Chapter 40. Cancer of the Rectum

At the beginning of the 21st century, rectal cancer continues to be a significant medical and social problem. Currently, there are approximately 149,000 cases of colorectal cancer diagnosed in the United States each year. Adenocarcinoma of the rectum accounts for nearly 30% of these cancers. This translates into 41,000 new diagnoses of rectal cancer each year and greater than 10,000 deaths attributable to this disease within the same time period.1

The history of modern rectal cancer resection dates back to 1884, when Czérny described the first abdominoperineal resection (APR). In 1885, Kraske pioneered the transsacral approach of rectal resection and anastomosis. In 1908, Miles improved on the APR by understanding that there was a “zone of upward spread.”2 He emphasized the importance of performing a wide perineal excision. Furthermore, he advocated removal of the rectum with a high ligation of the superior hemorrhoidal artery as well as excision of the abdominal attachments of the rectum and the iliac lymph nodes. Despite the improvements in oncologic resection, operative mortality in Miles' first series exceeded 42%. Over the next 80 years through the late 1980s, mortality and morbidity for rectal cancer surgery improved markedly in pace with improvements in intra-, peri-, and postoperative care. Unfortunately, there were few, if any, advancements in oncologic techniques during this period. Then, in the late 1980s, William Heald described and began popularizing total mesorectal excision (TME) for carcinoma of the rectum.3 In this technique he advocates using sharp dissection to perform the complete excision of the mesorectum and its associated lymphatics along the subtle fascial planes that encompass the rectum. Moreover, Heald described a “zone of downward spread” within the mesorectum that requires complete excision in order to reduce local recurrence. Finally, local excision of small rectal cancers has been used for a 100 years in selected patients. More recently, local excision is being combined with neoadjuvant and adjuvant chemoradiotherapy to maximize local control with a minimally invasive approach.

In Western industrialized nations the average lifetime risk for an individual to develop colorectal cancer is approximately 6%. This risk increases two- to fourfold if the patient has a personal history of a first-degree relative with colorectal cancer. Inflammatory bowel disease (IBD) is another risk factor. In the first 10 years after the initial diagnosis of ulcerative colitis (UC), the incidence of colorectal cancer ranges from 2 to 5%; however, this risk increases 1% for each year of disease thereafter. Pancolitis is associated with both an earlier and an increased risk for colorectal cancer when compared to left-sided colitis alone. For all patients with UC, the cumulative risk for colorectal cancer at 25 years is 25%. Screening the colon yearly starting at 10 years after the diagnosis with colonoscopy and multiple biopsies in four quadrants every 10 cm from the cecum to the distal rectum is used to predict when a patient is at risk for developing colorectal cancer. If high-grade dysplasia is detected in any of the biopsies, the patient needs to have a total proctocolectomy. Some practitioners advocate a surgical resection for low-grade dysplasia as well, whereas some are willing to repeat a colonoscopy with multiple biopsies. If low-grade dysplasia is found on the second short-interval colonoscopy, total proctocolectomy is indicated. Ultimately, the most effective method for preventing colon cancer in patients with UC is to remove the colon once any type of dysplasia has been identified. Crohn's colitis is associated with a similar increased risk for colorectal cancer. This is often not appreciated by clinicians because patients with severe Crohn's colitis often undergo proctocolectomy before their long-term risk becomes an issue.

Genetic risk factors also have been implicated in the development of colorectal cancer. One is familial adenomatous polyposis (FAP), an autosomal dominant syndrome with 100% risk of developing colorectal cancer. The abnormality is caused by a defect in the APC gene located on chromosome 5q21. Patients with FAP develop hundreds or thousands of adenomas by their 20s, and colorectal cancer develops in all patients by age 50 years if untreated. Extraintestinal manifestations of this genetic defect include desmoid tumors, periampullary masses, osteomas, and medulloblastomas. A second genetic abnormality associated with the development of colorectal cancer is related to defects in the mismatch repair genes MSH2 and MLH1. Mismatch repair genes affect the repair of DNA replication errors and spontaneous base repair loss and contribute to hereditary nonpolyposis colorectal cancer (HNPCC) that is also known as Lynch syndrome. Despite the name, these cancers arise from adenomas and may account for 5% of all colorectal malignancies. In this autosomal dominant syndrome, cancers occur more often on the right side of the colon. Despite developing at a younger age, there is a better prognosis with these cancers when compared with age-matched controls with a non-HNPCC colorectal cancer. In theory, a patient with HNPCC living to age 80 years would have an 80% risk for developing colorectal cancer; additionally, there is a risk of endometrial cancer (40%), gastric cancer (20%), biliary tract cancer (18%), urinary tract cancer (10%), and ovarian cancer (10%). Family members should be screened initially at age 20 years with colonoscopy for the presence of polyps or colon cancer. After age 40 years, colonoscopies should be performed yearly. If a polyp or cancer is detected, a total abdominal colectomy with an ileorectal anastomosis is recommended. Urine cytology to rule out dysplastic cells in the genitourinary tract (which is at risk for transitional cell carcinoma) is recommended. Women should get regular transvaginal pelvic ultrasounds and CA-125 levels. Any affected woman who has finished childbearing and requires a colectomy should give strong consideration to a prophylactic total abdominal hysterectomy and bilateral salpingo-oophorectomy.

Dietary fats, especially red-meat fats, have been implicated as a risk factor for colon and rectal cancer.4 People who consume less than 15% of their diet as fat have a lower incidence of colorectal cancer, whereas those who take in 20% of their diet as fat, either as unsaturated animal fat or as highly saturated vegetable oils, have an increased risk of colorectal malignancy.

In the past few decades, several studies have linked alcohol consumption and tobacco use with an increased risk of colorectal neoplasia. Moreover, there appears to be a synergistic effect with an even greater increased risk of adenomatous polyps in people who are both smokers and drinkers.5

The concept that colorectal cancers develop from polyps, or the “adenoma-to-carcinoma sequence,” was first described by Dukes in 1926. The majority of patients with rectal cancer have no inherited component; instead, there is an initiating genetic mutation, such as of an oncogene like ras, that leads to abnormal cell growth. Subsequently, mutations resulting in inactivation of tumor suppressor genes, such as p53, allow for progression to cancer.

The time course for polyp development and transformation to cancer is thought to be 5–10 years. Most adenomas remain benign; however, histologic type, polyp size, and evidence of dysplasia are associated with transformation. Data from the National Polyp Study and St. Mark's Hospital show that approximately 75–85% of adenomas are tubular, 8–15% are tubulovillous, and 5–10% are villous. Tubular adenomas usually form a stalk, whereas villous adenomas have a broad base (Fig. 40-1). Villous histology is associated with an increased risk of cancer development. Only 1% of polyps less than 1 cm in diameter show evidence of malignant transformation, whereas 50% of polyps greater than 2 cm in diameter harbor areas of carcinoma.

Clinically, it is important to diagnose the type, size, and number of polyps to risk-stratify patients for treatment and future surveillance. Endoscopic treatment likely reduces or eliminates the risk of colorectal cancer in patients. Rigid sigmoidoscopy and flexible sigmoidoscopy are all that are necessary to screen the rectum. Sigmoidoscopic screening should be followed by a complete colonoscopy if biopsy of a small rectal or sigmoid polyp shows adenomatous changes. Colonoscopic screening as the first study is indicated in high-risk populations. Autopsy studies have reported that adenomas are present in 20–60% of patients with a colorectal cancer, and synchronous cancers are found in 3–9% of patients. In patients who cannot undergo a preoperative colonoscopy, either a virtual colonoscopy or barium enema should be performed. If both procedures are contraindicated in these patients, colonoscopic evaluation should be performed 3 months after resection.

Treatment of the malignant rectal polyp is becoming more common with the increase in colonoscopic screening and the early diagnosis of small distal rectal cancers. Surgical treatment in part depends on the morphology of the polyp and the histologic evaluation of the resected lesion. Pedunculated malignant polyps are classified by Haggitt according to the depth of invasion of the cancer within the head of the polyp and stalk6 (see Fig. 40-1). Malignant polyps completely resected with greater than 2-mm margins and without stalk invasion are considered adequately treated with colonoscopic removal, provided there are no poor prognostic histologic features; tumors with poor differentiation or lymphatic or venous invasion are associated with an increased incidence of involved lymph nodes.7

Anatomic Landmarks

The type of therapy offered to a patient with rectal cancer depends not only on the stage of the tumor but also on its location within the pelvis and its relation to the anal sphincters. Compared with colon cancer, knowledge and appreciation of anatomic landmarks are critical in determining resectability and sphincter preservation.

The rectum, usually 15–20 cm in length, extends from the rectosigmoid junction, marked by fusion of the taeniae coli into a completely circumferential muscular layer, to the anal canal. The rectum transitions from being intraperitoneal to being completely extraperitoneal 10–12 cm from the anus and the root of the sigmoid mesentery is approximately 19 cm from the anal verge on rigid sigmoidoscopy.8 The rectum is “fixed” posteriorly and laterally by Waldeyer's fascia and the lateral stalks, respectively. In the male patient, the anterior rectum is fixed to Denonvilliers' fascia, a fold of two layers of peritoneum that separates the rectum from the posterior prostate and seminal vesicles. In the female patient, the peritoneal cavity descends to the pouch of Douglas, with its most dependent point being adjacent to the cervix anteriorly and midrectum posteriorly.9 When seen endoscopically, the rectum has three valves of Houston, the middle of which corresponds to the anterior peritoneal reflection (Fig. 40-2A).

While many surgical descriptions for rectal cancer refer to the distance of the lesion from the anal verge or the dentate line, a more accurate description for distal (palpable lesions) is the distance above the anorectal ring as palpated by the examining surgeon. For nonpalpable lesions, we use a rigid sigmoidoscope to localize the lesion and then ascertain the distance from the anal verge to the mass. At the muscular level, the anal canal starts at the top of the “high-pressure zone” that is at the proximal aspect of the anorectal ring, a muscular structure consisting of the internal sphincter, external sphincter, and puborectalis (Figs. 40-2A and 40-2B). The high-pressure zone descends beyond the dentate line to the junction of the anal mucosa and the perianal skin; this junction is often referred to as the anal verge. In order to achieve an adequate distal margin (≥2 cm) with sphincter preservation, the lower border of a tumor must be located high enough above the top of the anorectal ring. If curative resection compromises perfect function of the sphincter apparatus, or if an adequate distal margin cannot be obtained while preserving the anorectal ring, an APR with a permanent colostomy should be constructed. Although a patient may assume that a colostomy indicates a hopelessly incurable cancer, we must emphasize that the colostomy is necessary because of the anatomic location, not necessarily the severity of the rectal cancer.

Vascular Supply

Arteriography demonstrates extensive intramural anastomoses between the superior, middle, and inferior rectal arteries. The superior rectal artery originates from the inferior mesenteric artery and descends in the mesorectum to supply the upper and middle rectum (Fig. 40-3). The inferior rectal arteries, branches of the internal pudendal arteries, enter posterolaterally and provide blood supply to the anal sphincters and epithelium. The middle rectal artery, often depicted in anatomic drawings as a large and significant artery branching off the internal iliac artery on each side, is seldom greater than 1 mm in diameter.10 In one study, the middle rectal artery was observed in only 22% of cadaver specimens.9 When actually present, the middle rectal artery is located near the lateral rectal stalks. These stalks are primarily nerves but have been confused previously with arterial supply.

The superior rectal vein drains the upper and middle thirds of the rectum and empties into the portal system via the inferior mesenteric vein. The middle rectal veins drain the lower rectum and upper anal canal into the internal iliac veins. The inferior rectal veins drain the lower anal canal, emptying into the internal iliac veins via the pudendal veins. Because the venous systems communicate, low rectal cancers may spread via the portal and systemic circulations.

Lymphatic Drainage

Local recurrence after resection is common and can occur with and without distant metastatic disease. Rectal cancer can spread locally via lymphatics that follow cranially along the superior hemorrhoidal vessels. This “zone of upward spread” was described initially by Miles in his landmark paper describing the APR. Heald has described a “zone of downward spread” within the mesorectum3; this zone can encompass as much as 4 cm beyond the distal mucosal edge of the tumor.11,12 Although some surgeons and pathologists describe tumor within this zone of downward spread as tumor implants, others believe that these implants are replaced nodes. Appreciation of the zones of upward and downward spread has influenced the extent of dissection surgeons now perform for curative resection of rectal cancers.

Lymph from the upper and middle rectum drains into the inferior mesenteric nodes (Fig. 40-4). Lymph from the lower rectum may drain into the inferior mesenteric system or into the network along the middle and inferior rectal arteries, posteriorly along the middle sacral artery, and anteriorly through the channels to the retrovesical or rectovaginal septum, to the iliac nodes, and ultimately, to the periaortic nodes. In a Japanese study, the obturator nodes, external to the hypogastric nerve plexus, were found to be involved with cancer in 8% of tumors located in the distal rectum, whereas these nodes were rarely, if ever, involved with proximal tumors.13 Lymphatics from the anal canal above the dentate line usually drain via the superior rectal lymphatics to the inferior mesenteric lymph nodes and laterally to the obturator and internal iliac nodes. Below the dentate line, lymph drains primarily to the inguinal nodes but may empty into the inferior or superior rectal lymph nodes.

Innervation

The pelvic autonomic nerves consist of the paired hypogastric (sympathetic), sacral (parasympathetic), and inferior hypogastric nerves (Fig. 40-5). Sympathetic nerves originate from L1–L3, form the inferior mesenteric plexus, travel through the superior hypogastric plexus, and descend as the hypogastric nerves to the pelvic plexus. The parasympathetic nerves, or nervi erigentes, arise from S2–S4 and join the hypogastric nerves anterior and lateral to the rectum to form the pelvic plexus and ultimately the periprostatic plexus. The inferior hypogastric nerve plexus arises from interlacing sympathetic and parasympathetic nerve fibers and forms a fenestrated rhomboid plate on the lateral pelvic sidewall. Fibers from this plexus innervate the rectum as well as the bladder, ureter, prostate, seminal vesicles, membranous urethra, and corpora cavernosa. Therefore, injury to these autonomic nerves can lead to impotence, bladder dysfunction, and loss of normal defecatory mechanisms.

Fascial Planes

The walls and floor of the pelvis are covered by the endopelvic, or parietal, fascia (Fig. 40-6). The fascia propria, an extension of the endopelvic fascia, encloses the rectum and its mesorectal fat, lymphatics, and vascular supply as a single unit; forms the lateral stalks of the rectum; and connects to the parietal fascia on the pelvic sidewall. The presacral fascia is the parietal fascia that covers the sacrum and coccyx, presacral plexus, pelvic autonomic nerves, and the middle sacral artery. Posteriorly, a thickening of this fascia, called Waldeyer's fascia, is the anteroinferior fascial reflection from the presacral fascia at the level of S4. Anteriorly, Denonvilliers' fascia separates the anterior rectal wall from the prostate and seminal vesicles in the male and is thought to be an entrapped extension of the peritoneum.14

The preoperative evaluation is critically important to treat the cancer optimally and achieve sphincter preservation. With this information, surgeons must individualize the treatment and care of each patient.

History

The patient with rectal cancer usually presents to the surgeon after a definitive endoscopic diagnosis. The patient's initial complaint may include rectal bleeding, a change in bowel habits or stool caliber, rectal pain, a sense of rectal “fullness,” weight loss, nausea, vomiting, fatigue, or anorexia; however, many patients are completely asymptomatic. Specific symptoms may assist the surgeon in deciding on the optimal approach to therapy. Tenesmus, the constant sensation of needing to move one's bowels, usually is indicative of a large and possibly fixed stage II or III cancer. Pain with defecation suggests involvement of the anal sphincters; cancers growing directly into the anal sphincter usually are not amenable to sphincter-sparing procedures. Information pertaining to anal sphincter function is invaluable when one is contemplating a low anastomosis. If patients are incontinent, they are better served with an ostomy. Preoperative sexual function is important to know because one must discuss the risks of the procedure and possible diminution of sexual function postoperatively.

A comprehensive medical history should be aimed at identifying other medical conditions, such as cardiopulmonary, renal, and nutrition, that may require additional evaluation before surgical intervention and allow appropriate risk stratification. For patients with a cardiac history or symptoms, a stress test and cardiology evaluation are indicated.

Family history or factors predisposing the patient to rectal cancer, such as FAP, HNPCC, and IBD, are important to take into account as one plans the operative procedure.

Physical Examination

A careful and accurate digital rectal examination (DRE) is critical in determining the clinical stage and any plans for neoadjuvant therapy. DRE of a palpable lesion allows for the assessment of tumor size, mobility and fixation, anterior or posterior location, relationship to the sphincter mechanism and top of the anorectal ring, and distance from the anal verge.

Rigid proctoscopy is also essential to the evaluation of patients with rectal cancer because it demonstrates the proximal and distal levels of the mass from anal verge; extent of circumferential involvement; orientation within the lumen; and relationship to the vagina, prostate, or peritoneal reflection. All this information aids in determining the feasibility of local excision. Rigid proctoscopy also allows one to obtain an adequate tissue biopsy. Flexible sigmoidoscopy is not used routinely because the flexibility of the instrument can give a false distance between the tumor and the dentate line. Furthermore, a mass will often be described as being a sigmoid or rectosigmoid tumor on flexible colonoscopy, and, when the patient is evaluated in the office with rigid sigmoidoscopy, the lesion is often found to be much lower and in fact is often a true rectal cancer that qualifies for neoadjuvant chemoradiotherapy. Hence, rigid sigmoidoscopy is mandatory for all distal left sided lesions.

A complete colonoscopy to the cecum is essential to rule out synchronous cancers, which occur 2–8% of the time. We prefer colonoscopy over virtual colonoscopy so that we may not only diagnose but also excise any amenable polyps.

Women should undergo a complete pelvic examination in order to determine vaginal invasion or spread to the ovaries. Men should be evaluated for extension into the prostate or bladder.

Preoperative Staging

Following the initial history, DRE, and rigid proctoscopy, additional preoperative staging studies can help to determine the appropriate treatment for each patient, whether radical resection or local excision is warranted, and whether preoperative chemoradiation is recommended. Accurate preoperative staging is gaining increasing importance as combined-modality therapy and sphincter-preserving surgical approaches are considered.

Abdominal and pelvic computed tomographic (CT) scans can demonstrate regional tumor extension, lymphatic and distant metastases, and tumor-related complications such as perforation or fistula formation. Its accuracy in determining the depth of invasion, however, is less than that of endorectal ultrasound (ERUS) or specialized magnetic resonance imaging (MRI). Pelvic CT scan therefore is not recommended as the only modality for evaluation of a patient's primary tumor. For example, the sensitivity of CT scan for detecting distant metastasis is higher (75–87%) than that for detecting perirectal nodal involvement (45%) or the depth of transmural invasion (70%). If a node is seen on CT scan, it should be presumed to be malignant because benign adenopathy is not normally seen around the rectum.

Intravenous contrast material at the time of a CT scan is important to assess the liver for metastatic disease, as well as to evaluate the size and function of the kidneys. Ureteral involvement by the tumor can be assessed and allows for planning of ureteral stent placement preoperatively.

All patients should undergo a chest x-ray or chest CT scan to exclude pulmonary metastases. Although useful information for assessing long-term prognosis, the extra information obtained from a chest CT scan often does not influence the decisions that need to be made to treat the local/regional disease.

Laboratory Studies

Complete blood count and electrolytes often are obtained. Liver enzymes may be normal in the setting of small hepatic metastases and are not a reliable marker for liver involvement.

Guidelines published by the American Society for Clinical Oncology (ASCO) recommend that serum carcinoembryonic antigen (CEA) levels be obtained preoperatively in patients with rectal cancer to aid in staging, surgical treatment planning, and assessment of prognosis. Although neither sensitive nor specific enough to serve as a screening method for the detection of colorectal cancer, preoperative CEA levels greater than 5 ng/mL signify a worse prognosis, stage for stage, than those with lower levels. Additionally, elevated preoperative CEA levels that do not normalize following surgical resection imply the presence of persistent disease and the need for further evaluation. Furthermore, CEA is most helpful in identifying recurrent disease with an overall sensitivity rate of 70–80%.

Endoluminal Ultrasound

Compared with CT scanning, transrectal endoluminal or endoscopic ultrasound (TRUS) permits a more accurate characterization of the primary tumor and the status of the perirectal lymph nodes. Localized cancers involving only the mucosa and submucosa usually can be distinguished from tumors that penetrate the muscularis propria or extend through the rectal wall into the perirectal fat.

ERUS is an office-based procedure that is well tolerated and can be performed by the surgeon for preoperative planning. Figure 40-7 shows the schematic layers seen in TRUS.

T Stage.

Several studies comparing the accuracy of TRUS with CT scan and MRI suggest that TRUS is superior for T staging of rectal cancer. The range of the accuracy of TRUS is 80–95% compared with 65–75% for CT scan, 75–85% for MRI, and 62% for DRE. In one review, the accuracy of TRUS was greatest (95%) in distinguishing whether a tumor was confined to the rectal wall (T1, T2) versus invading into the perirectal fat (T3 or greater) and less able to distinguish accurately T1 from T2 cancers.15Figure 40-8 demonstrates a uT2 lesion. Additionally, in patients who have received prior radiation, the accuracy decreases owing to edema and fibrosis.

Despite these data, there is considerable interobserver variability and a significant learning curve associated with performing TRUS. For these reasons, TRUS understages more frequently than overstages the primary rectal tumor. However, TRUS understages the cancer less often than CT scan (15 vs 39%). A modified TNM (tumor-node-metastasis) classification for rectal cancer has been proposed based on TRUS-derived T stage (Table 40-1).

N Stage.

TRUS is less useful in predicting the status of perirectal lymph nodes. In a number of comparative studies, the accuracy of TRUS (70–75%) was similar to that of CT scan (55–65%) and MRI (60–65%). The accuracy of nodal staging with TRUS requires the nodes to be larger than 5 mm. The contribution of TRUS-guided fine-needle aspiration (FNA) biopsy to N-staging accuracy for rectal cancer is controversial.

Magnetic Resonance Imaging

Endorectal coil magnetic resonance imaging (ecMRI) and surface coil MRI are becoming more useful in the pretreatment evaluation of patients with rectal cancer. MRI offers some advantages compared with TRUS: It permits a larger field of view, it may be less operator- and technique-dependent, and it allows study of stenotic tumors that may not be even amenable to DRE.16Figure 40-9 illustrates a T3 lesion. Like TRUS, ecMRI or phased-array MRI can discriminate small-volume nodal disease and subtle transmural invasion. These specialized MRI techniques can identify involved perirectal nodes on the basis of characteristics other than size, with reported accuracy rates of up to 95%. Another advantage over TRUS is identification of foci not only within the mesorectum but also outside the mesorectal fascia, such as the pelvic sidewall. We currently prefer phased-array MRI for staging of rectal cancers because it provides equal accuracy in staging compared to ecMRI but without the intrarectal coil.

Double-contrast MRI may permit more accurate T staging of rectal cancer by allowing better distinction between normal rectal wall, mucosa, muscularis, and perirectal tissues. In one report, the specificity and sensitivity of ecMRI with combined intravenous and endorectal contrast material to predict infiltration of the anal sphincter were 100 and 90%, respectively. However, N staging was not improved with this approach.

Phased-array surface coil MRI also may be beneficial in predicting the likelihood of a tumor-free resection margin by visualizing tumor involvement of the mesorectal fascia. If confirmed in other series, preoperative MRI may prove useful in selecting patients at high risk of local recurrence for therapy prior to resection.

Positron Emission Tomography

Fluorine-18 fluorodeoxyglucose–positron emission tomography (FDG-PET) is effective in assessing the extent of pathologic response of primary rectal cancer to preoperative chemoradiation and may predict long-term outcome.17 Additionally, it has an accuracy of 87% for detecting recurrence of rectal cancer after surgical resection and full-dose external-beam radiation therapy.18 While PET scans are positive in 90% of primary and recurrent tumors and in distant metastatic disease, they are relatively inaccurate for nodal metastases. Rectal cancer rarely metastasizes to the bones or to the brain, and without symptoms these two areas are not included routinely in surveillance imaging. They will, however, light up on PET scan. Current guidelines recommend that PET scans not be used routinely in the standard workup of a rectal cancer.

The purpose of staging any cancer is to describe the anatomic extent of the lesion. Staging by clinical examination, radiology, and pathology aids in planning treatment, evaluating response to treatment, comparing the results of various treatment regimens, and determining prognosis. Currently, the most widely accepted staging system for rectal cancer in the United States is the TNM classification system.

In 1987, the American Joint Committee on Cancer (AJCC) and the International Union Against Cancer (IUC) introduced the TNM staging system for colorectal cancer; this system was updated in 2010 (Tables 40-2 and 40-3). The TNM staging system is based on depth of tumor invasion as well as presence of lymph node or distant metastases. In stage I disease, the tumor may invade into the muscularis propria. In stage II disease, the tumor invades completely through this layer into the perirectal fat (T3) or adjacent organs (T4). Any lymph node metastasis represents stage III disease, and metastatic spread denotes stage IV disease. Depth of invasion (T stage) of the primary tumor is an important prognostic variable as increasing depth of invasion is correlated with an increasing chance of lymph node metastases. For instance, early-stage cancers extending into the muscularis mucosa (T1) will have a 10–13% incidence of metastasizing to perirectal lymph nodes.19,20 In 805 pathology specimens Sitzler noted that 5.7% of T1 lesions, 19.6% of T2 lesions, 65.7% of T3 lesions, and 78.8% of T4 lesions had lymph node metastases.21

Generally, the biologic behavior of rectal cancer cannot be predicted by its location or size although there is a general consensus among experts that the more distal cancers have a poorer outcome when compared stage for stage with more proximal lesions. Poorly differentiated cancers have a worse long-term prognosis than well- or moderately differentiated tumors. Other factors that portend a poor prognosis include direct tumor extension into adjacent structures (T4 lesions); lymph node metastases; lymphatic, vascular, or perineural invasion; and bowel obstruction.

Surgical resection is the cornerstone of curative therapy. Following a potentially curative resection, the 5-year survival rate varies according to disease extent22,23 (Table 40-4). However, these survival figures may improve with the increased use of adjuvant therapy.

Surgical and oncologic management varies greatly depending on the stage and location of the tumor within the rectum. Superficially invasive, small cancers may be managed effectively with local excision. However, most patients have more deeply invasive tumors that require major surgery, such as low anterior resection (LAR) or abdominoperineal resection (APR). Yet others present with locally advanced tumors adherent to adjoining structures such as the sacrum, pelvic sidewall, vagina, uterus, cervix, prostate, or bladder, requiring an even more extensive operation.

After establishing the diagnosis and completing the staging workup, a decision is made whether to pursue immediate resection or administer preoperative chemoradiotherapy. For patients with stage II and III rectal cancer the authors advocate for combined preoperative chemoradiotherapy. The authors recommend this for all stage II and III patients with tumors located in the distal two-third of the rectum. For patients with rectal cancer in the proximal one-third of the rectum, the authors use preoperative chemoradiotherapy on a case by case basis depending on the size and bulkiness of the tumor as well as the patient's medical and surgical history.

Bowel Preparation

The high bacterial load in the intestinal tract requires preoperative bowel decontamination to reduce the incidence of infectious complications. Prior to the routine use of mechanical bowel preparation and preoperative antibiotics, the reported rate of infection following colorectal surgery was 60%.24 A standard bowel preparation includes a clear-liquid diet 1–3 days prior to surgery, laxatives and/or enemas, and gastrointestinal tract irrigation with a solution of polyethylene glycol electrolyte lavage (GoLYTELY) or saline cathartics. In two separate surveys of North American colorectal surgeons, almost two-thirds preferred the polyethylene glycol electrolyte solutions because of the reliability of the cleansing results.25,26 Certain preparations are contraindicated in patients with certain medical conditions. For example, patients with elevated creatinine or congestive heart failure should avoid the magnesium citrate preparation, whereas patients with gastroparesis should not take GoLYTELY.

Studies have shown that mechanical bowel preparation provides little, if any, additional benefit to reducing the perioperative infection rate. However, we still recommend to our patients that a mechanical bowel preparation be performed in large part because it allows for easier manipulation of the colon and rectum with both open and laparoscopic surgery.27

Oral antibiotics are also used to further decrease the incidence of postoperative infectious complications. Although mechanical cleansing decreases the total volume of stool in the colon, it does not affect the concentration of bacteria per milliliter of effluent. The most commonly used regimen is the Nichols/Condon preparation: neomycin 1 g and erythromycin base 1 g, both non-absorbable antibiotics, by mouth at 1:00 pm, 2:00 pm, and 10:00 pm on the day prior to surgery. Many surgeons substitute metronidazole 500 mg for the erythromycin base because it is active against a greater percentage of gastrointestinal anaerobes.

Instead of an oral antibiotic preparation most surgeons use perioperative systemic antibiotics. A typical choice to cover both aerobic and anaerobic intestinal bacteria is cefazolin and metronidazole administered intravenously just prior to the skin incision. A second dose of cefazolin is administered 4 hours into the case. Postoperative antibiotic prophylaxis usually is continued for 24 hours, although the perioperative dose is more critical. Some surgeons do “double” prophylaxis with oral and systemic antibiotics in all surgeries below the peritoneal reflection. Currently it is our practice to just use intravenous systemic antibiotics.

Perioperative systemic antibiotic coverage is broadened in patients with high-risk cardiac lesions such as prosthetic heart valves, a previous history of endocarditis, or a surgically constructed systemic-pulmonary shunt and with intermediate-risk cardiac lesions such as mitral valve prolapse, valvular heart disease, or idiopathic hypertrophic subaortic stenosis. Intravenous ampicillin 2 g and gentamycin 1.5 mg/kg are administered 30–60 minutes before the procedure, and ampicillin is repeated once 6 hours postoperatively in place of cefazolin; metronidazole is administered as usual. Vancomycin is substituted for ampicillin if the patient is allergic to penicillin or cephalosporin.

Goals of Surgery for Rectal Cancer

The primary goal of surgical treatment for rectal cancer is complete eradication of the primary tumor along with the adjacent mesorectal tissue and the superior hemorrhoidal artery pedicle. Although reestablishment of bowel continuity at the time of surgery has become routine, cancer removal should not be compromised in an attempt to avoid a permanent colostomy.

For tumors located in the extraperitoneal rectum, resection margins are limited by the bony confines of the pelvis and the proximity of the bladder, prostate, and seminal vesicles in men and vagina in women. Although locoregional recurrence may be inevitable, local recurrence, cure, mortality, anastomotic leaks, and colostomy rates after rectal cancer surgery are related to surgical technique as well as to the experience and volume of the individual surgeon and institution.

Resection Margins

Distal Margins

The optimal distal resection margin for surgically treated rectal cancer remains controversial. Although the first line of rectal cancer spread is upward along the lymphatics, tumors below the peritoneal reflection can spread distally via intra- or extramural lymphatic and vascular routes.

The use of APR for low rectal cancers traditionally has been based on the need for a 5-cm distal margin of normal tissue. However, in retrospective studies, margins as short as 1 cm have not been associated with an increased risk of local recurrence.28–30 Distal intramural spread usually is limited to within 2.0 cm of the tumor unless the lesion is poorly differentiated or widely metastatic. Data from a randomized, prospective trial conducted by the National Surgical Adjuvant Breast and Bowel Project demonstrated no significant differences in survival or local recurrence when comparing distal rectal margins of less than 2, 2–2.9, and greater than 3 cm.28 Therefore, a 2-cm distal margin is acceptable for resection of rectal carcinoma, although a 5-cm proximal margin is still recommended.31

Radial Margins

The importance of obtaining an adequate circumferential or radial margin has been appreciated more in the last decade. In fact, the circumferential radial margin (CRM) is more critical than the proximal or distal margin for local control. Tumor involvement of the circumferential margin has been shown to be an independent predictor of both local recurrence and survival. The Norwegian Rectal Cancer group reported on circumferential resection margins with 29-month median follow-up in 686 patients who had curative intent LAR with TME alone (no adjuvant radiotherapy) for rectal adenocarcinoma. The Norwegian group found that the overall local recurrence rate was 7% (22% with positive CRM and 5% with a negative CRM). Moreover, 40% of patients with a positive CRM developed distant metastases whereas only 12% of those with negative CRM developed distant disease.32 In this study a positive CRM clearly affected survival. In another report of 90 patients undergoing resection for rectal cancer, when the radial margins were histologically positive, the hazard ratio (HR) for local recurrence was 12.2, and the HR for death was 3.2 when compared with those with clear circumferential margins. Furthermore, the length of mesorectum beyond the primary tumor that needs to be removed is thought to be between 3 and 5 cm because tumor implants usually are seen no further than 4 cm from the distal edge of the tumor within the mesorectum.6,12 Therefore, in proximal rectal cancers, distal mesorectal excision 5 cm below the lower border of the tumor should be the goal.

Oncologic Results

A number of retrospective studies of local excision since the 1970s have demonstrated a local recurrence rate of 7–33% and survival rates of 57–87%. Many of these reviews are limited, small, single-institution studies, often combining patients with tumors of different depths, including T3 lesions, positive margins, or who underwent different forms of local therapy, such as fulguration and snare cautery. Despite these limitations, many of these studies have demonstrated that local excision for superficial tumors with negative margins may provide similar survival and local control but without the morbidity of the APR. Major risk factors for local recurrence include positive surgical margins, transmural extension, and poorly differentiated histology. These retrospective studies suggest that local excision of selected distal rectal adenocarcinomas may provide adequate oncologic control at considerably less morbidity than APR.

Several prospective studies have been published (Table 40-5). In a study from the M.D. Anderson Cancer Center, 46 patients underwent transanal excision of small distal rectal cancer followed by postoperative radiation treatment.33 Patients with T3 lesions also were given chemotherapy. For patients with negative margins, there was only a 6.5% local recurrence rate (all were T3 tumors) with a 93% overall 3-year survival. Local treatments combined with radiation provided similar oncologic control for T1 or T2 small distal rectal adenocarcinomas as compared with APR.

From the New England Deaconess Hospital in Boston, patients with small distal cancers (<4 cm in diameter and <10 cm from the dentate line) with no evidence of metastatic disease were entered in a prospective study.34 Patients with T1 lesions were observed after local excision. Patients with T2 lesions treated with local excision were given postoperative chemoradiation. Several patients were found to have T3 lesions and all were recommended further radical surgery. Those who refused had adjuvant chemoradiation therapy and were followed. All patients were followed every 3 months for 2 years and then every 6 months for 5 years. The local recurrence rate in this study was 8%, and the cancer-specific mortality rate was 4%. Risk factors associated with recurrence were T3 cancers or lymphatic invasion. Surgery alone was adequate for T1 lesions, and surgery combined with chemoradiation was appropriate for T2 lesions excised with negative margins. Radical resection was and still is appropriate for tumors with positive margins after local excision or for T3 cancers. Patients with lymphovenous invasion deserve further therapy, although that therapy was not defined.

In his initial report Steele and colleagues published the only large multicenter prospective trial of local excision (CALGB 8984 [Cancer and Leukemia Group B]).35 Patients were eligible for the study if their cancer was within 10 cm of the dentate line and was less than 4 cm and involved less than 40% of the luminal circumference. All patients preoperatively were thought to have N0M0 disease, as determined clinically and by CT scan. All study patients had negative margins. T1 lesions had no further treatment, and T2 lesions were treated with chemoradiation. After 6 years of follow-up, the overall survival (OS) and the disease-free survival (DFS) were 85 and 78%, respectively. DFS was 84 and 71% for T1 and T2 lesions, respectively. Seven patients recurred with local disease only and underwent APR with a 70% salvage rate. This approach was no worse than that of radical resection. Longer-term follow-up (median 7.1 years) of CALGB 8984 revealed that 10-year overall survival rates were 84% for T1 lesions and 66% for T2 lesions. DFS was reported at 75% in T1 patients and 64% in those with a T2 lesion. Furthermore, local and distant recurrence rates were 8 and 5% versus 18 and 12% in T1 and T2 lesions, respectively. The longer-term follow-up of CALGB 8984 showed that the rates of local recurrence, OS, and DFS didn't change significantly for T1 lesions, but that there was a significant decrease in OS and DFS in the T2 lesions even though these patients received adjuvant therapy. It is clear that local excision is indicated for appropriately selected patients and that local excision with adjuvant therapy should be used more judiciously especially in medically fit patients.36

Patient Selection and Choice of Operation

Preoperative staging, primarily with ERUS or MRI, is most helpful in identifying appropriate patients for a local excision. Criteria for consideration for local excision are listed in Table 40-6. Patients with T3 or N1 disease are inappropriate for local excision. Given the low probability of microscopic nodal disease in T1 lesions, these patients are the best candidates for local excision. T3 and T4 lesions have a high probability of nodal involvement and therefore should be treated with radical resection. Controversy remains over the best therapy for T2 lesions. Most colorectal surgeons still believe that radical surgery with anLAR or APR remains the standard for T2 lesions. However, local excision combined with postoperative chemoradiation achieves similar rates of survival but not necessarily similar rates of DFS. In patients with a T2 lesion who undergo treatment with local excision and adjuvant chemoradiation, those who have a recurrence ultimately require a salvage APR for cure. The American College of Surgeons Oncology Cooperative group has just finished enrolling patients in a trial examining the benefit of preoperative chemoradiotherapy followed by local excision for appropriately staged T2N0Mx distal rectal adenocarcinoma. Preliminary results have just been reported and there was a 44% pathologic complete response rate, 64% pathologic downstaging rate, 5% had ypT3 tumors and 1–2% positive radial margin rate. These preliminary results show that there is excellent pathologic complete response and downstaging as well as good surgical outcomes with nearly all margins being negative. The ultimate efficacy of this technique will depend on long-term oncologic outcomes which are still pending at the time of this publication.37 If the long-term oncologic outcomes are satisfactory, this approach of accurate staging and then the application of neoadjuvant therapy, followed by a local excision for small T2N0 lesions, may be the new paradigm for this subset of rectal cancer patients.

Tumors less than 3 cm from the dentate line but not invading the sphincters usually can be resected via a transanal procedure. Tumors 5 cm from the dentate line may need a transcoccygeal approach or transanal endoscopic microsurgery (TEM). Tumors 7–10 cm from the dentate line require TEM or should be considered for an LAR. Clearly, tumors tethered to the mesorectum or pelvic floor on physical examination, suggesting transmural involvement, are not amenable to local excision. Patients with such lesions should undergo preoperative radiation followed by a radical resection.

Patients considered medically unfit for a major resection are good candidates for local treatment of most small, mobile tumors, including T2 and T3 lesions, accepting a higher rate of local recurrence. In these circumstances, adjuvant chemoradiation is advocated, and close follow-up is mandatory.

After local excision, if the pathology is unfavorable, the patient should be counseled to have further therapy, including chemoradiation therapy and either an LAR or APR with TME. Local excision in these circumstances can be considered an open biopsy and not the definitive therapy.

Technique

There are four approaches to local excision: transsphincteric, transanal, transcoccygeal, and TEM. The transsphincteric technique, however, leads to significant dysfunction of the anal sphincters with subsequent moderate to severe fecal incontinence. Therefore, the transanal, transcoccygeal, and TEM approaches are the preferred techniques.

Transanal Excision

The majority of small distal rectal cancers can be excised locally via a transanal excision. Tumors amenable to this approach usually range from 6–8 cm above the anal verge which is the same as 3–4 cm above the anorectal ring.

Prior to the procedure, all patients should receive a full mechanical and antibiotic bowel preparation. Pre- and perioperative medications are similar to those administered for radical resection. Most patients are placed in the prone jackknife position, and the buttocks are taped apart. For lesions that are directly posterior, the lithotomy position can be used. The surgeon wears a fiberoptic headlight. A pudendal nerve block using 0.5% Marcaine (bupivacaine) with 1:100,000 units of epinephrine is administered to relax the sphincters and facilitate postoperative pain control. A Lone Star retractor (CooperSurgical, Inc., Trumbull, CT) can be used to expose the dentate line. A Pratt bivalve retractor (Pilling-Weck Instruments, Ft. Washington, PA), a Fansler operating speculum (HaydenMedical, Inc., Santa Clarita, CA) or Parks anal retractor (CS Surgical, Inc., Slidell, LA) is inserted to dilate the anus and expose the lesion. Once the tumor is viewed adequately, traction sutures using 2-0 Vicryl (Ethicon, Somerville, NJ) are placed 2 cm proximal to the tumor. The circumferential dissection line is outlined on the mucosa using the cautery with a pinpoint Bovie tip approximately 1 cm from the border of the tumor; careful attention should be paid to maintaining a wide proximal margin. If an adequate view of the lesion cannot be obtained initially, serial traction sutures starting distally are used to prolapse the lesion into the field. Additional local anesthetic is injected submucosally circumferentially along the Bovie markings to provide hemostasis. Starting proximally and proceeding circumferentially, a full-thickness incision of bowel wall is made down to perirectal fat using the cautery along the previously marked mucosa (Fig. 40-10). Once fat is reached, the dissection is made through the fat to undercut the specimen. Anteriorly in a female patient, one must not injure the posterior wall of the vagina. In a male patient, one must avoid the prostate. Once the specimen is free, carefully maintain and mark the orientation for the pathologist (eg, proximal, anterior, left, right). Irrigate and check for hemostasis. After excision, the defect in the bowel wall is closed transversely with full-thickness bites using interrupted 3-0 Vicryl sutures. One stitch is placed in the center of the incision. One-half is closed, followed by the other. A rigid sigmoidoscope is inserted to visualize the suture line and to ensure patency of the rectal lumen. The patient then is placed supine. A pad is applied to the rectal area and secured with mesh rectal shorts. A pack in the anal canal or rectum is not used. These procedures can be done either as an outpatient or with a short stay. Potential complications include urinary retention, urinary tract infections, fecal impaction, infections in the perirectal and ischiorectal spaces, and delayed hemorrhage. The incidence of these complications is quite low; mortality in most series is zero.

Transcoccygeal Excision

Originally popularized by Kraske, the transcoccygeal excision is used for larger or more proximal lesions within the middle or distal third of the rectum. Bleday et al reported that the average distance of the distal margin of an appropriate tumor that was selected for the posterior or Kraske approach was approximately 4.8 cm from the dentate line.34 This approach is useful for lesions on the posterior wall of the rectum but can be used for anterior lesions.

Patients undergo similar bowel preparation and thrombosis precautions as the transanal excision patients. The patient is placed in the prone jackknife position. The buttocks are taped apart for better exposure, but at closure the tape is released to facilitate approximation of the subcutaneous tissues and skin. After prepping the skin, the rectum is irrigated with a Betadine (povidone/iodine) solution. The incision is made in the intergluteal fold over the sacrum and coccyx down to the upper border of the posterior aspect of the external sphincter. After division of the skin and subcutaneous tissues, one encounters the coccyx and anal coccygeal ligament. In order to obtain optimal exposure, the coccyx is removed by cauterizing its attachments, including the anal coccygeal ligament, from each side and from its lower edge and then proceeding with the dissection on its undersurface. A cutting wire is used to transect the sacral coccygeal joint. With removal of the coccyx, bleeding from an extension of the middle sacral artery is controlled with electrocautery. The levator ani muscles are separated in the midline, exposing a membrane that is just outside the mesorectal fat. Once this membrane is divided, the rectum can be completely mobilized within the intraperitoneal pelvis. For anterior lesions, a posterior proctotomy is made; the anterior rectum is approached under direct vision, with removal of the tumor along with a 1-cm margin (Fig. 40-11). For posterior-based lesions, after complete mobilization of the mesorectum, the distal margin of the tumor can be palpated via a rectal examination; the mesorectum and rectum are transected approximately 1 cm distal to the tumor (Fig. 40-12). The tumor is excised with a 1-cm margin of normal tissue. The advantage of the posterior approach is that the immediate mesorectal tissue adjacent to the tumor is removed along with perirectal nodes. After removal, the specimen is oriented for the pathologist. The incision is closed in a transverse manner using an absorbable suture such as 3-0 Vicryl or 3-0 PDS (Ethicon, Somerville, NJ). After closure of the rectum, an air test is performed by insufflating the rectum with air and filling the operative field with sterile saline. After all air leaks are controlled, the levator ani musculature is reapproximated and the anal coccygeal ligament is reattached to the sacrum, followed by closure of the subcutaneous tissues and skin.

One of the most troubling complications of the transcoccygeal excision is a fecal fistula extending from the rectum to the posterior incision. The incidence of this complication ranges from 5 to 20%.35 These fistulas usually heal after temporary fecal diversion.

Transanal Endoscopic Microsurgery

The transanal endoscopic microsurgery (TEM) technique was first described by Gerhard Buess of Tubingen, Germany in 1980. It is especially useful for small benign and malignant lesions in the mid and proximal rectum that are too high for a traditional transanal excision. This technique is widely used in Europe but over the years has been underused in North America until recently. It is gradually becoming standard practice for early mid to upper rectal lesions. The specialized instrumentation includes a 4-cm Wolf operating proctoscope (Richard Wolf Company, Frankfurt am Main, Germany) in lengths of 12 and 20 cm with a flat or beveled end. The operating proctoscope is equipped with a binocular microscope and videoscope attachment for viewing on a standard laparoscopy tower. A CO2 insufflator and long operating surgical instruments are needed as well. The surgeon must be trained in the technique, which follows the same principles as transanal excision described earlier using the pinpoint tip on the Bovie electrocautery. Preoperative localization in the office with a rigid sigmoidoscope is essential so that the patient can be appropriately positioned. The patient is positioned using a beanbag and fixation to the table with tape, which allows the patient to be rotated laterally during the procedure. For an anterior lesion, the patient is placed in the prone jackknife position. For a posterior lesion, the patient is placed in a modified lithotomy position. For lateral lesions, the patient can be placed on the appropriate side so that the lesion is at the inferior quadrant of the visual field. After the patient is appropriately positioned, the operating proctoscope is fixed to the table with a rigid support arm and a glass faceplate. The faceplate is equipped with two operating ports and a suction port. The rectum is distended with carbon dioxide anywhere from 15- to 26-cm water pressure so that the tumor can be visualized and the resection and closure of the rectum can be completed. After full-thickness excision of the lesion is completed, the defect is endoscopically closed with interrupted 3-0 PDS figure-of-eight sutures. If unable, the defect may be left open as in a standard transanal excision. The one caveat, however, is that extreme care must be taken to identify the peritoneal reflection, especially anteriorly. If dissection carries into the peritoneal cavity, the defect must be closed. If we enter the peritoneal cavity, after we close the defect, our practice is to observe these patients in hospital until they are passing flatus. In selected patients, temporary diversion is needed after entering the peritoneal cavity. Furthermore, patients in whom TEM is contemplated should be made aware that because of technical considerations (proctoscope won't fit or pass and/or poor visualization or entry into the peritoneal cavity), an LAR may need to be performed. This is especially true in patients with a known malignancy.

Unfortunately, the literature describing oncologic outcomes for TEM resection of early-stage rectal adenocarcinoma is mainly single institution, small series, and with short-term follow-up. Most of these studies make a comparison with radical surgery (LAR, APR) but never make a direct comparison with transanal resection. For the most part, the comparison of TEM to traditional transanal excision is made with historical data alone. This is in part because very distal lesions near the sphincter are difficult to excise with the TEM and a traditional transanal excision is easier, whereas the more proximal lesions are not amenable to a traditional approach and a TEM is more likely to succeed in removing these lesions per rectum. Hence, only a small number of tumors thatare above 8 cm and below 10 cm from the anal verge could ever be enrolled in a trial to make a direct comparison. To answer this question, a multicenter randomized trial comparing TEM to traditional transanal excision for early-stage rectal cancer with and without adjuvant radiotherapy needs to be performed. To date this trial has not been done.

TEM resection of low-risk T1 rectal adenocarcinoma results in a 0–11% local recurrence rate whereas local recurrence for T2 lesions without adjuvant therapy is approximately 19–35%. When T2 and T3 lesions are treated with adjuvant or neoadjuvant therapy and TEM resection, the local recurrence rates decrease to 14 and 3%, respectively. One caveat is that all of these studies have short-term follow-up with the longest being 4 years. Indirect comparisons of local excision with TEM for T1 lesions have similar local recurrence rates (7–18% TAE vs 0–11% TEM), and as a result the decision to perform traditional transanal excision versus TEM should depend on the location of the tumor and the individual surgeon's expertise. Local recurrence rates without chemoradiotherapy for either TEM or transanal excision, on the other hand, are not satisfactory. Local recurrence rates for T2 lesions excised by TEM range from 19 to 35% versus 26 to 47% for traditional transanal excision (see Table 12-2) In either case, the results are not adequate, and as a result medically fit patients with T2 lesions should not have either a TEM or a transanal excision without the addition of radiotherapy.

In summary, the results of TEM resection are as good as or better than traditional transanal resection for early rectal cancer. When deciding whether to utilize transanal excision or TEM, the surgeon should remember that TEM offers better visualization, almost complete intact excision, and access to lesions that are higher in the rectum and otherwise would need radical surgery. Cataldo's group from the University of Vermont has shown that TEM resection resulted in intact nonfragmented excision 94% of the time whereas transanal excision only accomplished intact nonfragmented excision 65% of the time (p < .001) and tumor-free margins were 98% with TEM versus 78% with TAE when resecting a rectal cancer (p = .03). Furthermore, they showed a nonstatistically significant trend to lower recurrence rates (22% for TAE and 3% for TEM).38

Oncologic Results

Local failures most often result from inadequate surgical clearance of the radial margin. The concept of total mesorectal excision (TME) proposed by Heald et al has been shown to improve both disease-free and overall survival.3 TME in conjunction with an LAR or APR involves precise dissection and removal of the entire rectal mesentery, including that distal to the tumor, as an intact unit. Unlike conventional blunt dissection, which may leave residual mesorectum in the pelvis, TME involves sharp dissection under direct vision in the avascular, areolar plane between the fascia propria of the rectum, which encompasses the mesorectum, and the parietal fascia overlying the pelvic wall structures. This procedure emphasizes autonomic nerve preservation (ANP) and complete hemostasis and avoids violation of the mesorectal envelope. This results in a characteristic bilobed, smooth, glistening surface of the excised mesorectum.

Because rectal cancer spread appears to be limited to the mesorectal envelope, its total removal should encompass virtually every tumor satellite, thus improving the likelihood of local control. The excellent results with TME may be attributed to improved lateral clearance with removal of potential tumor deposits in the mesentery and decreased risk of tumor spillage from a disrupted mesentery.39 The completeness of the mesorectal excision influences local control, even if the surgical margins are uninvolved. In one report, both local (11.4 vs 5.5%) and distant recurrence rates (19.2 vs 12.2%) were higher in patients with an incomplete, as compared with a complete or nearly complete, mesorectal resection. These favorable results have led some to question the need for routine postoperative radiation in patients undergoing complete resection of rectal cancer with TME. However, the Dutch neoadjuvant trial that randomly assigned 1861 patients with resectable rectal cancer to TME alone or a short course of preoperative radiation (5 Gy daily for 5 days, in the “Swedish style”) followed by TME demonstrated a significantly decreased rate of local recurrence 8.2 versus 2.4% at 2 years.40

Of greatest importance is that improved local control appears to result in better overall survival. In one of the earliest reports, Heald et al noted a local recurrence rate of 3.6% and a survival rate of 86% after 9 years of follow-up.41 In 1994, the Norwegian Rectal Cancer Group was founded to improve the surgical standard by implementing TME on a national level and to evaluate the results; courses were arranged to teach surgeons the technique of TME. Optimized TME reduced the rate of local recurrence (6% TME vs 12% non-TME) and increased overall survival (73% TME vs 60% non-TME) within 2 years.42 This led to a strategic change in both Norway and the United States to initiate quality assessment in the surgical treatment of rectal cancer.

Guillem et al recently demonstrated an improved overall and disease-free survival in patients with T3 or N1 tumors who underwent TME after preoperative combined-modality therapy.43 With a median follow-up of 44 months, the estimated 10-year overall survival was 58% (Fig. 40-13) and 10-year recurrence-free survival was 62% (Fig. 40-14). On multivariate analysis, pathologic response greater than 95%, lack of lymphovascular invasion and/or perineural invasion (PNI), and lack of postoperative positive lymph nodes were significantly associated with improved overall and disease-free survival.

Lateral Nodal Dissection

Despite the advent of TME and the addition of neoadjuvant chemoradiotherapy to the treatment of patients with rectal cancer, there is still a risk of local pelvic recurrence and the appearance of distant metastatic disease. Lateral nodal spread, especially in distal rectal cancers, is one possible culprit for treatment failures in rectal cancer. It is well established that distal rectal adenocarcinomas have a worse prognosis than more proximally based lesions. Most surgeons attribute this to three factors: (1) Distal tumors require a more difficult low dissection in a narrow pelvis; (2) there are probably biologic differences in tumors with the low-lying tumors probably having a poorer biology; and (3) the more distal tumors have a predilection to more complex lymphatic channels and the possibility of lateral spread into the systemic circulation as well as the portal circulation. Takahashi et al performed a retrospective analysis of 764 patients over a 20-year period (1975–1995) who underwent a curative three-space dissection. The three spaces are defined as follows: (1) The inner space is encircled by the visceral pelvic fascia posteriorly and Denonvilliers' fascia anteriorly, and laterally the three spaces unite near the pelvic nerve plexus; (2) the intermediate space is defined by the parietal pelvic fascia posteriorly and the internal iliac arteries and branches laterally and anteriorly; and (3) the outer space is lateral to the internal iliac arteries. Takahashi found that 66 of 764 patients (8.6%) had lateral nodal spread of their rectal cancer. More importantly, 16.4% of the low-lying rectal cancers had their lower margins less than 5 cm above the dentate line. Lateral nodal spread is outside the traditional TME resection plane but can be encompassed by a three-space lateral nodal dissection in appropriate patients. When this was achieved, they had a 42.4% 5-year survival in their subgroup of patients who had lateral spread and a curative three-space dissection.44

A comparative study of Japanese and Dutch patients examined local recurrence in Dutch patients who received TME-alone versus TME-plus preoperative radiotherapy and Japanese patients who were treated with TME-plus lateral nodal dissection (LAR or APR). Most Japanese patients did not receive neoadjuvant therapy. Local recurrence, lateral pelvic recurrence, and presacral recurrence rates were analyzed and are shown in Table 40-7.

In summary, both TME with radiotherapy and lateral nodal dissection without radiotherapy result in excellent local control and have improved local control over TME alone. The conclusion is that the radiotherapy sterilizes the lateral space that has microscopic tumor extension beyond the traditional TME resection plane.45 The major caveat is that patients who have TME alone have much better postoperative sexual and urinary function than those who have TME-plus lateral nodal dissection.46

Quality of Life

Quality of life has improved with TME and ANP. Conventional rectal surgery is associated with a significant incidence of impotence, retrograde ejaculation, and urinary incontinence, presumably owing to damage to the pelvic autonomic parasympathetic and sympathetic nerves by blunt dissection.47 Postoperative impotence, retrograde ejaculation, or both are observed in 25–75% of conventionally treated patients compared with only 10–29% of patients after TME with its careful nerve-sparing dissection.47

Erectile capacity and normal ejaculation may be preserved in most male patients, especially those 60 years of age or younger. In one retrospective study of patients undergoing TME with ANP, 86% of male patients younger than 60 years and 67% of those 60 years or older were able both to engage in postoperative sexual intercourse and to achieve orgasm.47 In female patients, sexual activity was maintained in 86%, sexual arousal with vaginal lubrication in 98%, and the ability to achieve orgasm in 91%. With the advent of pelvic dissections that preserve autonomic nerves, postoperative sexual dysfunction rates have been reduced from greater than 50% to 10–28%.47

Isolated urinary dysfunction is uncommon with preservation of the pelvic autonomic nerves. In a prospective study of rectal cancer patients undergoing TME with ANP, only 2 of 35 had difficulty with bladder emptying and possessed evidence of bladder denervation on postoperative studies.

Some studies, however, have demonstrated impaired quality of life owing to LAR with TME in part because of a temporary loop ileostomy or preoperative radiotherapy. Yet cost-utility analysis estimates that improved survival outweighs impaired quality of life.48

Technique of Total Mesorectal Excision

Prior to the procedure, all patients receive a full mechanical and antibiotic bowel preparation. The patient's abdomen is marked preoperatively by the enterostomal therapy nurse for potential stoma sites. An epidural catheter is placed by the anesthesia team for postoperative pain control. Sequential compression devices are applied to the lower extremities before general anesthesia is induced for deep vein thrombosis (DVT) prophylaxis. One dose of 5000 units of heparin is administered subcutaneously. Cefazolin and metronidazole are infused. After anesthesia is induced, the patient is brought down on the table so that the buttocks are at the edge; a gel pad placed under the buttocks facilitates access to the anus. The patient is placed in a modified lithotomy position using Allen or Yellow Fin stirrups (Fig. 40-15). The hips are minimally flexed and abducted. The feet are positioned flat in the stirrups; an imaginary line is visualized keeping the ankle, knee, and contralateral shoulder in a straight line. Care is paid to having no pressure on the peroneal nerve or bony prominences; a hand should be able to be placed easily between the posterolateral aspect of each lower leg and its respective stirrup. If the patient has had previous pelvic surgery or evidence of hydronephrosis on CT scan, consider bilateral ureteral stent placement. A Foley catheter is placed and is draped over one leg. A nasogastric tube is inserted by the anesthesia team. A DRE is performed. If there is any question regarding the distal or proximal limits of the tumor, rigid proctoscopy may be performed at this time. Preoperatively, the lesion may have been marked by an injection of India Ink. The surgeon should wear a headlight to help with visualization in the lower pelvis. Most surgeons stand on the patient's left, which allows them to operate more efficiently with their right hand in the lower pelvis. A low midline incision is made between the umbilicus and the pubis, keeping in mind potential stoma sites; cephalad extension may be necessary to mobilize the splenic flexure. The abdomen is explored to search for metastatic disease in the liver or peritoneal surfaces. The rectum is palpated to assess the primary mass. The colon is palpated for any synchronous lesions.

The abdominal self-retractor is set up. The patient is placed in slight Trendelenburg's position. The sigmoid is mobilized laterally by scoring the white line of Toldt (Fig. 40-16A). The left ureter is identified by several ways: visualizing it cross over the bifurcation of the common iliac artery, palpating the external iliac artery and pinching the tissue above it, locating it at the level where the sigmoid turns, or incising the peritoneum over the psoas muscle and finding the ureter on the medial aspect of the peritoneum (Fig. 40-16B). If it is clear that much length will be necessary for reconstruction, the splenic flexure is mobilized. Tension on the colon should be gentle but firm; too much traction on the colon or omentum can cause splenic injury. The transverse colon is freed from the omentum by sharp dissection along the avascular plane between the two structures. The bowel is packed into the upper abdomen. The sigmoid is held up in the air at the junction between the descending colon and sigmoid. Both sides of the mesentery are scored from this point down to the sacral promontory. The right ureter is identified. The colon usually is divided at the sigmoid-descending colon junction using a linear stapler (or may be divided between two bowel clamps, which would require a hand-sewn anastomosis). The sigmoidal vessels are isolated and divided using large Kelly clamps, two proximally and one distally. Metzenbaum scissors are used to divide the vessels. The vessels are doubly ligated. The colon is packed cephalad, out of the field. The superior hemorrhoidal artery is then divided at the junction with the left colic artery (Fig. 40-16C). A more proximal ligation of the inferior mesenteric vessel can also be performed if extra length on the colon is needed, but it is not necessary to ligate the IMA flush with the aorta for oncologic reasons. One usually suture-ligates the superior hemorrhoidal vessels so as to ensure hemostasis.

After dividing the superior hemorrhoidal artery, it is important to find the proper plane of dissection at the sacral promontory. One first locates the sympathetic nerves along the pelvic brim. The rectum is retracted anteriorly. Electrocautery with a long Bovie tip is used to develop the loose areolar plane of avascular issue posteriorly (Fig 40-17B). The nerves are visualized and kept posterior to the plane of resection. The presacral fascia is incised down to Waldeyer's fascia, and the dissection is carried inferiorly to the coccyx. The St. Mark's abdominal retractor facilitates the deep pelvic dissection.

The anterior and lateral dissections are then started after the posterior dissection has been partially completed. The peritoneum is incised on each side and then across the anterior midline to meet at the deepest point in the cul-de-sac, the groove between the rectum and the anterior structures (uterus/vagina in women, seminal vesicles in men) (Fig. 40-17A). The mesorectum is separated from the pelvic sidewall using the cautery to divide the thin areolar tissue that is found when one is dissecting in the proper plane. The dissection is carried down anterolaterally to the lateral ligaments or “stalks” (Fig. 40-17C). Only 25% of patients have distinct branches of the middle rectal vessels in these ligaments. They can be divided flush with the pelvic sidewall, but care should be taken to preserve the hypogastric plexus that lies on the pelvic sidewall just lateral to the seminal vesicles in men or just lateral to the cardinal ligaments in women. Preservation of the plexus helps with avoiding postoperative potency or urinary problems, and resection of the plexus is rarely helpful for oncologic reasons. Throughout the lateral dissection, one should be aware of the nerves and vessels along the pelvic sidewall. Too lateral a dissection causes bleeding from the pelvic sidewall.

Anteriorly, the planes are less distinct, and the fat of the mesorectum is thin. The vaginal wall or seminal vesicles are elevated anteriorly using the lipped St. Mark's retractor while the surgeon places posterior traction on the rectum. In the male patient, the dissection is continued through or anterior to Denonvilliers' fascia (Fig. 40-17D). This fascia is often two layers of a thin membrane. When performing a cancer resection, one should take both layers of this membranous fascia off the seminal vesicles and upper prostate if possible.

Point of Transection

For middle to low rectal cancers, TME involves removing the entire mesorectum with its enveloping fascia as an intact unit. For tumors in the upper rectum (>10 cm from the anal verge), TME is extended to 5–6 cm below the level of the tumor, dividing the rectum and mesorectum at the same level. A number of pathologic studies demonstrate that tumor spread within the mesorectum rarely extends beyond 4 cm distal to the caudal edge of the tumor; usually most nodes or mesorectal implants are within 3 cm of the distal edge of the tumor.6,12 However, multiple studies have shown that a 2-cm margin is adequate on the mucosa. Fewer than 2–4% of tumors will have mucosal or submucosal spread beyond 2 cm distally. Rigid sigmoidoscopy may be used to identify the appropriate site for transection if the cancer is not palpable, especially after neoadjuvant therapy.

Once the rectum has been mobilized, a tumor measured at 5 cm by rigid proctoscopy often may be moved to 8 cm from the dentate line, a distance that permits an adequate resection margin and sphincter preservation (Fig. 40-18).

When the distal extent of the tumor and the site of transection have been established, electrocautery is used to dissect the mesorectal fat away from the rectum. Vessels require ligation with 2-0 Vicryl ties. It is important to keep the dissection of the mesorectum perpendicular to the site of transection. “Coning in” as one divides the mesorectum prior to transection should be avoided.

Once the bowel has been cleared of mesorectal fat, a 30-, 45-, or 60-mm TA linear stapler is used to staple the rectum (Fig. 40-19A). This is the first staple line in the “double-stapling technique.” The bowel is clamped just proximal to this point. A no. 10 blade on a long handle is used to transect the bowel. The specimen is handed off the field.

Reconstruction: Double-Stapling Technique

The proximal colon is unpacked, and the length required for a tension-free anastomosis is determined. If more colon is needed, the splenic flexure is mobilized further. This may require an extension of the incision cephalad. Proximal ligation of the inferior mesenteric vein also adds extra mobility (see Fig. 40-16D). The proximal bowel is cleaned by resecting residual fat and small vessels approximately 1 cm proximal to the staple line. The staple line is excised with Bovie electrocautery. Sizers may be inserted to select an end-to-end anastomosis (EEA) stapler diameter (25, 29, or 31 mm). A circular stapler is then chosen. The anvil is placed within the opened bowel. A 3-0 Prolene is used to take full-thickness, 1- to 2-mm bites to fashion a purse-string stitch around the anvil. The purse-string suture is tied gently but firmly around the shaft so that the shaft is completely encircled by bowel (Fig. 40-19C). If there are any gaps, an additional 3-0 Prolene suture can be used to take another full-thickness bite, and this suture may be tied around the shaft as well. The serosa of the bowel is cleaned further of fat and small vessels within 1 cm of the shaft of the anvil to optimize bowel-to-bowel contact when the circular stapler is applied. One can also perform a similar placement of the anvil on the antimesenteric side of the colon for a side-to-end anastomosis. The optimal placement of the anvil in this case is such that only a small blind end of colon remains distal to the anastomosis (1–5 cm).

Attention then is turned to the pelvis, which is irrigated and inspected for hemostasis. This is one's truly last opportunity to inspect this area because one's exposure will be compromised once the anastomosis is completed.

One member of the team then stands between the patient's legs. The circular stapler tip is coated with lubricant on the outside of the stapler; we do not place lubricant on the staples themselves. The tip is retracted fully. A rectal examination is performed, and the anus is dilated gently with two to three fingers in order to accommodate the stapler. The circular stapler is inserted gently following the curve of the rectum—initially straight in and then the stapler is tilted posteriorly. Using close communication with the surgeon overlooking the abdomen, the assistant positions the circular stapler tip so that the trocar will exit either 2–3 mm anterior or posterior (we elect to do this posteriorly in women in order to avoid the vaginal wall) to the staple line (Fig. 40-19B). The trocar then is advanced slowly; the bowel continues to be adjusted as necessary. When the trocar protrudes through the bowel wall, be sure that the trocar is fully advanced so that its bottom is visualized (see Fig. 40-19C). The trocar is removed. Ensuring that the proximal bowel is not twisted and that the remaining bowel, mesentery, and epiploicae are held away, the anvil is brought down gently to the stapler and connected. The colon is inspected again to verify that no adjacent tissue is entrapped. The stapler is closed slowly until both pieces of colon are fully approximated (Fig. 40-19D). The stapler is fired, opened slightly, and gently removed as directed according to the type of stapler. This is the second staple line in the double-stapling technique (Fig. 40-19E). The stapler is opened, and the tissues from the proximal and distal bowel are inspected to make sure that the two rings of tissue, or “donuts,” are intact (Fig. 40-19F). If they are not intact, additional sutures are placed if a visible gap is apparent. All anastomoses are checked for integrity. The surgeon fills the pelvis with saline and clamps the bowel proximal to the anastomosis gently with the hands; the assistant introduces a rigid sigmoidoscope through the rectum and insufflates air. If bubbles cannot be detected, one can be confident that the anastomosis is intact. If bubbles are detected, additional sutures are placed in suspected areas, and a diverting loop ileostomy is constructed. If the anastomosis is disrupted completely, it must be refashioned.

Diverting Loop Ileostomy

A diverting loop ileostomy should be considered in any low anastomoses (<5 cm) from the dentate line, which are associated with anastomotic leak rates of up to 17%. Other risk factors for anastomotic breakdown include a history of radiation, perioperative steroid use, malnutrition, elderly women with a thin rectovaginal septum, or elderly patients undergoing preoperative combined-modality therapy with planned postoperative chemotherapy. Additionally, if there is any question regarding the integrity of the anastomosis, an ileostomy should be created.

Ileostomies can be closed within 8 weeks but often are left in place until the patient completes adjuvant chemotherapy. A Gastrografin (diatrizoate meglumine) enema is used to check the patency and integrity of the anastomosis prior to takedown of the anastomosis.

Drain Placement

Most surgeons continue to advocate routine use of drains after pelvic anastomoses. One prospective, randomized trial of 100 patients to receive either no drains or closed-suction drains demonstrated that the presence or absence of a drain did not influence the rate of morbidity and mortality. Although there is no evidence for the use of drains when an anastomosis has been made outside the pelvis, pelvic drainage may be important after anterior resection. We recommend placing a drain in extremely low resections, especially where the anastomosis was hand sewn or in patients who undergo an APR. For all other resections, placement of a drain may be determined on a case-by-case basis.

Closure

We prefer to close the abdominal fascia with a looped no. 0 PDS suture starting at the cephalad and caudad ends and to run the suture toward the middle. The deep dermal layer is closed with 3-0 Vicryl. The skin is closed with either staples or a 4-0 Vicryl subcuticular suture followed by benzoin and Steri-Strips. A 4 × 8 gauze dressing is applied and covered with Tegaderm (3M, St. Paul, MN).

Postoperative Care

The nasogastric tube is removed at the end of the procedure or on postoperative day 1, and the patient can drink sips of clear liquids. The diet is advanced to low residue after flatus is passed. Cefazolin and metronidazole are continued for 24 hours postoperatively. Heparin is administered subcutaneously at a dose of 5000 units BID or TID depending on the patient's weight. Low-molecular-weight heparin also can be used in appropriate doses. Sequential compression devices are worn by the patient unless the patient is ambulating well. Most patients ambulate on postoperative day 1. The Foley catheter is kept in place for 3–5 days. If an epidural has been used for postoperative pain control, it is usually left in place until the patient is started on oral pain medication when he or she is tolerating clear liquids well.

Coloanal Anastomosis

Anastomoses at or just above the anorectal ring often result in increased frequency of stool, incontinence or soilage, and impaired quality of life owing to an insufficient reservoir. Diet restrictions and time after surgery usually will improve these symptoms, but two alternative techniques of reconstruction address these postoperative problems and often allow for improved function to be attained more quickly.

Colonic Pouch

A 6-cm limb of sigmoid or descending colon is folded, and the apex is brought down to reach the rectal stump without tension. The splenic flexure may require additional mobilization. A colotomy is made at the apex with Bovie electrocautery, and a no. 75 GIA linear cutter is used to staple the pouch on itself to create a common lumen. A second fire of the stapler may be necessary. This pouch now serves as the neorectum. A double-stapled anastomosis as described or a hand-sewn anastomosis then is performed. A diverting loop ileostomy is used routinely for these ultralow anastomoses.

Multiple prospective, randomized studies have demonstrated superior function of a coloanal J-pouch over a straight coloanal anastomosis, especially in the first 6 months after ileostomy takedown.

Transverse Coloplasty

When the pelvis is too narrow for a J-pouch or the length of the pouch is inadequate, a transverse coloplasty may be fashioned. This is performed by placing the anvil of a 29- or 33-mm circular stapler into the cut end of the sigmoid as described under the section Low Anterior Resection. The colon should be mobilized to the level of the middle colic vessels. Beginning 2 cm proximal to the anvil, a 7- to 8-cm longitudinal colotomy is made. This colotomy then is closed transversely. The anastomosis is completed as described under Low Anterior Resection. A diverting loop ileostomy is created.

Traditionally, distal rectal cancers have been treated with an abdominoperineal resection (APR), as first described by Miles, who noted high failure rates after local excision.2 This procedure involves the en bloc resection of the tumor as well as the surrounding lymph nodes and the anal sphincters, resulting in a permanent colostomy.

The APR, although quite successful for early rectal cancers (stage I) in terms of survival, is associated with significant morbidity of 61% and mortality ranging from 0 to 6.3%.49 Urinary complications can be as high as 50% and perineal wound infections 16%. In addition to these perioperative problems, significant long-term morbidity is associated with a permanent colostomy. In a patient survey, 66% of patients complained of significant leaks from their stoma appliance, 67% experienced sexual dysfunction, and only 40% of patients working preoperatively ultimately returned to work.50 There is also a significant change in body image when compared with sphincter-saving procedures. The 5-year survival rates following an APR range from 78 to 100% for stage I, 45 to 73% for stage II, and 22 to 66% for stage III disease.51 Despite radical resection, 20% recur locally. Variations in recurrence rates depend on location of the tumor within the rectum, changes in surgical technique, and the addition of adjuvant therapy.

For patients with cancers that involve the sphincter apparatus or for those who are incontinent of feces, an APR is performed to remove the rectal specimen.

Technique

The patient is marked preoperatively by the enterostomal nurse for a permanent colostomy. Please see the section Low Anterior Resection With Total Mesorectal Excision for details regarding additional preoperative care, positioning, incision, and rectal mobilization. The dissection proceeds down to the striated muscles of the levator ani; one can confirm muscle contraction by using electrocautery. Once this level is reached, the perineal excision field can begin either by the surgeon or by a second team. The two-team approach saves time.

Perineal Dissection

The anus is closed with a no. 0 silk suture in a purse-string fashion (Fig. 40-20B). A marking pen is used to draw an ellipse 2 cm outside the superficial external sphincter and extending from the perineal body anteriorly, coccyx posteriorly, and ischial tuberosities laterally. The incision is made with a no. 10 blade and carried down through the dermis into ischiorectal fat (Fig. 40-20C). Two Gelpi retractors are placed at 45 degrees to the anus in order to facilitate deep dissection. The dissection is carried deep outside the external sphincter toward the tip of the coccyx, keeping in mind the planes of dissection (Figs. 40-20A and 40-20E). The anococcygeal ligament is palpated just anterior to the tip of the coccyx, and the palpating finger meets the fingers of a team member working from the abdominal field (Fig. 40-20D). A pair of large scissors is used to poke through the ligament; the scissors are fully spread and, while wide open, are pulled straight back. Hooking the index and middle fingers under the levator muscles and transecting with electrocautery frees the rectum laterally (Figs. 40-20F and 40-20G).

The anterior surface is dissected last (Figs. 40-20H and 40-20I). The rectum is delivered through the perineal opening. An assistant retracts the skin and subcutaneous tissue anteriorly with an army-navy retractor. Care is taken to keep the posterior wall of the vagina or the prostate anterior to the plane of dissection. The surgeon cups the hand around the rectum with traction posteriorly and inferiorly and uses cautery between the rectum and the anterior structures, often reassessing the plane of transection. Once freed circumferentially, the specimen is passed off the field.

The pelvic floor is irrigated and checked carefully for hemostasis. A tongue of omentum or omental pedicle flap may be used to cover the pelvis to prevent the small bowel from dropping deep into the pelvis if radiation is contemplated. Omentum also helps healing, especially in an irradiated perineum or when patients also have undergone prostate or vaginal resections. The descending colon is further mobilized in order to exit the skin without tension. The colostomy site is prepared similar to the diverting loop ileostomy as described under Low Anterior Resection. The colon is drawn through the colostomy site using a Babcock clamp, but it is not matured until the midline incision is closed.

Two no. 10 Jackson-Pratt drains are placed in the pelvis and are brought anteriorly out through the abdominal wall and secured to the skin using 3-0 nylon suture (Fig. 40-20J). The abdomen is closed as described under Low Anterior Resection. The colostomy is matured using interrupted 3-0 Vicryl suture.

If using a two-team approach, the perineal wound can be closed after the pelvis has been irrigated and hemostasis achieved. Two layers of interrupted 2-0 Vicryl suture are used to approximate the subcutaneous tissue. One layer of 3-0 Vicryl in a vertical mattress fashion is used to approximate the skin. Because this area is often radiated, multiple layers decrease the risk of the wound breakdown extending into the pelvis.

Initially the APR was performed utilizing an anterior approach and then flipping the patient for the perineal component. In this way, APR was done by completing the abdominal mobilization of the rectum to the levators circumferentially and then creating a colostomy and closing the abdomen and flipping the patient to complete the perineal portion of the operation. In most institutions two teams work simultaneously (as described previously) and two instrument tables are used with separate counts and often requiring additional OR support staff to assist. All of this increases utilization in order to save operative time. More recently, attention has reverted to the traditional anterior and posterior approach to the APR. In fact, there are recent data that support a better oncologic outcome using an anterior and posterior approach as described previously. This is known as the cylindrical technique. West et al reported more tissue excised in all pathologic resections and better margins from the muscularis to all resection margins. These results translated into a lower rate of positive circumferential resection margin with APE 14.8 versus 40.6% for traditional APR and a decrease in intraoperative perforations 22.8 to 3.7%, respectively.52 In our experience we have also found better short-term outcomes with lower perineal wound infections and improved perineal healing.

Postoperative Care

Postoperative care is similar to that described under Low Anterior Resection. The patient is not allowed to sit for 5 days but may only recline and ambulate. Thereafter, the patient may sit on a soft pillow; we do not advocate using a “donut” because perineum is not supported. The perineum is cleaned daily with dilute hydrogen peroxide. The Foley catheter remains in place for 3–5 days.

Complications

Perineal wound complications are common following APR and occur in up to 25% of patients. While most of these wound complications are minor, some may require operative débridement. We demonstrated previously that preoperative radiation and primary closure were not associated with an increased incidence of wound complications compared with nonirradiated patients following APR for rectal cancer.53

Stoma complications include ischemia, retraction, hernia, stenosis, and prolapse. The construction of a good colostomy will provide a patient with a superb quality of life after APR. Early education in the immediate postoperative period allows the patient to adjust to life with a stoma. The stoma shrinks to its final size approximately 3 weeks postoperatively when the edema has subsided. An end colostomy may be irrigated to establish regularity of bowel movements and further improve the patient's quality of life. The operative mortality for APRs is less than 2%.

Posterior Vaginectomy

Partial vaginectomy is indicated for locally advanced low rectal cancers involving the vagina. One study demonstrated a 5-year survival of 46% and a median survival of 44 months, with most favorable results from negative surgical margins and node-negative disease.54

If the patient undergoes an APR, the posterior wall of the vagina is removed as the anterior margin of the resection (Fig. 40-21). After completing the posterior and lateral dissections, the rectum is delivered through the perineum. The anterior aspect of the perineal incision includes the posterior introitus and is extended around the posterior third to half of the vagina only to avoid denervation of the urethra. To achieve hemostasis during the procedure, one can place interrupted 2-0 absorbable full-thickness sutures through the vagina from either side, starting at the apex of the incision, and tie the sutures as the specimen is being excised.

If the patient undergoes an LAR with a coloanal anastomosis, the partial vaginectomy may be performed through the abdominal approach. The involved area of the vagina is resected with a 1-cm margin and kept en bloc with the rectum. Subsequent closure of the vagina is completed by initially placing Allis clamps on the vaginal edges and then taking full-thickness bites with 2-0 Vicryl sutures in a figure-of-eight fashion.

Before abdominal closure, we recommend placing an omental flap around the vaginal cuff to prevent breakdown of the vaginal suture line. If a coloanal anastomosis is in place, we would position the omentum between the vaginal and the coloanal suture lines.

Prostatectomy

In locally advanced rectal cancer in which there appears to be possible involvement of the prostate, urethra, bladder, or ureterovesicular junction on CT scan of the abdomen and pelvis, an MRI of the pelvis should be obtained. Urology consult should be made because one must be prepared for radical prostatectomy and/or cystectomy with ileal conduit diversion. A prostatectomy en bloc with rectal resection is an alternative to total pelvic exenteration in patients whose rectal cancer is fixed only to the prostate. The reasons for involving urology are in part due to the vascularity of the prostate. In addition, one should be concerned about constructing any genitourinary anastomosis (eg, between bladder and urethra) in the presence of previous radiation and a rectal anastomosis. Attention should be paid to potential for autonomic nerve deficit if proximity and effacement of the neurovascular bundle are evident on MRI.

Pelvic Exenteration

Total pelvic exenteration is an alternative for patients with locally advanced rectal cancer in which the tumor is contiguous with adjacent organs, such as the prostate or bladder (Fig. 40-22). Long-term survival rates range from 20 to 70% and are improved in younger patients with no lymph node metastases.55 Local recurrence rates range from 3 to 8%. An argument against performing total pelvic exenteration is the considerable morbidity (20–40%) and 0–20% mortality associated with this procedure. The most common complications are infection, small bowel obstruction, and problems with urinary diversion.

Prophylactic Bilateral Oophorectomy

Carcinoma of the rectum metastasizes readily to the ovaries, and prophylactic oophorectomy during rectal resection may diminish the morbidity of carcinoma of the rectum in women. Additionally, the incidence of ovarian cancer in women with a history of colorectal cancer is roughly five times the incidence in women without such a history. Although prophylactic bilateral oophorectomy does not significantly affect survival, the prevention of primary ovarian cancer in postmenopausal women is considered to be the main benefit of this procedure. We discuss this with our postmenopausal patients and offer this to them at the time of operation.

Palliative resection of the primary colorectal tumor in stage IV rectal cancer depends on the degree of symptoms present. In patients who are symptomatic from bleeding, localized perforation and obstruction, there are several management options that can relieve the symptoms of the primary tumor:

1. Permanent diversion followed by chemotherapy (± radiotherapy depending on local symptoms)

2. Palliative resection with a permanent colostomy followed by chemotherapy (radiotherapy is not needed if the primary is successfully resected in the stage IV patient)

3. Palliative resection with restoration of GI continuity followed by chemotherapy (once again, radiotherapy is not needed if the primary is successfully resected in the stage IV patient)

In the symptomatic patient it is our practice to offer upfront surgical resection/diversion and additional therapy using one of the above three options. The surgical procedure depends on the performance status and the intraoperative findings at the time of exploration. In patients with distant spread to solid organs alone, it is our inclination to perform resection of the primary with restoration of GI continuity. In patients with significant peritoneal and pelvic carcinomatosis, we will offer either resection with end colostomy or just a diversion depending on the degree of pelvic peritoneal carcinomatosis. In the patient with a large burden of disease, it is better to just divert the patient and start chemotherapy and the offer palliative radiotherapy if the bulky primary continues to bleed or causes pain from infiltration of the sacral nerve roots.

On the other hand, resection in the asymptomatic patient with incurable stage IV disease is controversial. Resections for rectal cancer often have significant morbidity and measurable mortality but at the same time offer the best palliation of local symptoms. Moreover, in the last decade chemotherapy regimens have significantly extended the life expectancy of patients with colorectal cancer. The advent and widespread use of the FOLFOX (5-fluorouracil [5-FU], leucovorin, and oxaliplatin) and FOLFIRI (folinic acid, fluorouracil, and irinotecan) chemotherapy regimens have extended the median life expectancy of patients with stage IV disease from approximately 8 months to nearly 2 years (see Chap. 36). Many patients live much longer than this. Hence, it is our opinion that upfront palliative resection is warranted in the medically fit patient with a low burden of distant but traditionally incurable disease. In those who have significant distant disease, we prefer systemic therapy with restaging after several courses of chemotherapy.

In the medically fit patient with a good performance status that has substantial improvement in the metastatic burden, we will once again offer palliative resection. The choice of operation (see above choices) depends on the intraoperative findings. This approach has never been studied prospectively and the sparse literature is based solely on retrospective reviews. Despite this, patients who have had asymptomatic primary tumors resected have a substantial survival advantage over those who were never resected. Ruo et al showed that those who were resected versus the nonresected had a prolonged median survival of 16 versus 9 months and a 2-year survival of 25 versus 6% with both measures reaching statistical significance.56

Minimally invasive laparoscopically assisted surgery was first considered in 1990 for patients undergoing colectomy for cancer. The technical feasibility of performing laparoscopic TME was demonstrated in several prospective studies. Preservation of the autonomic nerves is also possible during laparoscopic TME. Early results confirmed complete resection of the mesorectum with intact visceral fascia in all patients.57 Because Nelson and other showed equivalent outcome, quality of life, and survival in laparoscopic surgery for colon cancer, there has been renewed interest in laparoscopic LAR for rectal cancer.58 Currently a randomized prospective trial is ongoing in the United States. The Conventional versus Laparoscopic-Assisted Surgery In Colorectal Cancer (CLASSIC) trial group from the United Kingdom published the results of their randomized comparison of laparoscopic versus open LAR for rectal cancer in 794 patients (526 lap vs 268 open) with a median follow-up of 36.8 months and equivalent numbers or patients receiving chemoradiotherapy. The overall survival (open 66.7 vs lap 74.6%) and disease-free survival (open 70.4 vs lap 70.9%) were similar in both groups and were without statistical significance. The overall local recurrence rates were 7% in open resection and 7.8% in the laparoscopic group. Even though the laparoscopic group had a higher positive circumferential resection margin (CRM), there was no difference in local recurrence at 3-year follow-up. This increased positive CRM is certainly concerning with only 3-year follow-up and longer-term follow-up is certainly needed to ensure that this is not significant.59,60 In addition to increased positive CRM, the above CLASSIC trial reported overall worse sexual function in men (not women) undergoing laparoscopic rectal cancer resection.61 Both a positive CRM and worse sexual function are major potential complications, and, as a result, laparoscopic surgery for rectal cancer should still be limited to ongoing trials (especially in male patients) until long-term data show at least equivalent recurrence and complication rates. This is especially true because the same CLASSIC trial measured quality-of-life outcome for these patients and there was no difference between laparoscopic resection and open surgery.

Besides surgical resection for rectal cancer, there are other options for patients who may not be candidates for surgery owing to their comorbidities, extent of disease, or preference. Endocavitary radiation may be delivered at doses of 50 cGy for palliation and for curative intent. Performed as an outpatient and well tolerated by patients, endocavitary radiation is delivered with sedation and local perineal block.

Electrocoagulation may be used via a transanal approach after administering general anesthesia and placing the patient in the lithotomy position. The rectal lesion and a 1-cm margin are fulgurated. Recurrence rates approach 50–80%; therefore patients may require repeat treatments.

Cryotherapy, another alternative modality, results in a large amount of foul-smelling discharge. Photodynamic therapy has limited availability. Laser vaporization using neodymium:yttrium-aluminum-garnet laser provides palliation but is associated with a 14% recurrence rate and is costly.

Complications of surgical management of rectal cancer may include those common to any major intra-abdominal operation, such as infection, bleeding, wound problems, deep venous thrombosis/pulmonary embolism, myocardial infarction, pneumonia, and renal failure. There are, however, several complications that are related to rectal cancer. There is a 50% incidence of impotence in men following resection for rectal cancer. Therefore, it is critical to discuss this situation with the patient before the resection and to record the preoperative status of his sexual function. If a man is impotent after surgery, it is advisable to wait 1 year before undergoing implantation of a penile prosthetic device. This delay is recommended not only to ensure that the malignancy has been cleared but also to allow the patient sufficient time to overcome psychological impediments such as a change in body image from pelvic surgery or from a colostomy. Women may also suffer from impaired sexual function, especially if the vagina is distorted during the rectal resection.

A possible permanent colostomy is often not preferred by patients. Its placement, however, must be explained in a way that the patient understands that he or she may be left with this if reconstruction is not technically possible.

Anastomotic leak, which occurs in up to 20% of patients, can be avoided by constructing the anastomosis with well-vascularized tissue without tension. Interestingly, young, muscular men have a higher incidence of anastomotic leaks, which may result from the technical challenge of operating in a narrow pelvis or from strong sphincters that may stress the anastomosis. The latter may be addressed by dilating the anal sphincter in the operating room at the end of the procedure. Anastomotic leaks usually present between 4 and 7 days postoperatively. Symptoms may include fever, tachycardia, arrhythmias, tachypnea, enterocutaneous fistula, or diffuse peritonitis. When a leak is suspected, the patient should be made NPO and blood should be sent for a complete blood count, electrolytes, and type and cross-match. An upright chest x-ray will diagnose pneumoperitoneum. Abdominal series may demonstrate extraluminal air. CT scan of the abdomen and pelvis with water-soluble contrast material may demonstrate abscess formation, extraluminal air, and the actual leak. Barium should be avoided because leakage of barium creates a destructive peritonitis. A leak may be managed with intravenous antibiotics and bowel rest in a patient without peritonitis. An abscess may be drained percutaneously. An enterocutaneous fistula may be treated with total parenteral nutrition and local wound care. If a large leak is demonstrated or the patient experiences peritonitis, exploratory laparotomy with diverting ileostomy or colostomy should be performed. The anastomosis is rarely taken down and should not be reconstructed in the presence of sepsis.

Massive venous bleeding from the presacral space may result intraoperatively from lateral dissection onto the pelvic sidewall or sacrum. Ligation of the iliac vessels is discouraged and may be hazardous. If massive bleeding is encountered, a surgical metal “tack” may be driven into the sacrum to compress the venous space. Additionally, the pelvis may be packed for 24–48 hours, at which time the patient is returned to the operating room for pack removal and closure.

Urinary dysfunction may occur after rectal resection. Many men have coexisting prostatic hypertrophy. Because low rectal dissection approaches the membranous urethra, Foley catheters usually are kept in place for 5 days. Patients may be discharged with indwelling catheters, especially if they have undergone partial prostatectomies or seminal vesiculectomies. Women may experience urinary incontinence if the anterior aspect of the vagina, which contains the neurologic control of the urethra, is transected.

Obstructing Cancer of the Rectum

For obstructing cancers of the rectum, a loop ileostomy, performed as an open or a laparoscopic procedure, is constructed for diversion. Usually, the tumor is staged as a T3 or N1 lesion; the patient is treated with neoadjuvant chemoradiation and considered for subsequent surgical resection.

Metastatic Rectal Cancer

The management of hepatic and pulmonary metastases is not described in this chapter (see Chap. 36). Nonetheless, if a patient presents with incurable metastatic disease and life expectancy is greater than 6 months, it is reasonable to consider a palliative rectal resection. If the rectal lesion is staged as T3 or N1, we recommend neoadjuvant chemoradiation because this addresses both the primary lesion and the metastasis and may provide some palliation of obstruction, bleeding, or pain. Other options include rectal stents or laser destruction of the tumor to maintain an adequate lumen. It is important to understand the patient, his or her desires, and general state of health when recommending treatment at this stage of cancer.

Recurrent Rectal Cancer

Local recurrence of rectal adenocarcinoma is seen in up to 30% of patients. Although recurrence may be seen at the distal margin of the anastomosis, most develop from residual cancer on the pelvic wall. The time course for recurrences to present through the anastomosis is approximately 18 months. By their nature, these tumors are fixed to the pelvic wall and surrounding viscera. They cause significant symptoms, such as intractable pelvic pain, bleeding, cramping or constipation, urinary tract dysfunction, and chronic pelvic sepsis.

When patients present with these symptoms or with a rising CEA level, a workup including CT scan of the abdomen and pelvis, ERUS, MRI of the pelvis, and PET scan may be helpful. A careful pelvic examination is mandatory. A biopsy, either via sigmoidoscopy or CT-guided, should be used to confirm the diagnosis pathologically. If external radiation has not been used before, it should be considered. The surgeon should review the imaging studies and determine which organs are involved, such as the vagina, uterus, prostate, bladder, sacrum, and small intestine, which will require en bloc resection. Urology consult should be obtained if there is any question of prostate or bladder involvement; ureteral stents should be placed preoperatively. Removal of the rectum and urinary bladder with surrounding lymphatic tissue results in a permanent colostomy and ileal conduit.

Intraoperative Radiation Therapy

Intraoperative radiation therapy (IORT) may be considered in patients with pelvic sidewall recurrence. This is performed in an operating room–radiation therapy suite. Resection with negative microscopic margins and absence of vascular invasion independently predicts improved local control and survival after resection and IORT.62 The major morbidities of IORT include peripheral neuropathy and ureteral stenosis.

Palliation

These tumors are difficult to palliate, let alone cure. Surgical resection combined with aggressive multimodality therapy is advocated to avert the morbidity of pelvic disease and to prolong survival in a subset of patients, with survival rates up to 30%.63 Most patients, however, will not be offered curative surgery on the basis of comorbidities, poor performance status, distant metastases, or locally unresectable disease on preoperative imaging. These patients may be offered palliative intervention. Miner and colleagues demonstrated that in patients who underwent surgery with palliative intent, improvement was noted in 40% with bleeding, 70% with obstruction, and 20% with pain.64 When considering the effective use of surgery for these patients, decision making is complex because one must balance palliation of symptoms, comorbidities, and patient desires and goals. Seeking the input of a multidisciplinary treatment group, including medical oncologists and radiation oncologists, is invaluable.

Patients with rectal cancer who undergo surgery with intention to cure and without evidence of gross disease postoperatively may still develop local recurrence or distant metastases. Up to 10% of patients who undergo TME with tumor-free radial and distal margins may develop local failure. The goal of adjuvant therapy is to eliminate the micrometastatic disease present at the time of surgery.

Adjuvant Chemoradiation

In 1990, the National Institutes of Health consensus statement concluded that “combined postoperative chemotherapy and radiotherapy improves local control and survival in stages II and III patients and is recommended.” Most of the information regarding chemotherapy for colorectal cancer comes from trials of colon cancer rather than for rectal cancer. The NSABP C-04 (National Surgical Adjuvant Breast and Bowel Project C-04) trial studied stages II and III colon cancer patients and demonstrated that 5-fluorouracil (5-FU) and leucovorin treatment had a significantly better 5-year survival rate (74 vs 69%) compared with 5-FU and levamisole.65

Several trials have suggested a benefit for adjuvant chemoradiation for rectal cancer in patients with resected stage II or III cancers. The GITSG (Gastrointestinal Tumor Study Group) trial demonstrated that combined chemoradiation resulted in an improvement in overall survival as well as a decrease in local recurrence.66 The NCCTG (North Central Cancer Treatment Group) trial demonstrated that the addition of chemotherapy to radiation reduced both local recurrence (13 vs 25%) and distant metastases (28 vs 46%) and improved survival.67

Radiation therapy used alone as adjuvant therapy may improve local recurrence and survival rates. A theoretical reason to use postoperative radiation therapy is that more appropriate patient selection can be achieved because pathologic staging is performed prior to radiation. Disadvantages include radiating the neorectum and small bowel and a lower tendency of patients to complete their radiation. While none of the trials in the 1980s and 1990s demonstrated increased survival, one study did show a decrease in local recurrence.

Neoadjuvant Chemoradiation

There are a number of potential advantages for using neoadjuvant chemoradiation. They include the ability to deliver higher doses of chemotherapy with radiation. Another advantage is not only to downstage the tumor, which has been noted in 60–80% of patients, but also to achieve a pathologic complete response, which occurs in 15–30% of patients. The ability to “shrink” the tumor facilitates surgical resection, thereby allowing one to achieve negative margins and perform a sphincter-preserving operation in patients who otherwise would require an APR. Additional advantages include radiating tissues with a greater oxygen supply, not radiating the anastomosis, and decreased likelihood of developing radiation enteritis because small bowel is less likely to enter the pelvis. Finally, patients are more likely to complete the course of radiation therapy because it precedes their surgical resection.

The Dutch Colorectal Cancer Group demonstrated a significantly decreased rate of local recurrence at 2 years in patients who received preoperative radiotherapy (20 Gy over 5 days) followed by TME compared with TME alone (2.4 vs 8%).39 The Swedish trial was the first and only study to demonstrate a survival benefit (58%) for stage III rectal cancer patients receiving preoperative radiation (short course of 5 Gy over 5 days) followed by surgery compared with patients who underwent surgery alone (48%).68 The Swedish trial also demonstrated a decreased rate of local recurrence in the radiation-treated group (11%) compared with 27% in the surgery-alone group. Furthermore, a meta-analysis concluded that preoperative radiation therapy plus surgery compared with surgery alone significantly reduced the 5-year overall mortality rate, cancer-related mortality rate, and local recurrence rate.69

In the German Rectal Cancer Trial published in the New England Journal of Medicine, Sauer et al randomly assigned patients with clinical stage II or III rectal cancer to preoperative (421 patients) or postoperative (402 patients) chemoradiotherapy based on a concurrent long course of radiotherapy (5040 cGy delivered in fractions of 180 cGy per day, 5 days per week) and 5-FU (120-hour continuous intravenous infusion during the first and fifth weeks).70 Six weeks later, TME was performed, followed by four cycles of 5-FU 1 month postoperatively. Despite the preponderance of distal tumors in the preoperative chemoradiation group, there was no difference in overall survival or disease-free survival at 4 years. On the other hand, patients receiving preoperative chemoradiotherapy had a 6% local recurrence rate as compared to a 13% local recurrence rate in those receiving postoperative chemoradiotherapy. Moreover, the group treated with preoperative chemoradiotherapy had a higher incidence of sphincter preservation and lower treatment-related toxicities (27 vs 40%). Differences in local recurrence, sphincter preservation, and treatment toxicities were all statistically significant. A Polish rectal cancer trial from 2004 compared preoperative short-course radiotherapy (5 days of 5 Gy) versus conventional radiotherapy (28 fractions of 1.8 Gy for a total of 50.4 Gy) to ascertain whether there was a difference in sphincter preservation. The surgical resection was based on the tumor status at the time of surgery not before the radiotherapy. This allowed for the surgical decision to be made after tumor shrinkage for patients who received the longer course of radiotherapy. Between 1999 and 2002 the study enrolled 316 patients. Tumor shrinkage was on average 1.9 cm greater in the long-course group and this was statistically significant. However, sphincter preservation in the short-course group was 61% and in the long-course group 58%. In other words, whether the patient received short- or long-course radiotherapy, it did not impact sphincter preservation.71 There was also no difference in survival, local control or late complications. Furthermore, this Polish trial reported no differences in anorectal or sexual function between the short- or long-course radiotherapy.72 A French group in conjunction with the EORTC (European Organisation for Research and Treatment of Cancer) group studied “the addition of chemotherapy to preoperative radiotherapy and the use of postoperative chemotherapy in the treatment of rectal cancer.” Patients with clinical stage T3 or T4 rectal adenocarcinoma were randomized to four groups: preoperative radiotherapy, preoperative chemoradiotherapy, preoperative radiotherapy with postoperative chemotherapy, and preoperative chemoradiotherapy with postoperative chemotherapy. The primary end point of this study, which enrolled 1011 patients, was overall survival. The main secondary end point was local recurrence. The results showed that there was no difference in overall survival between the groups that received chemotherapy preoperatively or those that received it postoperatively. There was, however, a difference in local recurrence. In the patients who received pre-op, post-op, or pre-op and post-op chemotherapy, the local recurrence rates were 8.7, 9.6, and 7.6%, respectively whereas the radiotherapy-alone group had a local recurrence rate of 17.1%. This was statistically significant and it suggests that there is a benefit to local control by adding preoperative chemotherapy to the regimen. It is not clear whether the addition of postoperative chemotherapy to a patient who has already received preoperative chemotherapy with the radiation treatment has any survival benefit.73

Our current practice is to recommend preoperative staging with ERUS or MRI to all patients with rectal adenocarcinoma and then to offer chemoradiation to medically fit patients with curative intent who have T3-T4 or N-positive rectal carcinoma. Some patients with bulky T2 lesions near the sphincters should also be considered for neoadjuvant chemoradiotherapy in order to improve sphincter preservation (Table 40-8). Neoadjuvant therapy then is followed by TME with APR or TME with an end-to-side or colonic J-pouch reconstruction. Postoperatively, patients who have had involved lymph nodes either by preoperative staging or on the final pathology report are encouraged to have additional postoperative chemotherapy. Postoperative chemotherapy in node-negative patients or patients who have had a complete response is determined on a case by case basis.

After curative resection, long-term follow-up includes routine screening for rectal recurrence and metachronous colorectal neoplasms. Between 60 and 84% of recurrences are seen in the first 24 months and 90% within 48 months. Median time to recurrence is 11–22 months. Local recurrence rates range between 4 and 50%. Survival rates vary according to stage (see Table 40-4). Median survival after recurrences are detected is 40 months.

Patients are seen postoperatively at 2 weeks and then every 3 months for 2 years. At each visit, the patient undergoes DRE and sigmoidoscopy, and a CEA level is obtained. As per the National Comprehensive Cancer Network (NCCN) guidelines we recommend at 1 year postresection, a colonoscopy and abdominopelvic CT. Either a chest CT or chest x-ray is also performed. A CT scan is performed annually until 3 years postoperatively. Colonoscopy frequency is determined by the findings at 1 year. If there are no polyps and no recurrence, the follow-up interval can be lengthened. After 2 years, patients continue to be followed every 6 months with CEA levels and physical examinations until 5 years after the surgery. At 5 years, if the patient has had no recurrence, he or she may be followed yearly with clinic visits and may undergo colonoscopy every 3 to 5 years. Of course, closer observation is indicated for patients at high risk for subsequent cancer formation, such as patients with IBD, polyposis syndromes, or a strong family history of colorectal cancer.