transanal LE with regard to safety and local control. ¹⁷ Traditionally, suitable candidates for TEM have benign rectal lesions and early T1 rectal cancer with good prognostic factors and negative lymph nodes. TEM for patients with more advanced (T2 and T3) rectal cancer after neoadjuvant CRT has gained greater acceptance over the past few years. ¹⁸

Many studies have demonstrated significantly lower morbidity rates after LE and TEM compared with radical surgery in patients with rectal tumors. Because neoadjuvant CRT for rectal cancer may be associated with an increased rate of complications, the suitability of TEM following CRT is still questioned by some authors.^19,20 In this study, we aimed to evaluate the outcomes of patients with rectal cancer undergoing TEM after neoadjuvant CRT and to compare them with the outcomes of patients with rectal cancer managed by TEM alone.

Patients and Methods

From May 2004 to September 2010, all patients diagnosed as having benign or malignant rectal tumors who were treated by TEM at the Department of Surgery, Hasharon Hospital, Rabin Medical Center, Petach Tikva, Israel, were evaluated for this study. Their complete medical records were reviewed for data pertinent to this study. In our department, patients are considered suitable for TEM if they have benign rectal lesions, superficial adenocarcinoma arising from an adenoma, or early T1 cancer with negative lymph node. They are also candidates for TEM if they had a cCR together with disappearance of the rectal tumor after neoadjuvant CRT. Patients for whom the preoperative or postoperative pathologic examination of the specimen showed malignant rectal tumors were included in the study. Excluded were patients who had benign rectal lesions treated by TEM, those with metastatic rectal carcinoma who were operated for palliation, and those whose data were incomplete.

All the study patients were evaluated according to a standard protocol that included clinical examination with digital rectal examination, colonoscopy with biopsy, rigid proctoscopy, and endorectal ultrasound (EUS). Both the distance and the location of the tumor were assessed, as was the tumor size. “Distance” was measured from the dentate line to the lower margin of the tumor. The “location” of the tumor was assessed because the patient needs to be positioned with the tumor facing downward during surgery.

The pretreatment evaluation of patients scheduled for neoadjuvant CRT included a chest radiograph, computed tomography (CT) of the abdomen and pelvis, and a fluorodeoxyglucose positron emission tomography scan in the later years of the study period. EUS was performed routinely. Depth of invasion through the rectal wall and the presence of abnormal lymph nodes were recorded for accurate T and N staging.

Preoperative radiotherapy was delivered at a dose of 45 Gy in 25 fractions. Most of the patients were treated with fluoropyrimidine-based chemotherapy consisting of leucovorin or capecitabine. Clinical response to CRT was assessed by digital rectal examination and rigid rectoscopy in addition to an EUS, which was carried out 4–6 weeks after the completion of the neoadjuvant therapy. Patients were also restaged before surgery with abdominal and pelvic CT, and some patients had fluorodeoxyglucose positron emission tomography as well. Patients who had no detectable tumor on rectal examination or rectoscopy were defined as having a cCR. A localized scar without rectal wall thickness in the EUS was also defined as a cCR. Lesions that had been reduced to small scars or ulcers were marked by ink spotting before surgery for accurate identification and localization.

All patients with malignant rectal tumors were offered radical and definitive surgery after the completion of CRT. They underwent TEM instead of radical resection if they refused to undergo a formal rectal resection, which carried the risk of stoma formation. TEM was also performed in patients with cCR who did not reject radical surgery altogether but who strongly preferred LE by TEM.

The procedure was performed with original Richard Wolf (Knittlingen, Germany) TEM equipment and with the patient under spinal or general anesthesia. The patients were placed in a prone jackknife or lithotomy position, depending on the tumor location. The lesion was removed by excising a full-thickness rectal wall with a 1-cm margin around the tumor. The underlying mesorectal fat was included with the specimen. The specimens were pinned and marked for orientation by the surgeon. The rectal defect was closed primarily in a transverse fashion with absorbable sutures.

All patients had a urinary catheter in place at the time of surgery, which was removed the day after surgery. Postoperative pain management for all patients included oral dipyrone or paracetamol and oral narcotics (tramadol) on demand. Patients were allowed to resume eating the following day and were discharged from the hospital 2–3 days after the operation. They were evaluated 2 weeks after surgery and reexamined at 3-month intervals for the first 2 postoperative years and every 6 months thereafter. Clinical examination and rectoscopy were performed during each of the follow-up sessions. A metastasis work-up was also performed periodically. Perioperative complications were defined as unexpected and untoward events during the procedure or the postoperative period. The patients were analyzed as two separate groups according to whether or not they had undergone preoperative CRT treatment (i.e., the neoadjuvant CRT group and the non-CRT group).

The local ethics committee of Rabin Medical Center approved the study protocol and the data collection. Informed consent was waived.

Results

In total, 97 patients underwent 98 TEM procedures during the study period. Forty-four (45%) patients underwent the procedure for rectal malignancy, and they comprised the study cohort. Thirteen (30%) of these patients underwent TEM following neoadjuvant CRT treatment (the CRT group), and 31 (70%) patients underwent TEM without it (the non-CRT group).

The pre-neoadjuvant CRT treatment staging by EUS in the CRT group was T2N1 in 4 patients, T2N0 in 2 patients, T3N0 in 6 patients, and T3N1 in 1 patient. TEM was performed 9 weeks (range, 5–18 weeks) after completion of the CRT. The preoperative tumor pathology in the non-CRT group was villous adenoma in 14 patients, carcinoma in polyp in 6 patients, undetermined excisional margins following endoscopic polypectomy in 7 patients, and T1 carcinoma in 4 patients. The EUS lymph node status was N0 in all the non-CRT patients.

The median age was 69 years (range, 59–86 years) for the CRT patients and 72 years (range, 35–90 years) for the non-CRT patients (P=.717). The two groups were well matched for demographic data and American Society of Anesthesiologists scores. There were no significant differences in their comorbidities, including ischemic heart diseases, diabetes, and chronic lung disease (Table 1).

The tumor diameter in the non-CRT patients was larger (2.9 cm; range, 0.5–4.2 cm) than that of the CRT patients (2.1 cm; range, 0.5–3.5 cm), but the difference was not significant (P=.125). The two groups were comparable for the tumor location in the rectal wall, and there were no significant differences in the distance of the lower part of the tumor from the anal verge (7.7 cm [range, 5–15 cm] and 6.7 cm [range, 5–10 cm] for the non-CRT and CRT groups, respectively; P=.285) (Table 1).

There were no significant group differences regarding the type of anesthesia. The mean operative time was shorter in the CRT group (109 minutes; range, 85–135 minutes) than in the non-CRT group (119 minutes; range, 75–185 minutes), but the difference was not significant (P=.28). The postoperative hospital stay was similar for the two groups. Table 1 summarizes the demographics and perioperative data for the two study groups. The overall perioperative complication rate was 25% in each of them. The non-CRT patients had more overall complications (n=8; 26%) than the CRT patients (n=3; 23%), but that difference was also not significant (P=.659) (Table 2).

Peritoneal entry occurred in 2 non-CRT patients, 1 of whom underwent a laparotomy and protective ileostomy for insecure suture of a large rectal defect. Re-operation was performed in another non-CRT patient because of immediate postoperative rectal bleeding; that surgery consisted of coagulation of the bleeding site and resuturing of the resectional margins. Two patients in each group had postoperative urinary retention that was resolved after the replacement of a urinary catheter. One non-CRT patient had pulmonary edema on the second postoperative day, and another non-CRT patient developed pneumonia on the third postoperative day; both were managed conservatively. There were two hospital re-admissions: 1 CRT patient was re-admitted with rectal pain 3 days after discharge, and 1 non-CRT patient was re-admitted with fever 5 days after discharge. Abdominal and pelvic CT studies failed to reveal any peritoneal free air, pelvic abscess, or any contrast material leakage in either of those patients, and both were managed conservatively. No mortality or major complications occurred in the early postoperative period in either group. Urge fecal incontinence developed in 2 non-CRT patients, and they were treated conservatively with biofeedback because the EUS did not reveal any sphincter tears. No symptomatic rectal stenosis or late complications related to TEM developed among any of the patients in both groups.

Seven patients in the CRT group had no residual tumor in the specimen and were considered to have mural pCR (ypT0); they had close surveillance without additional surgical interventions. Six patients in the CRT group had a residual cancer in the final pathology (3 patients had a T1 [ypT1], and the other 3 had a T2 [ypT2] in the TEM specimen); 4 of them underwent anterior resection of the rectum, and 2 other patients were followed without radical resection. After a median follow-up of 62 months the recurrence rate in the CRT group was 7.7%. A recurrent disease was detected in only 1 patient 8 months after TEM; the patient had a ypT2 and refused radical surgery, and he died with diffuse metastatic disease 2 months later. Another patient with ypT1 who also refused additional surgery died after 28 months with pneumonia.

In the non-CRT group 9 patients had no residual tumor in the specimen, and the other 5 patients had carcinoma in situ; they were followed without further surgical interventions. Fourteen patients had a T1 in the TEM specimen pathology; 11 of them had close surveillance without additional surgery, and the other 3 patients underwent anterior resection. Three patients had a T2 in the TEM specimen; 1 of them underwent abdominoperineal resection for involved resectional margins, and the 2 other patients underwent anterior resection. In a median follow-up of 64 months a local recurrence was detected in 2 (6.4%) patients, 13 and 18 months after the TEM; both had a T1 carcinoma in the specimen. Subsequently they underwent anterior resection of the rectum.

Discussion

The well-documented benefits of neoadjuvant CRT for rectal cancer include tumor regression and down-staging associated with increased tumor resectability and a high rate of sphincter preservation. Radical surgery for rectal cancer carries a high risk of morbidity and mortality and can also greatly alter a patient's quality of life. In light of the facts that significant response rates that can be achieved with preoperative CRT and that many patients are still subjected to an unnecessary radical procedure associated with high morbidity rates, some authors have suggested limiting further surgical therapy to LE alone or observation for patients with a cCR. There is therefore increasing interest in properly identifying these patients in order to be able to offer a less invasive and still adequate oncologic treatment.^5,9–11

Assessment of tumor response after chemoradiation is currently determined by clinical restaging using digital rectal examination and rectoscopy, as well as by subjective evaluation using EUS and magnetic resonance imaging. There is concern that clinical examination is unreliable in assessing tumor response following CRT. In one report, up to 75% of patients who appeared to have obtained a cCR still had microscopic disease at the time of full-thickness biopsy. ⁵ In another series, 76% of patients who were subsequently demonstrated to have had a pCR had palpable abnormalities with the rectum that were consistent with residual tumor. ²¹

Following CRT therapy, it is difficult to distinguish between residual tumor and radiation fibrosis with thickening of the rectal wall using EUS, the accuracy of which is markedly decreased in such cases. ²² Similarly, using magnetic resonance imaging to restage patients who had been treated with preoperative CRT also does not appear to be accurate. ²³ Positron emission tomography–CT is another diagnostic modality used to restage rectal cancer after CRT; however, it appears to be more accurate in identifying distant metastasis than in identifying residual rectal tumor and predicting local response. ²⁴ The difficulties of accurate prediction of pCR limit the use of clinical response alone as an end point for determining future management. Using imaging techniques to identify patients with a pCR is difficult, and accordingly transanal LE of the remaining tissue or residual tumor is probably the most reliable method to confirm the response to CRT pathologically.

Several studies have reported the feasibility of performing LE for rectal cancer following CRT and showed favorable outcomes for selected rectal cancer, which was down-staged to ypT0 post-CRT.^4,25 Smith et al. ²⁶ recently evaluated the clinical significance of residual mucosal abnormalities after neoadjuvant therapy for rectal cancer, and those authors showed that small mucosal abnormalities (i.e., less than 3 cm) were strongly associated with good pathological response (ypT0–1) and a low rate (2%) of lymph node metastasis. They came to the conclusion that these patients could be offered LE as a macrobiopsy to rule out the persistence of cancer within the rectal wall and to avoid the risks of an unnecessary total mesorectal excision.

TEM, an LE technique that enables excellent access and visualization of rectal lesions and precise, full-thickness excision of the lesion, is considered as being superior to traditional transanal LE in terms of safety and local control. ¹⁷ When it is intended as being the definitive and curative treatment, TEM should be limited to early-stage T1N0 cancer,^27,28 and, with recent advances in chemoradiation, LE can be used selectively for T2 lesions when augmented by neoadjuvant or adjuvant therapy.^18,29 Encouraging results were also reported on patients with T3 rectal cancer treated by LE or TEM following neoadjuvant treatment.^25,30 Lezoche et al. ¹⁸ compared TEM and laparoscopic total mesorectal excision in patients with T2 lesions after neoadjuvant CRT and found no significant difference in the outcome between the two groups during a minimum 5-year follow-up.

The overall complication rate for TEM for benign and for malignant lesions has been reported to range from 6% to 31%.^15,17,27,28 Perioperative complications include peritoneal entry and hemorrhage, which may require conversion to laparotomy. Perirectal abscess, urge fecal incontinence, and rectal stenoses have also been described after TEM. In our current group, the overall complication rate was 25%, it did not different significantly between the two groups, and most of the complications were minor. The reported intraperitoneal perforation rate varies from 0% to 9%.^27,31 However, perforation into the peritoneal cavity can be managed by primary closure of the defect and does not always necessitate conversion to laparotomy or anterior resection. ³² Moreover, some series reported no increase in major or minor complications or no significant increase in the hospital length of stay for patients with peritoneal perforation compared with those without perforation after TEM. ³³ In the present study, two non-CRT patients had peritoneal entry, both on the lateral rectal wall position at 10 and 11 cm from the anal verge. Both patients had primary closure of the defect, and a protective ileostomy was performed for insecure suture line in 1 of them.

Postoperative bleeding, which occurred in 1 non-CRT patient (3%) in the immediate postoperative period, was most likely due to inadequate hemostasis of resectional margins. Our complication rate was therefore within the range of 1%–13% reported in the literature.^27,33 The wound healing process after rectal defect closure in patients undergoing TEM following neoadjuvant chemoradiation has been a subject of debate that yielded different and sometimes contradictory recommendations. In fact, the reported complication rates related to rectal wounds range from 0% to 60.9%, possibly reflecting the inconsistencies of this issue.^19,20,34 Marks et al. ¹⁹ have raised some concerns about wound healing. They compared short-term outcomes of 43 rectal cancer patients who were treated by neoadjuvant CRT followed by TEM with those for 19 patients treated by TEM alone. The wound complication rate was 25.6% for the CRT group (11 cases) and 0% for the non-CRT group (P=.015). However, 10 of their CRT patients (91%) were treated conservatively, and only 1 patient required a diverting colostomy.

In contrast, Lezoche et al. ¹⁸ found no significant difference in complication rates between patients undergoing TEM or laparoscopic total mesorectal excision after neoadjuvant treatment for T2 rectal cancer. Those authors observed that CRT was well tolerated by the patients and that it did not increase the technical difficulty during laparoscopic resection or the suture leakage rate after TEM, which occurred only when tension on the suture line was present. Others also reported that there were no significant differences in morbidity and wound dehiscence between patients undergoing TEM for irradiated and nonirradiated rectal cancer. ²⁸ One possible explanation is that ischemia rather than irradiated tissue may play a major role in wound dehiscence, like other wound failure causes, such as anastomotic leaks after anterior resection of the rectum, which are thought to occur mainly in the ischemic area caused by a weakened zone at the crossing of at least two staples in the double-stapler technique. ³⁵

Perez et al. ²⁰ reported the 30-day results of 36 consecutive patients treated by TEM for rectal neoplasm: 23 underwent CRT followed by TEM, and 13 were managed by TEM alone for various rectal lesions. Those authors reported a significantly higher rate of suture line dehiscence (60.9% versus 23.1%; P=.032) and hospital re-admissions (43.5% versus 7%; P=.025) among patients who had undergone TEM after neoadjuvant CRT. However, no patient in their series required surgery to repair the dehiscence. It is important that the closure and suture of the rectal wound in 11 (31%) patients in this series were performed with a Hill–Ferguson retractor and not with the TEM tools.

The two re-admissions in our series included 1 CRT patient with rectal pain and 1 non-CRT patient with fever. Neither of them had evidence of pelvic abscess or obvious wound failure, and both were successfully managed medically without any surgical intervention.

Because we perform TEM after neoadjuvant CRT when cCR has been achieved, most lesions had been reduced to a small scar or even completely disappeared. The TEM was not offered as a curative treatment but as a "large biopsy" in order to determine the pathological status and response of the rectal wall tumor. Patients with incomplete response after CRT were offered radical resection only. Therefore, our CRT group was highly selective and homogeneous, in contrast to other series^19,20 where patients with various grades of response to CRT were included, a feature that might contribute to a high complication rate.

Because suturing of the rectal wall defect during TEM is one of the most important and challenging steps in the procedure, precision in executing tensionless sutures with good approximation is the key to successful closure of the rectal wall defect. Using different and "non-endoscopic" methods of wound suturing during TEM could influence the high wound dehiscence rate reported in some series. ²⁰ Technical aspects and level of the surgeon's skills are considered the most important determinants of clinical wound failure and leakage after colorectal surgery and TEM. Barendse et al. ³⁶ recently demonstrated that the surgical learning curve of TEM affected the complication rate of the procedure. Leakage rates have also been used as an indicator of surgical quality. ³⁷

Although we used interrupted absorbable sutures to close the rectal wall defects in all TEM cases, some reports ³⁸ have argued against the need for closing the rectal defect and came to conclusion that there was no difference in outcome between patients in whom the rectal wall defect was closed and those in whom the defect was left open after full-thickness LE of an extraperitoneal rectal lesion. Therefore, wound dehiscence as a major complication in TEM would be avoided simply by leaving the defect open in certain cases.

Paradoxically we observed fewer overall complications in our CRT group. One explanation could be the fact that the CRT patients had smaller rectal lesions (mean, 2.1 cm) compared with the non-CRT patients (mean, 2.9 cm). Also, patients in the non-CRT group had rectal masses located higher in the upper rectum, which may explain complications such as peritoneal entry and the need for a protective ileostomy. The lower rate of complications following TEM in the CRT patients might also be associated with the TEM procedures being reserved for resection of benign or early-stage carcinomas and that, only later, with greater experience, did we perform them on patients after CRT. Neoadjuvant CRT initially causes tissue edema and subsequent fibrosis in the rectal wall and mesorectum. Because this event is time dependent, it is important to perform surgery after enough time has elapsed since completion of the CRT. We performed TEM 9 weeks (range, 5–15 weeks) after the last CRT session, a feature that could have also influenced the results.

The main concern regarding TEM technique after CRT is local recurrence because regional lymph nodes are not removed, and the accurate pathologic stage cannot be assessed. This may potentially obscure the need for additional therapy that could reduce the chances of local and regional relapses in the future. Nevertheless, a 7.7% local recurrence in our CRT group is reasonable considering the patients with residual disease and pathological incomplete response who did not undergo formal rectal resection. However the local recurrence rates in the CRT and non-CRT group were in accordance with other reports.^39–42

Although TEM is still considered the primary form of treatment for rectal adenomas and early and selected rectal cancer, this strategy did not completely replace the radical surgery for rectal cancer with complete response after neoadjuvant CRT. However, benefits in terms of reduced morbidity and improved quality of life after TEM may outweigh the small increase in the risk of recurrence and the need for salvage radical surgery in the event of residual tumor. ⁴³ Therefore, in cases of cCR after neoadjuvant CRT, we propose that TEM should be considered as a means for providing a diagnostic full-thickness biopsy for allowing a pathological evaluation of the specimen and subsequently guide management decisions. Specifically, choosing between radical surgery and close surveillance should be based on pathological evaluations and tumor characteristics.

Our study had limitations in several selection criteria, including a small number of CRT patients and long accrual periods with different diagnostic and therapeutic techniques (particularly in the CRT group), as well as its retrospective nature.

Careful patient selection, appropriate timing of TEM after the completion of CRT, and a meticulous technique of properly placed suturing for assuring good approximation without tension may contribute to the reduction of the morbidity in this group of patients.

TEM can be performed safely in selected rectal cancer patients with cCR following neoadjuvant CRT without any increase in wound dehiscence, postoperative morbidity, or major complications.