ESTADIFICACION

Standard clinicopathological staging is the best indicator of prognosis for patients with rectal cancer. For rectal

cancer, it is increasingly common to use clinical staging as the basis for the decision for neoadjuvant

chemoradiation therapy. Therefore, the accuracy of that initial staging is critically important, both for

management and for prognosis. There have been a large number of studies that have evaluated other

prognostic markers, including pathologic, socioeconomic, molecular, as described more fully in the chapter on

colon cancer. However, even though many of these appear to have prognostic value, there are none that are

commonly used to define management. This is related to the large number of tests that could be used, the lack

of standardization of these tests, as well as the lack of knowledge as to how to incorporate them into the

patient management scheme. The molecular marker that has engendered the most interest is the deletion of

18q.7 These markers have been fully reviewed elsewhere.8,9

The staging system that should be used in the evaluation of patients with rectal cancer is the American Joint

Committee on Cancer/International Union Against Cancer TNM (tumor, node, metastases) staging system (fully

described in Chapter 12), which has been recently revised to subcategorize patients with stage III (node

positive) tumors. The Dukes staging system or its multiple modifications has been used for many years, but

provides less information than the TNM system and should not be used. There have been gradual changes in the

TNM system that primarily reflect the stage grouping rather than the system itself. The other systems should

be acknowledged for their historical interest and for initially defining many of the high-risk factors for this

disease.

Patients now often have both a clinical (preoperative) staging, which may define the use of neoadjuvant

therapy, and a postoperative surgical stage. However, it is important to remember that initial therapy with

radiation and chemotherapy can produce substantial down-staging (approximately 15% of patients will have a

pathological complete response), and that subsequent therapy should be based on the initial T and N staging

determination. Specifically, a good tumor response locally

to radiation and chemotherapy does not mean that a patient has any lower risk of having micrometastatic

disease, and thus does not have a lesser need for adjuvant postoperative chemotherapy. Put another way, in a

patient who receives preoperative radiation and chemotherapy, there is no decision point regarding whether or

not postoperative chemotherapy should be given on the basis of the surgical pathology result. Until the data

demonstrate otherwise, the plan for postoperative chemotherapy should be carried out even in the setting of a

pathological complete response.

The major change that has occurred in the newest version of the staging system is the acknowledgment that

both the T stage and the N stage have independent prognostic importance for local control, disease-free

survival, and overall survival.10,11 Thus, for N0 and N1 patients viewed separately, the extent of the primary

tumor in the rectum is of additional prognostic importance. Patients with T1-2N1 tumors have a relatively

favorable prognosis and an outcome superior to that of other stage III patients. In fact, patients with T3N0M0

disease (stage II) have outcomes slightly inferior to those with T1-2N1M0, demonstrating the independent

prognostic importance of T stage. These distinctions may allow future decisions to be more individualized as to

the adjuvant therapy required.

Although at one level staging is very straightforward, the actuality of proper staging is much more difficult as it

relies on multiple quality control issues that can mislead the clinician regarding proper therapy. For instance,

it has been well demonstrated that for patients who are pathologically staged as N0 that the prognosis is

markedly improved for those in whom more than 12 to 14 nodes were identified by the pathologist compared

to those in whom fewer nodes were identified.12 This could be a surgical issue (fewer nodes were removed) or

a pathologist issue (fewer nodes were identified), but it suggests that many patients were inappropriately

staged lower, which could result in inappropriate therapy. Others have shown that staging accuracy continues

to improve as the pathologist recovers more nodes, with accuracy leveling off at approximately 12 to 20 nodes

recovered.13,14 Likewise, as is true for colon cancer, the percentage of positive nodes is of substantial

prognostic importance (M. Meyers, 2007, personal communication). The same issue relates to T-stage

determination. If the pathologist does not look carefully for evidence of extension of tumor through the

muscularis propria, the patient can be understaged, resulting in inappropriate treatment. Close or positive

circumferential margins are a poor prognostic factor, which can only be found if the pathologist assiduously

evaluates the radial margins.15,16

The standard staging procedure for rectal cancer entails a history, physical examination, complete blood cell

count, liver and renal function studies, as well as carcinoembryonic antigen (CEA) evaluation. The routine

laboratory studies are quite insensitive to the presence of metastatic disease, but they are usually ordered as a

screen of organ function prior to surgery or chemoradiation therapy. High CEA levels are associated with poorer

survival (see Chapter 12) and give an indication as to whether follow-up CEA determinations are likely to be

useful. A careful rectal examination by an experienced examiner is an essential part of the pretherapy

evaluation in determining distance of the tumor from the anal verge or from the dentate line, involvement of

the anal sphincter, amount of circumferential involvement, clinical fixation, sphincter tone, and so forth, and

has not been replaced by imaging studies or endoscopy. Colonoscopy or barium enema to evaluate the

remainder of the large bowel is essential (if the patient is not obstructed) to rule out synchronous tumors or

the presence of polyp syndromes.

Imaging studies including computed tomography (CT) or magnetic resonance imaging (MRI) to evaluate the

pelvis, abdomen, and liver as well as a chest x-ray to screen for pulmonary metastases are routine. There has

been much debate about the relative value of CT versus MRI without a clear resolution. This decision is heavily

dependent on the institutional expertise and the equipment available. As the technology continues to change

with fine-cut 64 detector row CT scans, improved MRI contrast agents, and so forth, one technique may

become slightly better than another.

Conventional CT lacks sufficient accuracy to be used for preoperative staging of the primary tumor site. For

example, in one series using air insufflation of the rectal ampulla and intravenous contrast, the overall T-stage

accuracy was approximately 80%, with 18% overstaging in T2 disease and 21% understaging in T3 tumors.17 As

radiologists have usually defined node positivity on CT based on size (typically 1 cm or more) the overall

accuracy of N staging by CT scan has been less than 80%, as nodes involved with rectal cancer are usually not

markedly enlarged. The same study demonstrated the expected tradeoff between sensitivity and specificity,

but with overall accuracy in a similar range. CT colonography is of increasing interest for screening, but at the

present time has a minimal role in staging patients with known disease.

MRI suffers from some of the same limitations as CT for evaluating the primary tumor, although endorectal coil

MRI allows discernment of the layers of the bowel wall and is similar in accuracy to endorectal ultrasound (see

below). Thin section pelvic MRI with a surface coil allows for better visualization of the rectal wall layers and

allows one to visualize the mesorectal fascia and thus to predict the likely distance of the surgical resection

margin when using a total mesorectal excision (TME). Although there has been great interest in this technique,

the studies to date still show a disappointing overall accuracy. In one study of 96 patients who had MRI

followed by TME, of 22 patients classified as MRI T2 disease, 3 had T1 and 6 had T3 tumors. Of 61 patients

P.1288

classified as having T3 disease on MRI, 8 had T2 tumors and 2 had T4. Thus, 6 of 22 (27%) patients who might

have benefited from preoperative therapy for T3 disease would not have received that therapy. Eight of 61

patients (13%) would have received preoperative treatment inappropriately based on the MRI T stage.18 For

nodal status, 8 of 33 MRI positive nodes were clinically negative, and 7 of 57 MRI negative nodes were

pathologically positive.19 The presence of nodal disease identified by MRI is also primarily determined by size,

so the accuracy is similar to that of CT, although defining node positivity based on irregular border or mixed

signal intensity could help improve sensitivity and specificity.18

CT, MRI, and positron emission tomography (PET) are all useful for detecting metastatic disease, primarily in

the liver. CT has an overall sensitivity of 70% to 85%,20 which might be improved with multidetector improved

CT technology, although the data do not yet prove that contention. MRI is superior in characterizing liver

lesions and distinguishing cysts and hemangiomas from tumor,21 especially with the use of enhancement with

gadolinium or other agents. PET with 18F-fluorodeoxyglucose has engendered a great deal of interest as a

method of better defining patients with metastatic disease, especially in abdominal lymph nodes where CT and

MRI are relatively insensitive. However, PET is not standard in the preoperative staging at most institutions,

and the incremental gain from routine PET scan appears to be small.22 Newer studies have suggested that a

substantial number of patients will have management altered by the use of staging PET. Gearhart et al.23 have

suggested a 27% rate of change of treatment plans based on PET results, with most of this due to the presence

of lymph node metastases. PET clearly can be of value in restaging patients with recurrence or suspected

recurrence to detect additional metastatic sites prior to attempted resection of metastatic disease.

PET shows promise as the most sensitive study for the detection of metastatic disease in the liver and

elsewhere. A meta-analysis of whole body PET showed a sensitivity of 97% and a specificity of 76% in evaluating

for recurrent colorectal cancer.24 In the United States at the present time contrast CT of the pelvis and the

abdomen is the most commonly used imaging study, with MRI or PET being used to clarify abnormalities noted

in the liver or abdomen. However, primary use of MRI is acceptable and could become standard with changing

technology and availability.

There is much interest in the use of endorectal ultrasound (EUS) for staging of the primary tumor, and this, at

present, is the most effective preoperative staging technique for T and N stage. Endorectal MRI provides

similar information, but is not generally available and will not be discussed further here. EUS defines five

interface layers of the rectal wall: mucosa, muscularis mucosa, submucosa, muscularis propria, and perirectal

fat, as shown in (Fig. 39.13.2). Rectal tumors are generally hypoechoic and disrupt the interfaces dependent on

the level of tumor extension. The accuracy of EUS is heavily dependent on the experience and skill of the

operator. The results mentioned below will not be obtained by an inexperienced examiner. However, in

experienced hands EUS has an overall accuracy rate for T stage of 75% to 95% with an overstaging of

approximately 10% to 20% in T2 disease due to an inability to distinguish a desmoplastic response and

postbiopsy changes from local tumor invasion, and approximately a 10% rate of understaging because of an

inability to detect microscopic tumor extension.25,26,27

Figure 39.13.2. Endorectal ultrasound of a T3 tumor of the rectum, extension through the muscularis

propria, and into perirectal fat. (From Ginsberg GG, Ahmad N. Endoscopic ultrasound for rectal cancer.

Vis Hum J Endosc 2003;2(2), with permission.)

EUS is less accurate in determining N stage than for T stage, with an overall accuracy rate of 62% to 83%.25,26

Understaging occurs because many nodal metastases from rectal cancer are small, even micrometastatic, and

not easily detected by EUS. In addition, some nodes are located beyond the range of the ultrasound transducer

and thus cannot be seen during the procedure. Overstaging is often due to an inflammatory response, perhaps

secondary to previous biopsy or manipulation. EUS is not accurate for determining tumor regression after

preoperative radiation therapy and chemotherapy, as inflammatory changes and scarring can persist in the

rectal wall or in perirectal soft tissue and may not reflect persisting tumor. Newer ultrasound techniques, such

as three-dimensional ultrasound, are being explored but have not yet made it into standard practice.