Clinical manifestations, diagnosis, and staging of colorectal cancer
tables and images:
  • Endoscopic appearance of the normal sigmoid colonic mucosa. The fine vasculature is easily visible and the surface is shiny and smooth. The folds are of normal thickness: here
  • Adenocarcinoma of the colon may have a variety of appearances on endoscopy. Panel A: a typical exophytic mass; Panel B: a friable polypoid mass; Panel C: circumferential adenocarcinoma. here
  • Endoscopy showing necrotic colonic lesions that usually suggest an advanced stage of malignancy. Panel A: severe tissue destruction has led to necrosis and bleeding; Panel B: the longstanding tumor has extended deeply into the mucosa and become necrotic. here
  • Double-contrast barium enema shows an eccentric mass arising from the anterior wall of the rectum (arrow). here
  • Endoscopic ultrasonography showing a rectal carcinoma. The normal hypoechoic boundary of the muscularis propria is expanded and irregular (arrow), indicative of tumor infiltration into the perirectal tissue. here Ultrasonographic staging as a T3 cancer was confirmed by surgical pathology. here
  • Transrectal ultrasound image demonstrates a mass arising in the anterior wall of the rectum and extending into the muscularis propria layer (arrows). This finding is consistent with a stage T2 tumor here
  • Prognosis rectal cancer. here
  • Prognosis stage II rectal cancer.  here
INTRODUCTION Colorectal cancer (CRC) is a common and lethal disease. Approximately 148,610 new cases of large bowel cancer are diagnosed each year in the United States, of which 106,680 are colon and the remainder rectal cancers. In 2006, more than 55,170 Americans will die of large bowel cancer.

Surgical resection is the primary treatment modality for CRC, and outcome is most closely related to the extent of disease at presentation.. Screening with fecal occult blood testing and flexible sigmoidoscopy can lead to diagnosis at an earlier stage of disease, and thereby reduce cause-specific mortality.. However, compliance with screening guidelines is poor; in two-thirds of patients, the initial diagnosis of CRC is established after the onset of symptoms

CLINICAL MANIFESTATIONS The majority of patients with CRC have hematochezia or melena, abdominal pain, and/or a change in bowel habits  Two reports evaluated the frequency of initial symptoms in patients presenting with potentially resectable colon cancer. Some patients have more than one abnormality:

  • Abdominal pain 44 percent
  • Change in bowel habit 43 percent
  • Hematochezia or melena 40 percent
  • Weakness 20 percent
  • Anemia without other gastrointestinal symptoms 11 percent
  • Weight loss 6 percent

Abdominal pain can be caused by a partial obstruction, peritoneal dissemination, or intestinal perforation leading to generalized peritonitis. In addition, tenesmus caused by a rectal cancer may involve pelvic floor muscles, and a locally advanced lesion may involve the sciatic or obturator nerve, leading to a neuropathic pain syndrome.

Hematochezia is more often caused by a rectal rather than colon cancer. Iron deficiency anemia from blood loss is most commonly associated with a delayed diagnostic evaluation. Occult colonic bleeding is site, but not stage, dependent: cecal and ascending colon tumors have a fourfold higher mean daily blood loss (approximately 9 mL/day) than tumors at other colonic sites

A change in bowel habits is a more common presenting symptom for left-sided cancers. In contrast to right-sided lesions which tend to be polypoid or fungating, tumors involving the distal or left colon are more commonly annular or encircling lesions that produce an "apple-core" or "napkin-ring" appearance). The bowel lumen becomes progressively constricted and narrowed, producing constipation, diarrhea, or bowel obstruction. Other possible symptoms of CRC include abdominal distention, nausea, vomiting, weight loss, and fatigue.

Metastatic disease Approximately 15 to 20 percent of patients have distant metastatic disease at the time of presentation. CRCs can spread by lymphatic and hematogenous dissemination, as well as by contiguous and transperitoneal routes. The most common metastatic sites are the regional lymph nodes, liver, lungs, and peritoneum, and patients may present with signs or symptoms referable to any of these areas. The presence of right upper quadrant pain, abdominal distention, early satiety, supraclavicular adenopathy, or periumbilical nodules usually signals advanced disease.

Because the venous drainage of the intestinal tract is via the portal system, the first site of hematogenous dissemination is usually liver, followed by lungs, bone, and many other sites, including brain. However, tumors arising in the distal rectum may metastasize initially to the lungs because the inferior rectal vein drains into the inferior vena cava rather than into the portal venous system.

Unusual presentations There are also a variety of unusual presentations of CRC. These include:

  • Local invasion or a contained perforation causing malignant fistula formation into adjacent organs, such as bladder or small bowel. This is most common with cecal or sigmoid carcinomas; in the latter case, the condition can mimic diverticulitis.
  • Fever of unknown origin, intraabdominal, retroperitoneal, or abdominal wall abscesses.
  • Streptococcus bovis bacteremia and Clostridium septicum sepsis are due to underlying colonic malignancies in 10 to 25 percent of patients
  • CRC ultimately proves to be the origin of approximately 6 percent of adenocarcinomas of unknown primary sites

Impact of symptoms on prognosis The presence of symptoms, and their particular type appear to be of some prognostic importance:

  • Patients who are symptomatic at diagnosis have a somewhat worse prognosis. In one report, the five-year survival rate for symptomatic and asymptomatic patients was 49 versus 71 percent. The total number of symptoms may be inversely related to survival for colon, but not for rectal cancer. Surprisingly, however, the duration of symptoms is not an accurate predictor of prognosis.
  • Obstruction and/or perforation, although uncommon, carry a poor prognosis, independent of stage
  • Tumors presenting with hemorrhage have been thought to have a better prognosis because of their tendency to be diagnosed earlier; however, bleeding is not an independent predictor of outcome

Other determinants of prognosis, including clinicopathologic and molecular features, are discussed elsewhere.

LOCATION OF COLORECTAL MALIGNANCIES The location of the primary tumor may also have prognostic significance. For each stage, cancers arising at or below the peritoneal reflection (rectosigmoid and rectum) have a worse five-year survival rate than those arising more proximally. Furthermore, within the rectum, distal cancers have a worse prognosis than more proximal lesions.

Over the last 50 years, a gradual shift toward right-sided or proximal colon cancers has been observed both in the United States  and internationally. As an example,data obtained from the National Cancer Data Base, suggest that the incidence of tumors proximal to the splenic flexure increased from 51 to 55 percent of all colonic lesions between 1988 and 1993. The greatest increase in incidence is in cecal primaries

Synchronous cancers Synchronous CRCs, defined as two or more distinct primary tumors separated by normal bowel and not due to direct extension or metastasis, occur in 3 to 5 percent of patients with colon cancer. The incidence is somewhat lower (about 2.5 percent) when patients with hereditary nonpolyposis colorectal cancer (see below) are excluded. Synchronous primaries have the same prognosis as solitary malignancies when the highest stage of disease is compared.

Metachronous cancers Metachronous CRCs, defined as nonanastomotic new tumors developing at least six months after the initial diagnosis, develop in 1.5 to 3 percent of patients in the first five years postoperatively. This topic is discussed in detail elsewhere.

DIAGNOSIS CRC may be suspected from one or more of the symptoms and signs described above, or may be asymptomatic and discovered by routine screening of average and high risk subjects.

Colonoscopy The majority of colon and rectal cancers are endoluminal adenocarcinomas that arise from the mucosa. Colonoscopy is the single best diagnostic test in symptomatic individuals, since it can localize lesions throughout the large bowel, biopsy mass lesions, detect synchronous neoplasms, and remove polyps, The air contrast barium enema, supplemented with flexible sigmoidoscopy, is also used, but the diagnostic yield of this combination is less than that of colonoscopy for the evaluation of lower tract symptoms

In the 5 to 10 percent of patients in whom the colonoscope is not able to reach the tumor for technical reasons (eg, tortuous colon, poor prep), a double-contrast barium enema can provide a radiographic diagnosis. It is essential that the entire large bowel be examined for the presence of synchronous lesions. If malignant obstruction precludes a full colonoscopy preoperatively, it should be carried out at a later time. Virtual colonoscopy using CT (CT colonography) or MRI technology is under study in this setting

Differential diagnosis The majority of CRCs are adenocarcinomas.  Many conditions can cause signs or symptoms that are similar to CRC, including other malignancies, and a multitude of benign lesions such as hemorrhoids, diverticulitis, infection, or inflammatory bowel disease. A direct evaluation of the total colonic mucosa is necessary to exclude carcinoma with certainty.

Metastases from other primary cancers, most often ovarian cancer, can mimic a primary large bowel malignancy.

STAGING Once the diagnosis of CRC is established, the local and distant extent of disease spread should be determined in order to provide a framework for discussing therapy and prognosis. A review of the biopsy specimen is important prior to making a decision about the need for clinical staging studies and surgical resection, especially for a cancerous polyp. Polyps that contain an area of invasive malignancy without associated adverse histologic features have a low risk of lymphatic and distant metastases; in such patients, clinical staging is unnecessary, and polypectomy alone may represent appropriate therapy.

Clinical versus pathologic staging The Duke's classification was originally developed and then modified by Astler-Coller to provide an organized method to classify disease extent for treatment planning, estimating prognosis, and measuring therapeutic outcomes. Subsequently, the TNM system evolved, and the most recent 2002 version (which divides stage III disease into prognostically distinct A, B, and C categories depending upon depth of the primary tumor and number of involved nodes) is currently the preferred staging system for CRC

Although radiographic, endoscopic, and intraoperative findings can be used to assign a clinical stage, assessment of the pathologic stage (termed pT, pN, pM) requires histologic examination of the resection specimen.

At initial presentation, the distribution of disease extent for colon cancer is as follows

  • Localized to the mucosa and submucosa (Dukes' A or TNM stage I) 23 percent
  • Extending into or through the muscle layer without lymph node involvement (Dukes' B or TNM stage II) 31 percent
  • Lymph node involvement (Dukes' C or TNM stage III) 26 percent
  • Distant metastases (Dukes' D or TNM stage IV) 20 percent

The distribution of disease extent for rectal cancers in the United States is as follows]:

  • Localized to the mucosa and submucosa (Dukes' A or TNM stage I) 34 percent
  • Extending into or through the muscle layer without lymph node involvement (Dukes' B or TNM stage II) 25 percent
  • Lymph node involvement (Dukes' C or TNM stage III) 26 percent
  • Distant metastases (Dukes' D or TNM stage IV) 15 percent

Clinical staging evaluation Preoperative clinical staging is best accomplished by physical examination (with particular attention to ascites, hepatomegaly and lymphadenopathy), computed tomography (CT) scan of the abdomen and pelvis, and chest radiograph. Although frequently obtained during the preoperative evaluation, liver enzymes may be normal in the setting of small hepatic metastases, and are not a reliable marker for liver involvement . The single most common liver test abnormality associated with liver metastases is an elevation in the serum alkaline phosphatase level

CT scan In patients with newly diagnosed CRC, preoperative abdominal and pelvic CT scans can demonstrate regional tumor extension, regional lymphatic and distant metastases, and tumor-related complications (eg, obstruction, perforation, fistula formation). The sensitivity of CT for detecting distant metastasis is higher (75 to 87 percent) than for detecting nodal involvement (45 to 73 percent) or the depth of transmural invasion (approximately 50 percent). Compared to colon cancers, the sensitivity of CT for detection of malignant lymph nodes is higher for rectal cancers; any perirectal adenopathy is presumed to be malignant since benign adenopathy is not seen in this area

The necessity of preoperative CT scans for all patients with CRC is a matter of debate. In one retrospective review of 180 patients who underwent resection, only three of 67 patients had incidental findings on CT scan that altered the surgical approach. Assessment of hepatic metastases by intraoperative ultrasound with manual palpation of the liver may provide a better yield than preoperative CT, particularly for those patients who are found to have transmural tumor involvement (T3/4) at the time of exploration

Furthermore, the finding of liver metastases on preoperative radiographic evaluation may not alter the surgical approach to the primary tumor, particularly in patients who are symptomatic (eg, bleeding, impending obstruction). In patients with four or fewer hepatic lesions, resection may be curative, with five-year relapse-free survival rates of 24 to 38 percent.. Although most surgeons advocate resection of the primary tumor and synchronous hepatic metastases at two different operations, some approach both sites at the same time

MRI The efficacy of magnetic resonance imaging (MRI) for preoperative staging is similar to that of CT and is not routinely recommended. However, contrast-enhanced MRI may identify more hepatic lesions than are visualized by CT, and potentially narrow the available therapeutic options for patients with suspected liver metastases. Furthermore, preoperative MRI may prove to be of benefit for patients with rectal cancer because of its ability to identify mesorectal nodal involvement by characteristics other than size.

PET scans PET scans do not appear to add significant information to CT scans for preoperative staging of CRC. The established role of PET scanning in colorectal cancer is as an adjunct to other imaging modalities in the following settings:

  • Localizing site(s) of disease recurrence in patients who have a rising serum CEA level and nondiagnostic conventional imaging evaluation following primary treatment
  • Evaluation of patients who are thought to be candidates for resection of isolated colorectal cancer liver metastases

In one series of 51 patients with apparently isolated liver metastases, clinical management decisions based upon conventional diagnostic imaging were altered by PET scan findings in 10 (20 percent); six had unsuspected extrahepatic disease. Eight patients were spared an unnecessary laparotomy, while two others were identified as candidates for resection. Similar results have been noted by others

For patients with symptoms or a rising serum CEA level who are suspected of having metastatic disease but whose diagnostic workup is negative, PET scanning can potentially localize occult disease, permitting the selection of patients who may benefit from exploratory laparotomy . In one series,105 such patients underwent PET scanning and subsequent abdominopelvic CT scans. Compared to CT and other conventional diagnostic studies, PET scanning had a higher sensitivity (87 versus 66 percent) and specificity (68 versus 59 percent). In a second report, PET scan findings led to a potentially curative resection in 14 of 50 patients (28 percent) with elevated serum CEA levels and a completely normal or equivocal conventional diagnostic work-up

One caveat is that recent chemotherapy may alter the sensitivity of PET for the detection of colorectal metastases, thought related to decreased cellular metabolic activity of the tumor [55]. In a series of 42 patients undergoing surgery for colorectal cancer hepatic metastases, 15 of 41 liver lesions (37 percent) were undetected by PET among the 13 who had received preoperative chemotherapy. Three of these 13 patients had lesions confirmed pathologically (one single, two multiple liver lesions) that were all undetectable by PET. In contrast, of the 29 patients who did not receive preoperative chemotherapy, only 16 of 69 lesions (23 percent) were undetected by PET. Thus, the results of restaging PET scans (particularly if negative) must be interpreted in the context of recent therapy.

Intraoperative evaluation Even if the staging evaluation fails to show evidence of metastatic spread, intraoperative evaluation is an essential component of the clinical staging process. In particular, CT scan is not a reliable diagnostic test for low volume tumor on peritoneal surfaces

EUS and rectal cancer Preoperative knowledge of the depth of invasion and nodal status is critically important for planning therapy of a rectal cancer. Neoadjuvant combined modality approaches utilizing both chemotherapy and radiation are associated with less toxicity and a higher likelihood of sphincter preservation, particularly for distal transmural tumors., section on Pretreatment staging). Both transrectal or endorectal ultrasound (EUS) and MRI with endorectal coils can demonstrate the various layers of the rectal wall, but the EUS is less expensive and less time- consuming.

EUS exceeds the capability of CT scan to locally stage the depth of transmural invasion and the presence of perirectal nodal involvement for rectal cancers. The accuracy of EUS for correctly predicting tumor (T) stage ranges from 80 to 95 percent, compared to 65 to 75 percent for CT, and 75 to 85 percent for MRI. However, EUS has been less successful at correctly predicting nodal (N) status. A number of comparative studies have suggested that the accuracy of EUS for predicting nodal status (approximately 70 to 75 percent) is similar to that of CT (55 to 65 percent) and MRI (60 to 65 percent). These data are discussed in detail elsewhere.

The addition of EUS-guided fine needle aspiration (FNA) biopsy improves the accuracy of N staging

Tumor markers A variety of serum markers have been associated with CRC, particularly carcinoembryonic antigen (CEA), and carbohydrate antigen (CA) 19 9. However, these markers have a low diagnostic ability to detect primary CRC due to significant overlap with benign disease, and low sensitivity for early stage disease. Noncancer-related causes of an elevated CEA include gastritis, peptic ulcer disease, diverticulitis, liver disease, chronic obstructive pulmonary disease, diabetes, and any acute or chronic inflammatory state

An expert panel on tumor markers in breast and colorectal cancer convened by the American Society of Clinical Oncology (ASCO) recommended that serum CEA nor CA 19 9 levels not be used as a screening test for colorectal cancer. A similar recommendation has been made by the European Group on Tumor Markers.

However, serum levels of CEA do have prognostic utility in patients with newly diagnosed CRC. Patients with preoperative serum CEA > 5 ng/mL have a worse prognosis, stage for stage, than those with lower levels.  Furthermore, elevated preoperative CEA levels that do not normalize following surgical resection implies the presence of persistent disease and the need for further evaluation.  ASCO guidelines recommend that serum CEA levels be obtained preoperatively in patients with demonstrated colorectal cancer to aid in staging, surgical treatment planning, and in the assessment of prognosis

PROGNOSIS Despite the enormous number of correlative studies exploring the prognostic significance of various histologic, molecular, and clinical features, the pathologic stage at diagnosis remains the best indicator of long-term prognosis for both colon and rectal cancer. The most important characteristics are the presence of distant metastases, local tumor extent, nodal positivity (particularly the number of involved lymph nodes), and residual disease.

Colon cancer Five-year survival rates in a contemporary series of over 119,000 patients treated between 1991 and 2000 stratified according to the most recent modification of the TNM staging system were as follows

  • Stage I (T1-2N0) 93 percent
  • Stage IIA (T3N0) 85 percent
  • Stage IIB (T4N0) 72 percent
  • Stage IIIA (T1-2 N1) 83 percent
  • Stage IIIB (T3-4 N1) 64 percent
  • Stage IIIC (N2) 44 percent
  • Stage IV 8 percent

Rectal cancer Five-year survival rates for rectal cancer tend to be somewhat lower, and are heavily dependent upon case mix

In particular, the survival of patients with stage III disease is variable, and depends upon the T stage (ie, T1-2 versus 3-4) and the extent of nodal disease. As an example, when data from the National Cancer Database were stratified according to subsets of stage III disease, the following five-year survival rates were reported

  • Stage IIIA (T1-2, N1) 55.1 percent (n = 1043 patients)
  • Stage IIIB (T3-4, N1) 35.3 percent (n = 2856 patients)
  • Stage IIIC (Any T, N2) 24.5 percent (n = 2088 patients)

Similar results were noted in a pooled analysis of data from five randomized trials of adjuvant therapy performed by United States Cooperative Groups.  The recognition that these subgroups of stage III disease have different prognoses was reflected in the 2002 modification of the TNM staging system for rectal cancer, in which stage III disease was subdivided into A, B, and C categories

These outcome estimates were derived from studies in which surgical resection was not preceded by neoadjuvant therapy. Outcomes may differ in patients who undergo pathologic staging after preoperative treatment, and cannot be estimated with certainty until more data become available.

Although not yet incorporated into the TNM staging system, further stratification of outcome can be achieved by the identification of patients with lymphovascular invasion, and high preoperative serum CEA levels. Preliminary studies suggest that the molecular features of microsatellite instability and loss of the Deleted in Colon Cancer (DCC) gene may also influence outcome, independent of stage at presentation. However, until large, statistically robust studies are completed with multivariate analysis of an array of potential prognostic factors, and their interaction with each other, neither these nor other molecular or genetic markers should be used routinely to develop treatment recommendations or to estimate prognosis in patients with resected CRC. These issues are discussed in detail elsewhere.