Cancer of the Rectum

Externally, the upper extent of the rectum can be identified where the tenia spread to form a longitudinal coat of muscle. The upper third of the rectum is surrounded by peritoneum on its anterior and lateral surfaces but is retroperitoneal posteriorly without any serosal covering. At the rectovesical or rectouterine pouch, the rectum becomes completely extra-/retroperitoneal. The rectum follows the curve of the sacrum in its lower two-thirds. It enters the anal canal at the level of the levator ani. The anorectal ring is at the level of the puborectalis sling portion of the levator muscles.

The location of a rectal tumor is most commonly indicated by the distance between the anal verge, dentate (pectinate or mucocutaneous) line, or anorectal ring and the lower edge of the tumor. These points of reference are all different for different individuals. Also, these measurements differ depending on the method of measurement. This can be important clinically, as the measurement from a flexible endoscopy can substantially overestimate the distance to the tumor from the anal verge or other landmark. The distance from the anal sphincter musculature is clinically of more importance than the distance from the anal verge, as it has implications for the ability to perform sphincter-sparing surgery. The lack of a peritoneal covering over most of the rectum is a major reason for the higher risk of local failure after primary surgical management of rectal cancer compared to colon cancer. The mesorectum is usually used as the structure to define the extent of a total mesorectal excision (TME), with most of the perirectal fatty tissue and perirectal lymph nodes (LN) contained within its boundaries.

Lymphatic Drainage

The lymphatic drainage of the upper rectum follows the course of the superior hemorrhoidal artery toward the inferior mesenteric artery. Lymph nodes that are above the midrectum and therefore drain along the superior hemorrhoidal artery are often part of the mesentery that is removed during resections of the intraperitoneal portion of the colon. Lesions that arise in the rectum below approximately 6 cm are in a region of the rectum that is drained by lymphatics that follow the middle hemorrhoidal artery. Nodes involved from a cancer in this region can include the internal iliac nodes and the nodes of the obturator fossa. These regions deserve particular attention during the resection and irradiation of lesions in this location. When lesions occur below the dentate line, the lymphatic drainage is via the inguinal nodes and external iliac chain, which has major therapeutic implications, especially for the radiation fields. The corollary of this high risk of inguinal node involvement for the very low-lying tumors is that tumors located above the dentate line are at low risk of inguinal node involvement, and these nodes as well as the external iliacs do not need to be treated.

Bowel Function

Fecal continence is maintained through the function of both the sphincter mechanism and the preservation of the normal pelvic floor musculature, which creates a neorectal angle or rectal sling. The pelvic floor is composed of the levator ani muscles, which separate the pelvis from the perineum and ischiorectal fossa. The urethra, vagina, and anus pass through the levator muscles.

Preservation of fecal continence during surgery for rectal cancer is therefore dependent on a thorough understanding of the anatomic relationships of the musculature and the sphincter mechanism. Maintenance of the sphincter apparatus without preservation of the muscular angles will not have the desired result. These anatomic constraints, especially with respect to lateral margins, make the use of adjuvant chemotherapy and radiation therapy critical to a successful surgical outcome. This is true from both an oncologic as well as a bowel function perspective.

Autonomic Nerves

The preservation of both bladder and sexual function depends on the surgeon’s understanding of the autonomic nerve supply to the pelvic organs.^1,2 The hypogastric plexus is formed from the sympathetic trunks as they converge over the sacral promontory. These sympathetic nerves are found beneath the pelvic peritoneum along the lateral pelvic sidewalls lateral to the mesorectum. The second, third, and fourth sacral nerve roots give rise to parasympathetic fibers to the pelvic viscera. The parasympathetic fibers proceed laterally as the nervi erigentes to join the sympathetic fibers at the site of the pelvic plexus that is just lateral and somewhat anterior to the tips of the seminal vesicle in men.^1,2 In order to preserve these structures and, therefore, sexual and bladder function, a sharp rather than a blunt technique should be used to dissect the mesorectum.^3–6

STAGING

Standard clinicopathologic staging is the best indicator of prognosis for patients with rectal cancer. For rectal cancer, it is increasingly common to use clinical staging as the basis for the decision to initiate neoadjuvant chemoradiation therapy. Therefore, the accuracy of that initial staging is critically important, both for management and for prognosis. There have been a large number of studies that have evaluated other prognostic markers, including pathologic, socioeconomic, and molecular, as described more fully in Chapter 57. However, even though many of these appear to have prognostic value, there are none that are commonly used to define management. This is related to the large number of tests that could be used, the lack of standardization of these tests, as well as the lack of knowledge as to how to incorporate them into the patient management scheme. The molecular marker that has engendered the most interest is the deletion of 18q.⁷ These markers have been fully reviewed elsewhere.^8,9

The staging system that should be used in the evaluation of patients with rectal cancer is the American Joint Committee on Cancer/International Union Against Cancer TNM (tumor, node, metastases) staging system (fully described in Chapter 57), which has been recently revised to subcategorize patients with stage III (node-positive) tumors. The Dukes staging system or its multiple modifications has been used for many years, but provides less information than the TNM system and should not be used. There have been gradual changes in the TNM system that primarily reflect the stage grouping rather than the system itself. The other systems should be acknowledged for their historical interest and for initially defining many of the high-risk factors for this disease.

Patients now often have both a clinical (preoperative) stage, which may define the need for neoadjuvant therapy, and a pathologic (postoperative) stage. Initial therapy with chemoradiation can produce substantial downstaging (approximately 15% of patients will have a pathologic complete response, and as many as 40% in those with more favorable tumors). While some believe that the degree of response to neoadjuvant therapy should alter subsequent treatment, and this is in fact an area of active investigation (see later discussion), the current standard of care dictates that all surgical planning and adjuvant therapy be determined based on the initial clinical stage regardless of tumor response. This guideline is based in part on the idea that a good tumor response locally to chemoradiation does not translate into reduced risk of having micrometastatic disease, and thus does not lessen the need for adjuvant postoperative chemotherapy. Whether this will continue to be true in the face of newer data and more aggressive neoadjuvant regimens remains to be seen. Numerous studies have indicated that tumor response to neoadjuvant therapy is an important predictor of multiple oncologic end points for patients completing the full course of multimodal therapy. Patients with pathologic complete response in particular demonstrate excellent long-term results, with local recurrence rates as low as 0.7%, and significantly improved disease-free survival (DFS) and overall survival (OS) compared to nonresponders at 5 years (odds ratio [OR] = 3.28 and OR = 4.33, respectively).^10–12 It is unclear at this time whether such a favorable outcome can be maintained should the course of treatment be altered based on postneoadjuvant reassessment.

Although it is not standard practice to alter treatment based on local response to neoadjuvant therapy, preoperative restaging prior to surgery may still be valuable not only as a prognostic predictor but also for detecting interval metatastic progression. Multiple studies recommend repeating serum carcinoembryonic antigen (CEA) levels, for example, between chemoradiation and surgery as this value as well as the pre- to posttreatment ratio may be more important in predicting survival than the initial measurement.^13–15 In addition, Ayez et al.¹⁶ advocate restaging with chest and abdominal computed tomography (CT), as this changed management in 12% of their patients, and spared 8% from undergoing noncurative rectal surgery, due to new findings of progressive metastatic disease.¹⁶

The major change that has occurred in the newest version of the staging system is the acknowledgment that both the T stage and the N stage have independent prognostic importance for local control, DFS, and OS.^17,18 Thus, for patients with N0 and N1 tumors viewed separately, the extent of the primary tumor in the rectum is of additional prognostic importance. Patients with T1-2N1 tumors have a relatively favorable prognosis and an outcome superior to that of other stage III patients. In fact, patients with T3N0M0 disease (stage II) have outcomes slightly inferior to those with T1-2N1M0, demonstrating the independent prognostic importance of T stage. These distinctions may allow future decisions to be more individualized as to the adjuvant therapy required.

Although at one level staging is very straightforward, the actuality of proper staging is much more difficult as it relies on multiple quality control issues that can mislead the clinician regarding proper therapy. For instance, it has been well demonstrated that for patients with colon and rectal cancer who are pathologically staged as N0, the prognosis is markedly improved for those in whom more than 12 to 14 nodes were identified by the pathologist compared with those in whom fewer nodes were identified.¹⁹ This could be a surgical issue (fewer nodes were removed) or a pathologist issue (fewer nodes were identified), but it suggests that many patients were inappropriately understaged, which could result in inappropriate therapy. Others have shown that staging accuracy continues to improve as the pathologist recovers more nodes, with accuracy leveling off at approximately 12 to 20 nodes recovered.^20,21 (See discussion in Chapter 57). In rectal cancer, however, N staging presents a particular challenge as there are often fewer LNs in the specimen and preoperative radiotherapy is thought to reduce that number even further.^22–24 In fact, one recent report suggests that LN harvests <12 in pretreated specimens may be a marker of high tumor response and improved rather than compromised oncologic outcome. In this study of 237 patients, local recurrence rates were significantly higher in the LN >12 group as compared to those with “inadequate” LN retrieval (11% versus 0%; p = 0.004).²⁵ As with colon cancer, the percentage of positive nodes is likely of greater prognostic importance than total LN number (M. Meyers, 2007, personal communication).^26–28 The same issue relates to T-stage determination. If the pathologist does not look carefully for evidence of extension of tumor through the muscularis propria, the patient can be understaged, resulting in inappropriate treatment. Close or positive circumferential margins are a poor prognostic factor, which can only be found if the pathologist assiduously evaluates the radial margins.^29,30

The standard staging procedure for rectal cancer entails a history, physical examination, complete blood cell count, liver and renal function studies, as well as CEA evaluation. The routine laboratory studies are quite insensitive to the presence of metastatic disease, but they are usually ordered as a screen of organ function prior to surgery or chemoradiation therapy. High CEA levels are associated with poorer survival (see Chapter 57) and give an indication as to whether follow-up CEA determinations are likely to be useful. A careful rectal examination by an experienced examiner is an essential part of the pretherapy evaluation in determining distance of the tumor from the anal verge or from the dentate line, involvement of the anal sphincter, amount of circumferential involvement, clinical fixation, sphincter tone, and so forth, and has not been replaced by imaging studies or endoscopy. Colonoscopy or barium enema to evaluate the remainder of the large bowel is essential (if the patient is not obstructed) to rule out synchronous tumors or the presence of polyposis syndromes.

Local staging is completed with one of two imaging modalities, endorectal ultrasound (EUS) or pelvic magnetic resonance imaging (MRI). Each provides similar overall accuracy in T and N determination, and each has its advantages as well as drawbacks. The decision of which to use generally depends on local institutional expertise and resource availability.

EUS defines five interface layers of the rectal wall: mucosa, muscularis mucosa, submucosa, muscularis propria, and perirectal fat, as shown in Figure 60.2. Rectal tumors are generally hypoechoic and disrupt the interfaces depending on the level of tumor extension. The accuracy of EUS depends heavily on the experience and skill of the operator. In experienced hands, EUS has an overall accuracy rate for T stage of 75% to 95% with an overstaging of approximately 10% to 20% in T2 disease because of an inability to distinguish a desmoplastic response and postbiopsy changes from local tumor invasion, and approximately a 10% rate of understaging because of an inability to detect microscopic tumor extension.^31–33

EUS is less accurate in determining N stage than for T stage, with an overall accuracy rate of 62% to 83%.^31,32 Understaging occurs because many nodal metastases from rectal cancer are small, even micrometastatic, and not easily detected by EUS. In addition, some nodes are located beyond the range of the ultrasound transducer and thus cannot be seen during the procedure. Overstaging is often related to an inflammatory response, perhaps secondary to previous biopsy or manipulation. EUS is not accurate for determining tumor regression after preoperative radiation and chemotherapy, as inflammatory changes and scarring can persist in the rectal wall or in perirectal soft tissue and may not reflect persisting tumor. Newer ultrasound techniques, such as three-dimensional ultrasound, are being explored but have not yet made it into standard practice.

Endorectal coil MRI allows discernment of the layers of the bowel wall and is similar in accuracy to EUS. Thin-section pelvic MRI with a surface coil also allows one to visualize the mesorectal fascia and thus to predict the likely distance of the surgical resection margin when performing a TME. The MERCURY Study Group confirmed this key advantage of MRI in their landmark 2005 multicenter trial, where specificity for predicting clear margins in 408 patients with varying stages of rectal cancer was 92%.³⁴ Although there has been great interest in this technique since then, follow-up studies still show a disappointing overall accuracy for T and N staging, which fails to surpass EUS in experienced hands. In one study of 96 patients who had MRI followed by TME, of 22 patients classified as having T2 disease on MRI, 3 had T1 and 6 had T3 tumors. Of 61 patients classified as having T3 disease on MRI, 8 had T2 tumors and 2 had T4. Thus, 6 of 22 (27%) patients who might have benefited from preoperative therapy for T3 disease would not have received that therapy. Eight of sixty-one patients (13%) would have received preoperative treatment inappropriately based on the MRI T stage.³⁵ For nodal status, 8 of 33 MRI-positive nodes were clinically negative, and 7 of 57 MRI-negative nodes were pathologically positive.³⁶ The presence of nodal disease identified by MRI is also primarily determined by size, so the accuracy is similar to that of CT (<80%), although defining node positivity based on irregular border or mixed signal intensity could help improve sensitivity and specificity.³⁵

While the accuracy of MRI in determining T and N stage is imperfect, newer studies have focused on other radiographic features that may prove more relevant to prognosis and treatment planning than the traditional American Joint Committee on Cancer classification. In addition to defining the circumferential resection margin (CRM) of a low rectal cancer, high-resolution MRI can be used to predict tumor regression grade (TRG) after neoadjuvant therapy. TRG in the surgical specimen is a measure of response to preoperative chemoradiation and has been shown to correlate strongly with OS and DFS. In the first prospective study to address MRI-predicted TRG, Shihab et al.³⁷ found this too was significantly associated with long-term outcomes. A prognostic role was also demonstrated for pretreatment MRI, specifically in the characterization of tumor invasion into the pelvic floor muscles. Based on these results, the authors postulate that MRI-defined factors may be extraordinarily useful for modifying treatment in both the pre- and postneoadjuvant settings.³⁷ Patel et al.³⁸ offer a similar conclusion in their subgroup analysis from the original MERCURY study, which likewise found MRI-predicted TRG and CRM to be significant prognostic markers.

Two studies have taken this issue a step further by investigating exactly how MRI parameters can and should alter therapy. In another extension of the MERCURY trial, Taylor et al.³⁹ identified patients with “good prognosis” MRI (as defined by predicted negative CRM, absence of extramural venous invasion, and T2/T3a/T3b regardless of N stage) and referred them directly for TME resection without chemoradiation. Survival and recurrence outcomes were highly favorable, suggesting that early MRI can improve patient stratification for more selective and appropriate targeting of preoperative therapy.³⁹ In the postneoadjuvant setting, MRI can be used to identify poor responders who may require alternative treatments or more radical resection such as the extralevator abdominoperineal approach described by Shihab et al.⁴⁰

Finally, pelvic MRI may also help predict which patients are at increased risk for distant synchronous metastases and would therefore benefit from more extensive pretreatment imaging such as positron emission tomography (PET)/CT or liver MRI. Hunter et al.⁴¹ found that adverse features demonstrated on pelvic MRI (extramural venous invasion, extramural spread of >5 mm or T4, involved CRM or intersphincteric plane for low tumors) were significantly associated with a higher incidence of distant metastases (OR = 6.0; p <0.001). The authors recommend using MRI-based risk stratification to identify patients who may benefit not only from more meticulous staging but also from more aggressive treatment regimens.⁴¹

M staging for rectal cancer is determined in the same way as colon cancer: with a baseline CT scan to evaluate the chest, abdomen, and pelvis.⁴² There has been much debate about the relative value of CT versus MRI or PET, particularly in assessing the liver, without any clear resolution. This decision depends heavily on the institutional expertise and the equipment available.

CT has an overall sensitivity of 70% to 85%, which might be improved with multidetector-improved CT technology, although the data do not yet prove that contention.⁴³ MRI is superior in characterizing liver lesions and distinguishing cysts and hemangiomas from tumor, especially with the use of enhancement with gadolinium or other agents.⁴⁴ PET with [¹⁸F]fluorodeoxyglucose shows promise as the most sensitive study for the detection of metastatic disease in the liver and especially in abdominal LNs for which CT and MRI are relatively insensitive. In addition, a meta-analysis of whole-body PET showed a sensitivity of 97% and a specificity of 76% in evaluating for recurrent colorectal cancer.⁴⁵ However, PET is not standardly used in preoperative staging, or recommended by National Comprehensive Cancer Network guidelines, and the incremental gain from routine PET scan appears to be small.⁴⁶ A 2013 study has re-emphasized this point, reporting that preoperative PET/CT had no impact on disease management in 96.8% of enrolled patients and advocating against its routine use for primary staging.⁴⁷ PET is probably most valuable in restaging patients with recurrence or suspected recurrence to detect additional metastatic sites prior to attempted resection of metastatic disease.

SURGERY

The surgical management of primary rectal cancer presents unique problems for the surgeon based in large part on the anatomic constraints of the pelvis. The primary goal of achieving a complete oncologic resection must be balanced with the desire for optimal nerve and sphincter preservation, which can be quite challenging in such a confined space.

Stage I

The treatment of early-stage rectal cancer can be confusing as there are many approaches that can be used, and patient selection is critical to outcome. In addition, the risk of nerve injury and damage to the anal sphincter is substantial for low-lying tumors and must be taken into consideration, along with the desire not to have a permanent colostomy for early-stage disease. Thus, the options for these patients are primarily those of local therapies without abdominal surgery, abdominal resection of the rectum with anastomosis and retention of the anal sphincter, and abdominal-perineal resection. The last two options are discussed in detail in “Stages II and III Rectal Cancer.”

Small early-stage lesions of the rectum that are diagnosed on physical examination or by colonoscopy/proctoscopy can often be managed with local resection. Local resection can be performed colonoscopically (as described in the Chapter 57), or lesions can be removed via a transanal excision with the patient positioned in a prone or lithotomy position. Appropriate retractors can provide visualization, and resection should extend into the perirectal fat with a surrounding margin of normal tissue.⁴⁸ For selected T1 and T2 lesions without evidence of nodal disease, transanal excision often provides an adequate resection of the primary tumor mass and can spare the patient the morbidity of a more extensive rectal resection. However, it does not stage the nodal drainage areas and therefore cannot provide as complete staging and management of the tumor as a definitive resection. In the effort to minimize the risk of locoregional failure, criteria for local excision have been established: the tumor must be within 8 to 10 cm of the anal verge, be well or moderately well differentiated, encompass <40% of the circumference of the bowel wall, and contain no evidence of lymphovascular invasion on biopsy. For T2 lesions, local resection should be followed by adjuvant chemoradiation. While these criteria are not strongly evidence-based (and are evolving along with surgical technology), a growing body of literature supports this approach particularly for T1 lesions. In a review of 677 T1 and T2 cancers after TME, Saraste et al.⁴⁹ identified three significant risk factors for LN invasion (and hence relative contraindications for local excision): T2 stage (OR = 2.0), poor differentiation (OR = 6.5), and vascular infiltration (OR = 3.4) with likelihood of LN positivity ranging from 6% to 78% depending on how many were found. Further support for these criteria comes from a study of 25 high-risk T1 rectal cancers, half of which were treated by transanal excision only (due to comorbidities or patient refusal to undergo resection) and the remainder with immediate conventional reoperation after local excision. Local recurrence was significantly higher in patients undergoing local excision only (50% versus 7.7%, mean follow-up 62 months), and there was a trend toward decreased 5-year survival (63% versus 89%). There were no differences in age, gender, or tumor characteristics between the two groups.⁵⁰ On the other hand, for low-risk T1 lesions in the prospective phase 2 Cancer and Leukemia Group B study, local excision alone was associated with low recurrence and good survival rates that remained durable with long-term follow-up. For T2 lesions, however, even with adjuvant therapy, the role of local excision is less clear, as Saraste et al.⁴⁹ would predict, these were associated with higher recurrence rates.⁵¹ Whether the addition of neoadjuvant therapy might be helpful is a focus of multiple investigations. The American College of Surgeons Oncology Group has published preliminary results from its recently completed phase 2 trial of neoadjuvant capecitabine, oxaliplatin, and radiation therapy followed by local excision for ultrasound T2 tumors (ACOSOG Protocol Z6041).⁵² The authors report that 49 of 77 patients were downstaged and 44% achieved a complete pathologic response (pCR). There was one positive margin and one patient with a positive node. Rates of treatment-related toxicity and perioperative complications were high, however, requiring dose reduction and potentially compromising response. Follow-up trials are planned to improve upon the therapeutic ratio of this approach and better evaluate long-term efficacy.⁵³ More long-term results are reported from another prospective trial that supports the role of local excision following neoadjuvant therapy in selected T2N0 lesions with favorable features. Lezoche et al.⁵⁴ randomized 100 patients to either endoluminal resection or to laparoscopic or open TME, following neoadjuvant chemoradiation. Downstaging and pCR rates were similar in both groups, occurring in 51% and 28% of patients, respectively. With a mean follow-up of 9.6 years, oncologic outcomes were also essentially equivalent—with similar local recurrence rates and incidence of distant metastases (8% versus 6% and 4% versus 4%, respectively) and no difference in DFS.⁵⁴

Performing a good transanal excision requires substantial technical expertise as the surgeon must retain control over the primary tumor and obtain adequate mucosal margins as well as deep resection into the perirectal fat. Once removed, the tumor must be well laid out for the pathologist so that all relevant margins can be properly evaluated. There is some experience using preoperative radiation therapy and chemotherapy for small lesions, but care must be taken to have the site of the primary tumor well marked with a tattoo if this approach is taken, as excellent regression could make identification of the primary site difficult.

Newer techniques for transanal excision, including transanal endoscopic microsurgery (TEMS) and transanal minimally invasive surgery, have recently gained popularity based on improved visualization of the lesion. TEMS makes use of a standard laparoscopic light source and monitoring system combined with specialized instruments and scopes. The technique allows for videoscopic magnification and the placement of instruments through an operating sigmoidoscope. TEMS and its counterpart transanal minimally invasive surgery, which uses the more basic single port laparoscopic technology, may be applied, in general, to the same patients who are candidates for traditional transanal resection. However, these methods are most useful for excising more proximal lesions that are beyond the reach of standard surgical instruments and too large for removal through a colonoscope. Preliminary data supports the role of TEMS in both benign and early-stage malignant lesions with improved margin negativity and DFS compared to transanal resection for T1 and T2 lesions in a recent report.⁵⁵ Another meta-analysis found significant reductions in morbidity and mortality compared to conventional surgery and equivalent 5-year survival rates for T1 tumors.⁵⁶ Studies that include T2 lesions and selective use of adjuvant therapy have demonstrated 5-year OS and cancer-specific survivals over 90%, with recurrence rates between 4% to 9%.^57,58 Moreover, the TEMS procedure fairs quite favorably with respect to long-term quality of life and functional outcome as most defecatory parameters return to baseline by 5 years, according to prospective data.⁵⁹ Other reports are less encouraging with recurrence rates following TEMS resection as high as 30%,⁶⁰ and therefore close endoscopic surveillance is recommended.

Stages II and III Rectal Cancer

The primary treatment of patients with stages II and III rectal cancer (T3-4 and/or node-positive) is surgical. However, in contrast to the treatment of patients with stage I disease, there is a strong body of information to suggest that combined modality therapy with radiation therapy and chemotherapy should be used in conjunction with surgical resection. This conclusion is based on both patterns of failure data, which demonstrate a substantial incidence of local, regional, as well as distant disease failure, and the fact that this incidence of tumor recurrence at all sites is decreased with the use of trimodality therapy.

The desire when performing a resection for rectal cancer is to preserve intestinal continuity and the sphincter mechanism whenever possible while still maximizing tumor control. Therefore, careful preoperative screening is crucial in the determination of the location of the lesion and its depth of invasion. As previously described, it is convenient to think of the rectum as divided into thirds for the purposes of the evaluation and preoperative determination of the surgical approach for resection. The upper third of the rectum is often considered the region of large intestine from the sacral prominence to the peritoneal reflection. These lesions are in almost all cases managed with a low anterior resection in much the same way as a sigmoid colon cancer (see Chapter 57). An adequate 1- to 2-cm distal mucosal margin can be achieved for these lesions well above the sphincter mechanism, and intestinal continuity can be restored using either a hand-sewn technique or a circular stapling device inserted through the rectum.^61,62

Tumors in the middle and lower thirds of the rectum can be considered as lying entirely below the peritoneal reflection. The resection of these tumors can be challenging because of the confines of the pelvic skeletal structure, and the ability to perform a resection with an adequate distal margin is significantly influenced by the size of the lesion. Nevertheless, tumors of the middle third of the rectum in most cases can be safely resected with a low anterior resection, with restoration of intestinal continuity and preservation of a continent sphincter apparatus.

Lesions in the distal third of the rectum, defined as those within 6 cm of the anal verge, can present the greatest challenge to the surgeon with respect to sphincter preservation. This is often influenced by the extent of lateral invasion of the lesion into the muscles of the sphincter apparatus and how close distally the tumor is to the musculature of the anal canal. The abdominal perineal resection (APR) has historically been considered the standard treatment for patients with rectal cancers located within 6 cm of the anal verge. This procedure requires a transabdominal as well as a transperineal approach with removal of the entire rectum and sphincter complex. A permanent end colostomy is created and the perineal wound either closed primarily or left to granulate in after closure of the musculature.

Although an APR is associated with a relatively low rate of local recurrence, it is not without the obvious problems of the need for a permanent colostomy and loss of intestinal continuity and sphincter function. Therefore, intense interest has been focused on developing approaches to the resection of tumors in the distal third of the rectum that would both avoid local regional recurrence and preserve intestinal continuity and sphincter continence.

Traditionally, tumors within 1 to 2 cm of the dentate line—that is, those that can be removed with at least a 1-cm distal margin—have been considered candidates for sphincter preservation and restoration of intestinal continuity via a coloanal anastomosis, which is commonly protected by a diverting loop ileostomy that can be reversed in 6 to 12 weeks.^63,64 Newer data suggest that when TME and preoperative radiotherapy are routinely employed, even smaller margins are acceptable without oncologic compromise, as long as they are microscopically negative.⁶⁵ In fact, one of the advantages of neoadjuvant therapy is thought to be an increase in sphincter-sparing procedures due to reduction in tumor bulk, which would normally preclude identification of this slight but critical margin.^63,66 A recent systematic literature review identified seven studies addressing this topic, most of which implemented pre- or postoperative radiotherapy, and three of which reported results related to a margin of <5 mm. There were no statistically significant differences in local recurrence rates regardless of margin status. This data contributes to the growing evidence that a 1 cm (or even 5 mm) margin may be unnecessary and, more importantly, that strategies employed solely to achieve this margin (such as an APR or intersphincteric resection [ISR] for distal T1 lesions) may in fact be unnecessary as well.⁶⁷

While controversial in the United States, ISR has been described extensively abroad as a method involving at least partial resection of the internal sphincter designed to improve margin status without sacrificing sphincter function.⁶⁸ Recently, a large systematic review addressed the efficacy of this approach, identifying 14 (mostly retrospective) studies with 1,289 patients who underwent both open and laparoscopic ISR. Median follow-up was 56 (range, 1 to 227) months. Overall oncologic outcomes did not appear to be compromised with R0 resection achieved in 97% and a mean local recurrence rate of 6.7% (range, 0% to 23%). In addition, mean 5-year OS and DFS rates were 86.3% and 78.6%, respectively. Functional outcomes, however, were widely variable with only 51.2% of patients reporting “perfect continence,” while an average of 29.1% experienced fecal soiling, 23.8% incontinence to flatus, and 18.6% complained of urgency.⁶⁹

It has been postulated that neoadjuvant chemoradiation, while improving locoregional control and rates of margin-negative resection, has a deleterious effect on long-term functional outcomes, particularly after surgery for ultralow tumors. However, a recent multivariate analysis did not support this in ISR cases, finding the only significant predictors of continence were distance of the tumor from the anal ring and distance of the anastomosis from the anal verge. There was also no difference with age or extent of internal sphincter resection.⁷⁰ Another report did find significant functional differences when comparing partial ISR (resection above the dentate line), subtotal ISR (resection at the dentate line), and total ISR (resection from the intersphincteric groove). Patients with more extensive sphincter resection had higher fecal incontinence scores, more frequent nocturnal leakage, and more problems with discrimination. In addition, manometric studies at 12 months showed greater reductions in mean resting pressure. Overall though, quality of life was maintained in the majority of patients and function improved over time in both studies.⁷¹

Chemoradiation should be used preoperatively when performing sphincter-preserving resections for T3 or T4 rectal lesions or for any node-positive disease stages II or III. There is some evidence that preoperative radiation results in less morbidity than postoperative radiation therapy when a coloanal anastomosis is planned. In a study of 109 patients treated with a low anterior resection and a straight coloanal anastomosis, those receiving preoperative radiation therapy had a lower incidence of adverse effects on anal function than those receiving postoperative radiation.⁷² The authors attributed this to sparing of the neorectum from these effects. Relative benefits and outcomes for preoperative chemoradiation versus postoperative chemoradiation will be discussed in detail in following sections.

Total Mesorectal Resection

The goal of the resection of rectal tumors is the removal of the tumor with an adequate margin as well as removal of draining LNs and lymphatics to properly stage the tumor and to reduce the risk of recurrence and spread. For lesions in the intraperitoneal colon, the lymphatics and vascular supply are found in the mesentery associated with that region of bowel.

In the rectum, the mesorectum is the structure that contains the blood supply and lymphatics for the upper, middle, and lower rectum. Most involved LNs for rectal cancers are found within the mesorectum, with T1 lesions associated with positive LNs in 5.7% of cases, T2 lesions having positive LNs in 20% of cases, and T3 and T4 lesions having positive LNs in 65% and 78% of cases, respectively.⁷³

The anatomy and approach to mesorectal excision is depicted in Figure 60.3. This operation involves a sharp dissection occurring in an avascular plane between the fascia propria of the rectum and the presacral membrane, beyond the region where most of the nodes are located. After a TME, the specimen is typically shiny and bilobed in contrast to the irregular and rough surface after a blunt dissection, where much of the mesorectal fat is left behind. TME attempts not only to clear involved LNs but also to adequately manage the radial margins of the rectal tumor. These radial margins have been shown to be more important with respect to the risk of local regional recurrence than the distal mucosal margin.^66,74 Distal mucosal margins of ≥1 cm are adequate for local control; however, the margin on the mesorectum should extend beyond the distal mucosal margin in order to ensure a successful surgical outcome.^64,66 Numerous studies have demonstrated the benefit of TME, and it is now considered the standard of care for the surgical management of middle and lower third rectal cancers.^5,75–77 Although some studies have suggested that an adequate TME might in and of itself be sufficient management for T2 and T3 rectal cancers, the majority of the literature still supports the use of adjuvant chemoradiation for stages II and III disease even when combined with TME.

Large studies of proctectomy with TME have demonstrated a reduction in the overall incidence of local recurrence to <10%.⁴ The consequences of TME can be impairment in erectile and bladder function because of disruption of parasympathetic nerves that are located in proximity to the mesorectum. Several authors have stressed the importance of the experience of the surgeon performing the procedure, and some have suggested specific techniques for monitoring modalities that can be used during this procedure to minimize morbidity.^5,6 A careful understanding of the anatomy and adequate visualization during sharp dissection will help in minimizing injury to the parasympathetic nerves and the consequent morbidity.^3,4

Adequate visualization in the deep pelvis can often be a challenge. This may be a situation where the visual magnification and ability to enter tight spaces that are unique to the laparoscopic approach may be an advantage. Several groups have demonstrated the feasibility of laparoscopic TME for low rectal cancer as part of a sphincter-preserving operation.^78–80 Some of the larger series, while demonstrating that TME using laparoscopic techniques can be performed safely, do not have adequate follow-up to demonstrate whether there were any oncologic disadvantages to such an approach. Unfortunately, the prospective random assignment trial conducted in the United States to evaluate the role of laparoscopic surgery for colon cancer excluded patients with low rectal lesions. In addition, subgroup analysis from the UK CLASICC trial reported a 34% conversion rate and double the frequency of positive margins compared to open cases (12% versus 6%), prompting the authors to advise against routine practice of laparoscopic proctectomy outside of the research setting.⁸¹ While these results have raised serious concerns regarding oncologic outcomes, follow-up reports are more encouraging. Multiple single-institution experiences have now been published demonstrating not only similar surgical parameters (margin status, LN harvest numbers) but also comparable recurrence and 5-year survival data.^82–86 Furthermore, in a study based on National Surgical Quality Improvement Program data from 5,420 patients, Greenblatt et al.⁸⁷ reported significant short-term advantages to laparoscopy, including decreased length of stay (5 days versus 7 days; p <0.0001) and 30-day morbidity (20.5% versus 28.8%; p <0.0001). Smaller randomized trials as well as two recent large meta-analyses of randomized controlled trials also support the oncologic equivalence of the two approaches, although short-term benefits are mixed.^88–90 In 10-year follow-up data from a pooled analysis of three randomized controlled trials (including 136 laparoscopic and 142 open cases), continued long-term oncologic safety of the laparoscopic approach was demonstrated with no significant differences compared to open in terms of locoregional recurrence (5.5% versus 9.3%), cancer-specific survival (82.5% versus 77.6%), or OS (63% versus 61.1%). Additionally, there was a trend toward lower recurrence among stage III patients in the laparoscopic group (17.7% versus 25.3%), though this did not reach statistical significance.⁹¹ Lujan et al.⁹² also reported similar rates of local recurrence and OS in a prospective cohort of 4,405 patients but found decreased complication rates (38.3% versus 45.6%) and improved oncologic parameters with laparoscopy, including decreased margin involvement and more complete TME. Finally, in a smaller study by Westerholm et al.,⁹³ laparoscopic surgery was found to be an independent predictor of DFS on multivariate analysis with 5-year DFS rates of 50.3% compared to 71.0% after open resection. Definitive recommendations await the results of three ongoing multicenter phase 3 randomized trials: the European COLOR II, the Japanese JCOG 0404, and the ACOSOG Z6051 from the United States.⁹⁴

While laparoscopic TME may be technically feasible, it requires a high level of expertise and can be particularly challenging to perform within the confines of a deep and narrow pelvis. More recently, robotic technology has been applied to rectal dissection, overcoming many of the limitations associated with conventional laparoscopy including limited dexterity, inadequate visualization, and tremor. Robotic surgery offers the advantages of a stable, three-dimensional image, enhanced ergonomics and articulating instruments with seven degrees of freedom, in addition to operator-controlled camera and retraction.⁹⁵ Embraced by urologists and gynecologists over the past decade, this technology is ideally suited to pelvic procedures and has the potential to yield enhanced oncologic and functional outcomes in rectal cancer surgery as well. Limited studies so far have demonstrated feasibility and acceptable short-term outcomes.^95–99 In a case-control analysis of 118 patients undergoing laparoscopic versus robotic resection, Kwak et al.¹⁰⁰ reported no differences in surgical oncologic parameters, postoperative complications, or recurrence rates at a median of 15 months follow-up. When compared to open TME in another case-matched study, the robotic approach was superior in terms of LN harvest, distal margin length, blood loss, and length of stay.¹⁰¹ Other potential benefits include decreased conversion rates in three large meta-analyses as well as a trend toward reduced anastomotic leaks and CRM positivity with complete autonomic preservation in a recent systematic review of 1,549 patients.^102–105 Whether this data will translate into meaningful long-term advantages that justify the significantly higher cost of this approach remains to be seen.

The Robotic versus Laparoscopic Resection for Rectal Cancer (ROLARR) trial is a prospective, randomized, controlled, multicenter superiority trial that began enrollment in 2010, with a target recruitment of 400 patients. It will evaluate differences in conversion rates; CRM positivity, 3-year local recurrence, DFS and OS, as well as operative morbidity and mortality, quality of life, and cost-effectiveness. Investigators also wish to explore the purported clinical benefits of robotics including preservation of normal bladder and sexual function. Results from this ambitious trial are anxiously awaited.¹⁰⁶

Resection of Contiguous Organs and Total Pelvic Exenteration

Although aggressive surgical approaches to rectal cancer have resulted in improvement in locoregional recurrence rates, these rates can still be as high as 33%. Not infrequently, large rectal lesions will invade through the wall of the rectum into contiguous structures such as the bladder, prostate, vagina, and uterus. Carefully selected patients with recurrent or locally advanced rectal cancers may benefit from an aggressive approach such as a total pelvic exenteration. Local recurrences remain localized to the pelvis in a significant number of patients, with autopsy studies demonstrating the incidence of pelvic recurrence to be as high as 50%.¹⁰⁷

Recurrences in the pelvis can result in significant morbidity such as tenesmus, pain, bowel obstruction, and fistula. Although some of these can be ameliorated with radiation, these problems are best managed by preventing their occurrence. Although the impact of total pelvic exenteration on survival has been debated, the potential benefits on controlling locoregional disease and preventing morbidity keeps this technique as one of the tools in the surgeon’s armamentarium when approaching large rectal lesions.

Existing literature on multivisceral resection of both primary and recurrent tumors has been recently evaluated in a systematic review of 22 studies comprising 1,575 patients. The authors reported a 4.2% perioperative mortality rate with morbidity of 42.5%. The overall 5-year survival rate was 50.3% with, not surprisingly, worse outcomes in patients with recurrent compared to primary disease (19.5% versus 52.8%). R0 resection was achieved in 79.5% of cases and, also not surprisingly, was the strongest factor associated with long-term survival.¹⁰⁸ Another review focusing only on locally recurrent tumors, reported R0 resection rates from 30% to 45% and 5-year global survival ranging from 30% to 40%, with authors stressing the importance of careful patient selection.¹⁰⁹ To this end, a panel of 36 colorectal surgeons were recently recruited to establish a scoring system for determining patient suitability for pelvic exenteration. A comprehensive list of clinicopathologic and radiographic criteria were considered and ranked by importance and utility in predicting negative resection margin. The authors hope to apply this quantitatively toward improving outcomes for this highly invasive and morbid intervention.¹¹⁰

For symptomatic tumors that are not resectable, other palliative options to consider include debulking and ablation. Ripley et al.¹¹¹ reported some benefit associated with sequential open radiofrequency ablation and surgical debulking in 16 patients, achieving a median survival of 12 months, with OS 24% at 36 months, and 3 patients remaining with no evidence of disease at 9, 48, and 84 months. There were four cases of significant postoperative morbidity, however, and variable levels of symptom relief.¹¹¹ Pusceddu et al.¹¹² reported far better palliation with CT-guided radiofrequency ablation in 12 patients with painful pelvic recurrence. At the end of follow-up (23±23 months), 92% of patients were symptom free, with a 16% treatment-related morbidity (one rectovesical fistula and one rectal abscess).¹¹² Finally, transrectal high-intensity focused ultrasonography has now been described in the palliative treatment of rectal cancer. As the only completely noninvasive thermal therapy, it can be delivered by either an intracavitary or extracorporeal device, causing focal ablation via coagulative necrosis. In the first case report, it was well-tolerated and led to immediate symptom relief.¹¹³

Combined Modality Therapy (Stage II and III)

The use of adjuvant radiation therapy is based on the substantial incidence of locoregional failure with surgical therapy alone. Older studies demonstrate local failure rates of up to 50% in patients with T3-4 or node-positive disease (Table 60.1).^114–120 The locoregional recurrence rates in these studies are in the range of 25% to 50% for patients with T3-4 and/or node-positive disease and is a dominant pattern of failure, although distant recurrence is also of great importance. Local failure is related not just to the stage of the disease, but also the location of the tumor in the rectum (tumors located low in the rectum have a higher incidence of local failure) and the experience and ability of the surgeon. However, the relevance of these older local recurrence data has been brought into question with the advent of the use of TME, as previously described. It is important to realize that the data on local recurrence after primary surgical resection come from selected series with operations performed by experienced surgeons who have been specially trained in TME and may not be relevant to the operations performed by general surgeons who perform the operation only occasionally and who are not specially trained.

Although initial studies reported locoregional failure rates of <5% after TME without the use of any adjuvant therapy,^{75,77,121–123} there was concern that these excellent results could not be replicated in larger population-based studies. A number of European countries or regions have shown that the overall locoregional recurrence risks could be decreased by limiting the surgeons who were authorized to perform rectal surgery to those who were trained and certified in the procedure, and by having educational sessions for those who were performing the surgery.⁵ This raised the question of what is the true rate of local failure after TME to help define which patients really require adjuvant therapy.

The most important analysis on local recurrence rates with TME are the data from the Dutch TME study in which patients were randomized to receive either TME alone or a short course of preoperative radiation therapy followed by TME.⁷⁷ All patients with rectal cancer were eligible, including those with early-stage disease. Special attempts were made to have good surgical and pathology quality control. The early results (2 years) relating to local tumor recurrence have been reported and are summarized in Table 60.2. The study demonstrates that there are subsets of patients in whom TME alone is likely sufficient for obtaining good pelvic control, including patients with high rectal tumors (some of these may have been sigmoid cancers, rather than rectal) and low-stage tumors (T1-2, N0). On the other hand, low-lying rectal tumors that are moderately advanced (T3-4 and/or node-positive) had a higher incidence of locoregional failure. Local failure after TME alone was 15% in node-positive patients at 2 years, not corrected for site of the primary, and longer-term follow-up will undoubtedly demonstrate higher local failure rates. In addition, as these results were obtained in a controlled setting, one would likely not obtain similarly good results when surgery is done with less careful quality control. There was a consistent decrease in local failure rate by the addition of preoperative radiation therapy, but the absolute magnitude of the effect varied by the tumor characteristics previously discussed. Long-term results from the Dutch TME study have now been published demonstrating a stable, persistent >50% reduction in recurrence risk for the radiotherapy group after a median follow-up of 12 years. For patients with a negative circumferential margin, the benefit was even greater, with the 10-year cumulative incidence of local recurrence 3% after radiotherapy versus 9% after surgery alone (p <0.0001) and the incidence of distant recurrence 19% versus 24% (p = 0.06). In addition, the incidence of cancer-specific death at 10 years was 17% for the irradiated group versus 22% for surgery alone (p = 0.04). OS rates, however, were equivalent.¹²⁴

A trial similar to the Dutch TME study was recently reported. This phase 3 trial randomized 1,350 patients with operable adenocarcinoma of the rectum to short-course preoperative radiotherapy (25 Gy in five fractions; n = 674) or to initial surgery with selective postoperative chemoradiotherapy (CRT; 45 Gy in 25 fractions with concurrent 5-fluorouracil [5-FU]) restricted to patients CRM involvement (n = 676). The primary outcome measure was local recurrence. At the time of analysis, 330 patients had died (157 preoperative radiotherapy group versus 173 selective postoperative CRT), and median follow-up of surviving patients was 4 years. A total of 99 patients developed local recurrence (27 preoperative radiotherapy versus 72 selective postoperative CRT). A reduction was noted of 61% in the relative risk of local recurrence for patients receiving preoperative radiotherapy (hazard ratio [HR] = 0.39; 95% confidence interval [CI] = 0.27 to 0.58; p <0.0001) and an absolute difference at 3 years of 6.2% (4.4% preoperative radiotherapy versus 10.6% selective postoperative CRT; 95% CI = 5.3–7.1). A relative improvement in DFS of 24% for patients receiving preoperative radiotherapy (HR = 0.76; 95% CI = 0.62 to 0.94; p = 0.013), and an absolute difference at 3 years of 6.0% (77.5% versus 71.5%; 95% CI = 5.3 to 6.8) was observed. OS did not differ between the groups (HR = 0.91; 95% CI = 0.73 to 1.13; p = 0.40). These findings provide further evidence that short-course preoperative radiotherapy is an effective treatment for patients with operable rectal cancer.¹²⁵

The data are excellent that radiation therapy, especially when combined with chemotherapy, can decrease the local failure rate. This is shown by a Swedish study of preoperative radiation therapy compared with surgery,¹²⁶ the Dutch TME trial in the preoperative setting,⁷⁶ and by multiple studies in the postoperative setting.^127–132 There are also excellent data to show that locoregional failure is decreased by the use of radiation therapy and is further decreased by the use of concurrent 5-FU–based chemotherapy (Table 60.3).^127,128,133 Most studies have demonstrated that local failure decreases by about 50% with the use of adjuvant radiation therapy, with a greater effect when concurrent 5-FU is used with irradiation. This appears to provide a strong justification for the use of adjuvant radiation therapy. What is less clear is whether trimodality therapy with radiation therapy improves survival, if radiochemotherapy should be given preoperatively or postoperatively, and precisely which patients should be irradiated. To that effect, Schrag et al.¹³⁴ investigated the use of a neoadjuvant chemotherapy utilizing a 5-FU–leucovorin-oxaliplatin (FOLFOX)-based regimen, with selective use of chemoradiation therapy only in those patients who had failed to demonstrate tumor improvement on neoadjuvant chemotherapy. Of the 30 patients treated without radiation therapy in this small pilot trial, none experienced local recurrence with a minimum follow-up of 4 years.¹³⁴ Three patients experienced distant failure, all in the lungs. This interesting pilot trial has led to the current phase 3 cooperative group trial comparing this approach of neoadjuvant chemotherapy plus selective use of radiation versus standard neoadjuvant chemoradiation therapy. Pending any new information from this randomized trial, neoadjuvant chemoradiation therapy remains appropriate standard practice.

DOES ADJUVANT RADIATION THERAPY IMPACT SURVIVAL?

Although there have been multiple randomized trials addressing the use of adjuvant radiation therapy or chemoradiation therapy, and although they consistently show an improvement in local control with adjuvant radiation therapy, the survival outcome data have been mixed. In the past, there have been two meta-analyses performed.^135,136 Table 60.4 shows the results of a meta-analysis by Camma et al.¹³⁶ showing a decreased local recurrence rate, cancer mortality rate, and overall mortality rate with the use of preoperative radiation therapy, although without a decrease in distant metastasis rate. The Colorectal Cancer Collaborative Group study (Table 60.5) demonstrates no improvement in the likelihood of curative surgery with preoperative therapy or of OS with all types of radiation therapy combined.¹³⁵ Preoperative radiation therapy, however, was shown to improve local control, DFS, and OS compared with surgery alone, although deaths within the first year after surgery were higher after radiation therapy. Local recurrence with preoperative radiation therapy was 46% lower than surgery alone, and cancer deaths were decreased from 50% to 45%. Postoperative radiation therapy was shown to improve local control (although less than preoperative therapy), but did not impact long-term survival. Lending substantial strength to the conclusion that there was a true advantage to radiation therapy is the fact that there was a dose response demonstrated for the radiation effect on local control (i.e., better control was obtained with higher radiation dose). This observation strengthens the conclusion, as it demonstrates a direct correlation between the amount of therapy and outcome. The data from this analysis are heavily influenced by the results of a single Swedish study that showed a long-term survival advantage to the use of preoperative radiation therapy compared with surgery alone.¹²⁶ Thus, these data show that improving local control with the use of radiation therapy (and presumably with concurrent chemoradiation therapy) is beneficial, and that trimodality therapy, especially when chemoradiation therapy is used preoperatively, can improve survival.