Figure 68-11. Anatomy of the rectum. A: Coronal pelvis. B: Sagittal pelvis.

The blood supply of the rectum comes primarily from the superior rectal artery, which is the terminal

branch of the inferior mesenteric artery after it gives off the left colic artery. The superior rectal vein

has a parallel course to its homonymous artery, and joins the left colic vein to form the inferior

mesenteric vein draining into the splenic vein. The lower portion of the rectum and the anal canal also

receive blood supply from the internal iliac vessels through the middle rectal artery, an inconsistent

branch of the inferior vesical artery; the inferior rectal artery is a branch of the pudendal artery. The

middle and inferior rectal vessels anastomose with the upper rectal vessels, supplying enough blood to

the entire rectum. As in other locations, the middle and inferior rectal veins follow the course of the

homonymous arteries, draining into systemic circulation through the internal iliac veins.

The anatomy of the autonomic pelvic nerve system is very important when operating in the rectum,

because of its proximity to the plane of dissection during different parts of the operation. Damage to

these nerves can result in urinary and/or sexual dysfunction. The hypogastric plexus, located in front of

the aorta, contains predominantly preganglionic sympathetic fibers originating from the lumbar

sympathetic trunk. The fibers of the hypogastric plexus converge at the level of the aortic bifurcation

into well-defined hypogastric nerves, which course laterally and anteriorly over the internal iliac vessels

toward the lateral pelvic sidewall. There they join the splanchnic pelvic nerves, containing primarily

postganglionic parasympathetic fibers from the anterior rami of S2, S3, and S4, to form the pelvic

plexus. Branches of the pelvic plexus provide innervation to the distal ureter, the vas deferens, the

seminal vesicles, urinary bladder, and prostate. Branches of the pelvic plexus also provide innervation to

the distal rectum, passing through the lateral rectal ligaments, or lateral stalks. Finally, distal to the

lateral rectal ligaments, the distal pelvic plexus forms the urogenital neurovascular bundles that pass

close to the posterolateral aspect of the seminal vesicles or the vagina, extending toward the apex of the

prostate and the neck of the bladder.

Clinical Staging of Rectal Cancer

The preoperative evaluation of the rectal cancer patient follows the same principles as that of the colon

cancer patient, but the wider spectrum of therapeutic options for rectal cancer patients requires accurate

1788

information about the location and stage of the tumor. The clinical stage is important in making

treatment decisions, such as the intent of the treatment (palliative or curative), the need for

neoadjuvant therapy, or even the extent of the surgery (LE vs. TME).

Figure 68-12. Endorectal ultrasound. Role in rectal cancer staging.

Patients with rectal cancer should undergo a complete colonoscopy because synchronous polyps are

present in up to 30% of patients, and synchronous cancers in up to 5.3% of patients. They also require a

CT scan of the chest, abdomen, and pelvis to exclude distant metastasis, which is present in 20% of

patients at the time of diagnosis.

The preoperative locoregional staging of rectal cancer follows the clinical TNM system, based on

depth of tumor penetration in the rectal wall and the presence of regional lymph nodes. However, the

preoperative assessment of rectal cancer goes beyond determination of clinical tumor stage; it includes

the distance of tumor from the anal verge, its relationship to the sphincter complex and the levator

muscles, the proximity of tumor to the MRF, and the presence of extramural venous invasion (EMVI).

This information is essential in planning treatment and surgery. Gross morphologic features such as size,

morphology, and ulceration have been associated with prognosis. However, when stratified by stage,

these characteristics have not appeared to contribute independently to oncologic outcomes, and

therefore they are not incorporated into treatment decisions.

3 DRE and proctoscopy provide the surgeon with the first direct impression of the tumor. Tumor

mobility on DRE provides a gross estimation of the tumor depth of invasion. There is a four-category

clinical classification of rectal tumors, based on mobility on DRE, which supposedly corresponds to the

four T categories of the TNM system. However, the correlation between the clinical and pathologic T

classifications is not very accurate, and varies significantly with the experience of the examiner. In

general, DRE is able to distinguish between mobile tumors, most likely limited to the bowel wall, and

tethered or fixed tumors, likely penetrating beyond the bowel wall. However, DRE fails to identify

more than 50% of pathologically proven involved nodes.157 Finally, only very distal tumors are within

reach of the examining finger. Therefore, pretreatment clinical staging requires imaging studies;

namely, ERUS, multidetector CT (MDCT) scan, and MRI.

Ultrasound with an endorectal rotating probe is the imaging modality that best depicts the different

layers of the bowel wall, and it is most useful for staging early rectal cancer (Fig. 68-12).158 But ERUS

has a relatively short focal range, and cannot depict important anatomical structures such as the MRF.

MDCT scans provide accurate images of the rectum, adjacent pelvic structures, and even the MRF, but

have lower tissue resolution compared to MRI. A CT scan of the chest, abdomen, and pelvis should be

performed in every rectal cancer patient to exclude distant metastases. MRI with a surface phased-array

coil has become the preferred imaging modality for locoregional staging of rectal cancer. Using various

sequences and multiple planes, MRI provides high-tissue resolution and excellent anatomical depiction

of the rectum, the mesorectum, the MRF, the levator muscles, and other pelvic structures, relative to

1789

tumor location (Fig. 68-13).159 The addition of new contrast agents and advanced functional sequences

such as diffusion-weighted imaging (DWI) and dynamic contrast enhancement (DCE) permit the

quantification of tumor biologic processes such as microcirculation, vascular permeability, and tissue

cellularity. While still experimental, these images are potentially useful for early assessment of rectal

cancer response to neoadjuvant therapy.160,161

Comparative studies suggest that ERUS is at least as accurate as MRI and more accurate than CT scan

in assessing the depth of tumor penetration in the bowel wall. The accurate detection of involved

mesorectal lymph nodes remains a challenge for all three techniques, however. Two meta-analyses of a

number of case series assessing the accuracy of these three techniques reported a wide range of

sensitivities and specificities for each, reflecting not only differences in technology, but also wide

variation in the criteria used to define malignant lymph nodes.162,163 However, the older studies

included in these analyses were conducted before MRI techniques for rectal cancer staging were fully

developed. At this time, MRI with a rectal cancer protocol provides the most useful information in a

majority of rectal cancer patients, and has great utility for surgical planning, while ERUS is most useful

for staging early-stage tumors.

Figure 68-13. Magnetic resonance imaging. Role in rectal cancer staging.

Tumor distance from the MRF – the CRM when performing TME surgery – has prognostic

implications for local recurrence and patient survival, and has become one of the most important

parameters in the preoperative evaluation of rectal cancer patients.164 MRI is the most accurate imaging

modality in determining the distance of the tumor to the MRF, and predicting CRM involvement. The

MERCURY trial, a prospective observational study assessing the accuracy of MRI in predicting a

curative resection in rectal cancer, reported 92% specificity in predicting a negative CRM.165 Other

important tumor features accurately assessed by MRI, and associated with patient outcomes, are

extramural spread, EMVI, involvement of the peritoneal reflection, and distance of tumor from the

levator muscle and sphincter complex.166,167 Using information on these parameters for patients

registered to the MERCURY trial, clinicians are able to stratify rectal cancer patients with good

prognosis (e.g., clear CRM, no evidence of EMVI, T2 or T3 <5 mm and not involving the

intersphincteric plane) with a 3% LR rate and an 85% 5-year DFS after treatment with surgery alone.168

In European and Scandinavian countries, the information gained from imaging, rather than clinical TNM

staging, is used to determine treatment in patients with rectal cancer.169

4 Tumor response to neoadjuvant therapy has prognostic value, and re-staging after neoadjuvant

1790

therapy is becoming increasingly important in reassessing treatment options and planning the surgical

procedure. Some patients with a clinical complete response are now offered alternatives to rectal

resection, such as “wait-and-see,” in the context of clinical trials.170,171 However, endoscopy and DRE

tend to underestimate tumor response to CRT.172 Morphologic imaging modalities such as ERUS and CT

provide a rough estimate of tumor regression, but cannot reliably distinguish post-treatment edema,

fibrosis, and necrosis from residual tumor.173 Similarly, MRI assessment of tumor response based on the

reduction of signal intensity relative to pre-treatment images, that occurs when tumor is replaced by

fibrosis, correlates poorly with pCR.174 Functional studies such as FDG-PET with or without

simultaneous CT, are valuable in assessing partial tumor response, but are not sensitive enough to

identify pCR.175 MRI dynamic sequences such as DWI and DCE, which provide an estimate of tissue

perfusion and cellularity, are currently under investigation in assessing rectal cancer response to CRT.176

Local Excision for Early-Stage Rectal Cancer

Selected patients with stage I rectal cancer, those with tumors localized to the bowel wall (T1 or T2)

and without involvement of the mesorectal nodes (N0), can be hypothetically cured with a LE of the

portion of the rectal wall containing the tumor. This operation spares most of the mortality, morbidity,

and bowel, urinary and sexual dysfunction associated with TME. However, the success of LE requires

meticulous patient selection. With LE, the mesorectum is not inspected for pathologic nodal staging.

Therefore, LE should only be offered to patients with a low risk of nodal metastasis. The selection

criteria for LE include: tumor size smaller than 3 cm, involving less than 30% of the circumference of

the rectum, mobile on DRE, localized to the submucosa on ERUS, without evidence of metastatic lymph

nodes on ERUS and CT or MRI, and no high-risk histologic features (i.e., grade 3 or 4, lymphovascular

or perineural invasion).177

In the past, only tumors located in the distal rectum could be treated with conventional transanal

excision (TAE), and LE was recommended only for tumors located within 8 cm from the anal verge.

With the use of endoscopic instrumentation and large operating proctoscopes (TEMTM, Richard Wolf

Medical Instruments Corp., Vernon Hills, IL; TEOTM, Karl Stortz Corp., El Segundo, CA) or single-port

devices (TAMIS, Trans-Anal Minimally Invasive Surgery), LE can now be easily performed for tumors

located in the mid and upper rectum. However, the benefit of LE is more significant for very distal

tumors that otherwise would require a low anastomosis or a permanent stoma. The real benefit of LE

for tumors located in the upper rectum is a matter of debate.

Independent of the approach, a full-thickness excision of the portion of the bowel wall involved by

tumor, with a 1-cm negative margin is required to avoid local recurrence (Fig. 68-14). The final decision

regarding the suitability of LE is made after the pathologic evaluation of the surgical specimen. If the

resection margins are positive, depth of invasion is beyond the submucosa (T stage ≥2), or the

histology reveals high-risk features (grade 3 or 4, lymphovascular or perineural invasion), the patient

should be offered immediate salvage TME.

Even in patients with tumors meeting all selection criteria, LE is associated with a higher rate of LR

compared to TME. The recurrence rates for T1 tumors treated with LE have been described as up to

23%.178–180 While some patients who develop LR after a failed LE are candidates for salvage TME, only

half of these patients are ultimately cured of their tumors. Recurrences after LE generally present as

more advanced tumors, often requiring extended pelvic dissections.181,182 The oncologic outcome for

these patients is worse than those treated originally with TME; this reinforces the importance of patient

selection for this strategy.

1791

Figure 68-14. Local excision of early rectal cancer. A: Marking resection margins. B: Incision up to mesorectal fascia. C: Resection

lodge. D: Defect closed with interrupted absorbable sutures.

A number of retrospective case series have compared the results of LE versus radical surgery (RS) for

patients with early-stage rectal cancer. A cohort study from the National Cancer Database (NCDB)

reported a 12.5% LR rate for LE, versus 6.9% for T1 tumors treated with RS (p <0.001). Five-year

overall survival was 77.4% for LE versus 81.7% for RS (p = 0.09).183 Other series have reported similar

results.184 However, comparisons between LE and RS are limited by selection bias, since patients having

LE tend to be older, have more comorbid conditions, and lower-lying tumors; whereas patients having

RS tend to be younger and healthier with higher-lying – but usually larger – tumors.

A recent systematic review of one small randomized trial and 12 observational studies, including a

total of 2,855 patients with T1N0M0 rectal cancer treated by LE (TAE/TEM-TEO) versus RS, showed

that LE was associated with lower perioperative mortality (relative risk 0.31, 95% CI 0.14–0.71), major

postoperative complications (relative risk 0.20, 95% CI 0.10–0.41), and need for a permanent stoma

(relative risk 0.17, 95% 0.09–0.30). The LR rate was higher for patients treated with LE, and overall

survival significantly lower, compared to patients treated with RS. The relative risk of dying after LE

was 1.46 (95% CI 1.19–1.77), which corresponds to 72 more deaths per 1,000 patients at 5 years. Metaregression showed that the difference in overall survival observed between groups could be explained

by the higher use of LE for tumors located in the lower third of the rectum, which is associated with

worse prognosis. Survival was similar between groups when only tumors located in the distal rectum

were included in the analysis.185

A number of studies have compared the outcomes of different techniques – conventional transanal

excision, TEMS, TEO, and TAMIS – for LE of rectal cancer, with most comparisons limited to

conventional TAE and TEMS. A recent meta-analysis of 6 retrospective case series including 927

procedures, reported that TEMS was associated with less specimen fragmentation, less positive resection

margins, and lower LR rates. No difference was observed in complication rate.186

In summary, the literature suggests that LE may be an alternative to TME for patients with small,

distal T1N0 rectal cancers, without high-risk histologic features, that would otherwise require a CAA or

an APE. Transanal endoscopic techniques may be superior to conventional transanal excision. Patients

considering LE should be informed about the potential gains in improved quality of life, but also about

the higher risk of LR and the need for salvage surgery.

1792

Patients with T2N0 tumors are not candidates for treatment by LE alone, independent of tumor

histologic features. The rate of LR for these patients when treated with LE alone ranges from 13% to

47%, significantly higher compared to similar stage tumors treated with TAR. Five-year overall survival

is also significantly lower for patients with T2N0 tumors treated with LE (65% to 67.6%) compared to

TAR (76.7% to 81%).183,187 A recent survey of the NCDB found that patients with T2N0 tumors treated

with LE alone had a higher hazard ratio for death compared to similar patients treated with TAR (1.535,

95% CI 1.283–1.835).188 Similar results have been reported in an independent cohort from the SEER

database.189 Based on this evidence, LE alone cannot be recommended as an alternative to RS for

patients with T2N0 tumor treated with curative intent.

Radiation therapy has been used to reduce the rate of LR in patients with T2N0 rectal cancer treated

with LE. A number of retrospective case series have reported 15% to 21% LR rates and 66% to 76.1% 5-

year survival rates with LE followed by radiation or chemoradiation for T2N0 tumors.190–193 The CALGB

8984 trial investigated the role of LE followed by 5-FU-based adjuvant chemoradiation for patients with

T2N0 cancers, reporting a 66% (95% CI, 51% to 84%) overall survival and 64% (95% CI, 51% to 80%)

DFS.194 The trial compared the survival curves of patients with similar stage tumors treated with TME,

concluding that LE and postoperative chemoradiation may be an alternative for patients with T2N0

tumors. However, 32% of the patients initially accrued for this trial were ultimately excluded from

analysis due to tumor size, inaccurate staging, surgical specimen fragmentation, high-grade histologic

features, or involved resection margins. Therefore, the conclusion is applicable only for properly staged

early rectal cancers penetrating only into the muscularis propria (pT2), with full-thickness excision

negative resection margins, and no high-risk histologic features.

The results of several retrospective studies have suggested that chemoradiation before LE reduces the

risk of recurrence to rates similar to those observed in similar stage patients treated with TME

alone.195,196 However, these studies are limited by small sample size, variable clinical staging criteria

and imaging modalities, and the use of different CRT regimens. A retrospective study using the SEER

database showed that patients who had undergone neoadjuvant CRT and LE had equivalent oncologic

outcomes compared to patients who had a major resection.189 One prospective trial found that patients

with ERUS-staged T2N0 rectal cancer treated with neoadjuvant CRT had similar recurrence and survival

rates when randomized to LE or RS. More than one in four patients in the RS arm required a permanent

colostomy, and a similar number required a temporary ileostomy.197 The ACOSOG Z6014 trial, a single

arm study investigating the use of 5-FU and oxaliplatin-based neoadjuvant CRT and LE in T2N0 tumors,

recently reported a complete pathologic response in 44% of patients.198 Only 2 (3%) patients had

developed recurrence after 4 years of follow-up. The 3-year DFS was 87%, with a 3-year OS of 90%. At

the end of the follow-up 70 of 72 patients treated with this approach had organ preservation.199 In

summary, the optimal treatment of T2N0 rectal cancer is TME; LE alone is not an alternative to radical

resection in patients treated with curative intent. Postoperative CRT may be an alternative to TME in a

patient with a distal clinical stage T1N0 cancer who is unexpectedly found to have a T2 tumor after LE

and is interested in preserving the rectum. Neoadjuvant CRT followed by LE may be an alternative for

patients with very distal T2N0 tumors interested in organ preservation, who otherwise would require a

permanent colostomy.

LE and other forms of local therapy, such as endocavitary radiation, brachytherapy, or

electrofulguration, have been used for palliation in patients who have more advanced tumors or are

unfit for a major surgical procedure.

Transabdominal Resection for Rectal Cancer

Patient preparation for rectal cancer surgery is similar to that for colon cancer surgery. Stoma marking

is particularly important in these patients, as many will require a temporary or permanent stoma. In

addition to discuss fertility options with all individuals of child-bearing potential, female patients

receiving neoadjuvant radiation who are interested in preserving fertility should consult their

gynecologist regarding the possibility of ovarian transposition. Mechanical bowel preparation, antibiotic

and thromboembolic prophylaxis are similar to those described for colon cancer patients. Many

surgeons recommend a rectal washout immediately before surgery to eliminate exfoliated cancer cells

that could potentially implant in the anastomosis and cause LR. In fact, a recent meta-analysis with

5,012 patients found that this practice is associated with a lower incidence of LR.200 On the operative

table, the patient should be placed in lithotomy with the legs in stirrups to provide the surgeon with

simultaneous access to the abdomen and perineum. It is important to avoid pressure on the bony

prominences, which might cause peroneal nerve compression. Patients requiring prolonged

1793