Colorectal Cancer Screening and Surveillance: Current Standards and Future Trends
- First Online:
- Cite this article as:
- Martinez, S.R., Young, S.E., Hoedema, R.E. et al. Ann Surg Oncol (2006) 13: 768. doi:10.1245/ASO.2006.03.087
- 57 Views
Its prevalence, long premalignant course, and favorable response to early intervention make colorectal cancer an ideal target for screening regimens. The success of these regimens depends on accurate assessment of risk factors, patient compliance with scheduled visits and tests, and physician knowledge of screening strategies. We review the current recommendations for colorectal cancer screening in general and at-risk populations, comment on surveillance methods in high-risk patients, and examine current trends that will likely influence screening regimens in the future.
KeywordsColorectal cancerScreeningSurveillanceRisk factors
In 2006, an estimated 148,610 Americans will receive a diagnosis of colorectal cancer.1 Colorectal cancer is the second most common cause of cancer-related mortality, with an estimated 55,170 deaths in the United States annually. However, its mortality rate has been decreasing over the last decade, and this may be attributable to more effective screening and surveillance.
NONHEREDITARY COLORECTAL CANCER
Sporadic cases represent approximately 85% of colorectal cancers in the United States. Typically, these individuals are older than those with hereditary variants of colorectal cancer, and they lack a family history of colorectal cancer. The lifetime risk of developing nonhereditary colorectal cancer is 6%.3 Approximately 80% of colorectal cancers arise from adenomatous polyps. If these polyps are identified and treated early, malignant transformation can be avoided.
Fecal Occult Blood Testing
American Cancer Society guidelines for colorectal cancer screening in patients 50 years of age or older
Annual 6-sample, at-home fecal occult blood test (FOBT) or fecal immunochemical test (FIT) plus flexible sigmoidoscopy every 5 ya,b or
Complete colonoscopy every 10 y or
Double-contrast barium enema every 5–10 yb or
FOBT or FIT alone every yearb or
Sigmoidoscopy alone every yearb
In a recent prospective study of 3121 asymptomatic patients at 13 Veterans Affairs medical centers, Collins et al.8 reported that take-home FOBT using 6 sample cards was superior to in-office FOBT using 1 sample card. The sensitivity for detecting advanced neoplasia was 23.9% for take-home FOBT versus 4.9% for in-office FOBT. Unfortunately, a significant proportion of physicians use the less sensitive test. Nadel et al.9 surveyed 1,147 primary care physicians who ordered or performed FOBT and 11,365 adults who had undergone FOBT. The less sensitive single-card FOBT was the test of choice for 32.5% of physicians. Perhaps worse, a positive FOBT was followed up with inadequate testing, such as repeat FOBT or sigmoidoscopy, in approximately 30% of patients. Any positive FOBT should be followed up with a complete colonoscopy.
Despite these troubling data, FOBT remains an important part of a successful colorectal cancer screening regimen. The test must be used properly, however, and positive results must be acted on accordingly. Physicians and patients should abandon the single in-office FOBT and use the take-home, six-sample FOBT instead. For maximal sensitivity, FOBT should be used in conjunction with screening sigmoidoscopy.
Flexible sigmoidoscopy is less expensive than colonoscopy and requires neither a full-bowel preparation nor sedation; moreover, it can be performed in an office setting and it is associated with fewer iatrogenic injuries. Historically, 70% of colorectal cancers can be detected by using a 60-cm flexible sigmoidoscope. However, because sigmoidoscopy cannot evaluate more proximal lesions, many guidelines recommend combining sigmoidoscopy with FOBT every 5 years.
For more than two decades clinicians have proposed that the anatomical distribution of colon cancers has undergone a rightward shift, thereby increasing the percentage of colon cancers that cannot be evaluated with flexible sigmoidoscopy.10–14 Mamazza and Gordon10 detailed the anatomical distribution of 1044 cases of colon cancer over each 3-year period from 1955 to 1978. An increase in right colon lesions from 15.6% in the first 3-year period to 37.6% in the final 3-year period was statistically significant (P < .01). Recently, Mensink et al.14 reported changes in the incidence and anatomical distribution of colon cancers in The Netherlands during a 15-year period: between 1981 and 1996, the overall incidence of colon cancer increased, and the incidence of proximal colon cancers changed from 25% to 37%. These findings cast doubt on the role of sigmoidoscopy as a screening tool when used alone. Although data are currently insufficient to recommend discontinuing screening sigmoidoscopy, we recommend its use in conjunction with yearly FOBT beginning at age 50, in accordance with the American Cancer Society’s screening guidelines (Table 2). Although no randomized clinical trial data exist regarding the use of sigmoidoscopy as a screening tool, several case-control studies report significant reductions in colorectal cancer mortality.15–17 Selby et al.15 compared screening sigmoidoscopy in 261 patients who died of distal colorectal cancer and 868 control subjects matched for age and sex. Only 8.8% of those who died of colorectal cancer had received screening sigmoidoscopy within 10 years of their diagnosis, compared with 24.2% of the control group (odds ratio, .41; 95% confidence interval, .25–.69). Newcomb et al.17 demonstrated an even more impressive and durable decrease in the risk of developing distal colorectal cancer (odds ratio, .24; 95% confidence interval, .17–.33). The risk reduction was durable up to 16 years from the time of sigmoidoscopy, thus indicating that, perhaps, the recommendation of screening sigmoidoscopy every 5 years may be, if anything, overly aggressive.
Because progression from adenoma to carcinoma requires several years, effective screening measures have the potential to detect disease at multiple time points. Despite the perceived advantages of colonoscopy as a screening tool, many patients avoid it because of fear of or distaste for the procedure and its requisite bowel preparation or because of their physician’s lack of knowledge concerning appropriate screening guidelines.
Because it allows visualization of the entire colon, screening colonoscopy can be used to perform diagnostic and potentially therapeutic polypectomies. Experienced endoscopists can survey the entire colon in 92% to 97% of cases; polyps ≥1 cm are rarely missed.14,18 Efficient removal of premalignant polyps and earlier detection of colorectal cancer should improve survival, but no randomized clinical trial has shown that screening colonoscopy can decrease the mortality of colorectal cancer. Therefore, even though the American Cancer Society lists colonoscopy as an option for colorectal cancer screening, its use in the asymptomatic, low-risk population is somewhat controversial.
Double-Contrast Barium Enema
Before colonoscopy, DCBE was the preferred method of screening the colorectal mucosa, and this technique still has a lower rate of perforation and bleeding than the <1% associated with colonoscopy.19 However, whereas DCBE can detect 90% of colorectal polyps or cancers >1 cm,20 it may miss 20% to 50% of smaller lesions.20,21 The National Polyp Study Work Group22 found that, compared with colonoscopy, DCBE missed 68% of adenomas ≤.5 cm, 47% of adenomas between .6 and 1 cm, and 52% of adenomas >1 cm. Moreover, therapeutic colonoscopy is necessary to remove or biopsy polyps identified by DCBE. Colonoscopy is a better diagnostic and therapeutic tool and may be more cost-effective in patients who are symptomatic or at higher risk. DCBE provides a useful adjunct when colonoscopy cannot be completed or in addition to sigmoidoscopy.
Computed Tomographic Colonography (Virtual Colonoscopy)
Computed tomographic colonography, or virtual colonoscopy (VC), is the three-dimensional study of the colon with computed tomography. Patients undergo a bowel preparation the day before the procedure, and the colon is insufflated with air until fully distended. As with standard colonoscopy, sedation may be required. Intravenous contrast may supplement the examination, if requested.
A nonrandomized, multicenter study examined the sensitivity of VC in 615 patients who subsequently underwent standard colonoscopy.23 VC had a sensitivity of 39% and 55% for lesions at least 6 and 10 mm in diameter, respectively, but it missed 25% of cancers. Macari et al.24 compared results of VC and same-day conventional colonoscopy in 68 asymptomatic men; the overall sensitivity of VC was 21%. Not surprisingly, its sensitivity was directly related to lesion size: it was 11.5%, 52.9%, and 100% for detecting lesions 1 to 5 mm, 6 to 9 mm, and >10 mm, respectively.
More favorable results were reported by Pickhardt et al.25 in a nonrandomized study of 1233 asymptomatic adults undergoing same-day VC and standard colonoscopy. The sensitivity of virtual versus standard colonoscopy was 93.8% vs. 87.5% for polyps at least 10 mm in diameter, 93.9% vs. 91.5% for polyps at least 8 mm in diameter, and 88.7% vs. 92.3% for polyps at least 6 mm in diameter. The specificity of VC for adenomatous polyps was 96.0%, 92.2%, and 79.6% for lesions at least 10, 8, and 6 mm in diameter, respectively.
These disparate results reflect variations in interpretation and emphasize the need for randomized multicenter trials based on standardized protocols. At present, VC is an investigative tool that cannot be considered a viable colorectal cancer screening option.
In 1992, Sidransky et al.26 published the first report of K-ras mutations in colorectal cells obtained from stool specimens. Optimism for this assay was tempered by the discovery that up to 14% of patients without colorectal cancer may demonstrate K-ras mutations in fecal DNA.27 Since then, several investigators have proposed various panels of molecular markers to detect colorectal cancer via fecal analysis. It has been suggested that multimarker molecular fecal analysis may be preferable to single-marker analyses and FOBT, both of which are prone to false-positive results.
Dong et al.28 were able to detect 71% of cases of colorectal cancers by assessing the mutation status of p53, BAT26, and K-ras. By adding mutations of the adenomatous polyposis coli (APC) gene to this molecular panel, Ahlquist et al.29 improved the colorectal cancer detection rate to 91%. Koshiji et al.30 used a panel of seven microsatellite markers for APC, p53, deleted in colorectal carcinoma (DCC), mut L homolog 1 (hMLH1), D9S162, and interferon alpha (IFNA) to examine stool specimens from 41 colorectal cancer patients. Sensitivity and specificity were maximal (96.7% and 100%, respectively) with a combination of APC and p53.
Imperiale et al.31 compared standard FOBT with fecal DNA testing on the basis of a panel of 21 different mutations. FOBT detected 14.1% of invasive cancers, adenomas, and high-grade dysplasias, whereas the fecal DNA panel detected 40.8%.
Additional work is needed to determine the utility of this technique as a screening tool and to identify the most sensitive and specific molecular markers. Multicenter trials are necessary to establish the role of fecal DNA testing in the diagnostic armamentarium.
HEREDITARY COLORECTAL CANCER
Hereditary colorectal cancers represent approximately 20% of all large-bowel cancers.32 Unlike sporadic cases of colorectal cancer, most individuals with hereditary colorectal cancer have a family history of colorectal cancer, tend to develop disease before the age of 50, and have more proximally located lesions. Both familial adenomatous polyposis coli syndrome (FAP) and hereditary nonpolyposis colorectal cancer (HNPCC) are inherited in an autosomal dominant manner.
Familial Adenomatous Polyposis Coli Syndrome
Germline defects in the APC gene on chromosome 5 are responsible for FAP. FAP manifests at an early age as numerous colorectal adenomas that undergo malignant degeneration by the time the patient reaches 50 years of age. Genetic testing should be offered to those patients with an increased likelihood of harboring the genetic mutation.
Patients with a known APC gene mutation or one or more first-degree relatives with either confirmed FAP or an APC gene mutation should be screened as early as 10 to 12 years of age. Thereafter, these high-risk patients should be monitored by routine colonoscopy and polypectomy. Prophylactic subtotal colectomy with surveillance of the rectal stump or total colectomy with ileoanal pouch reconstruction is indicated when the number of polyps precludes safe colonoscopy and polypectomy, usually in the late teens or early 20s. Additional indications for prophylactic subtotal or total colectomy include bleeding, diarrhea refractory to treatment, or failure to thrive. In addition, FAP patients should be followed up with upper endoscopy because they have an increased risk of developing gastric, duodenal, and small-bowel adenomas. It is not clear when to begin upper gastrointestinal tract screening. Some recommend an initial upper endoscopy in the patient’s early 20s or around the time of prophylactic colectomy. Others have suggested that there is no need to perform endoscopy before age 30, because the incidence of duodenal cancers in this age group is extremely low, with most cases being diagnosed after age 40. If no polyps are found on initial upper endoscopy, it may be repeated in 3 years. If polyps are noted, however, the number, size, histology, and severity of dysplasia may be scored on a five-point scale (0 to IV) according to the Spigelman criteria, which can then be used to determine the frequency of upper gastrointestinal endoscopy (usually every 1 to 3 years).33 Endoscopy should be performed with a side-viewing scope to better visualize the ampulla of Vater, a common site for adenoma and cancer formation.
Hereditary Nonpolyposis Colorectal Cancer
The modified Amsterdam criteria
At least three relatives must have an HNPCC-related cancer (colon, rectal, endometrial, gastric, ovarian, ureteral, brain, small-bowel, hepatobiliary, or sebaceous tumors)
One must be a first-degree relative of the other two
At least two successive generations must be affected
At least one of the relatives with HNPCC should have received the diagnosis of colorectal cancer before the age of 50 y
FAP should have been excluded in any relative with colorectal cancer
Tumors should be pathologically verified when possible
Individuals with a clinical history suggestive of HNPCC according to the modified Amsterdam criteria or those with a documented DNA mismatch-repair gene mutation should undergo surveillance with colonoscopy at suggested intervals of 1 year,33 2 years,35 or 3 years.36 At the John Wayne Cancer Institute, we favor colonoscopic surveillance annually beginning at the age of 20. Extracolonic surveillance, particularly of the endometrium and ovaries, may be performed annually beginning at the age of 30 with the aid of transvaginal ultrasonography in combination with endometrial biopsies and serum CA-125.33 Such a regimen, however, has not been proven to reduce mortality from gynecological malignancies.
Peutz-Jeghers Syndrome and Juvenile Polyposis
Other hereditary syndromes that increase the risk for colorectal cancer include Peutz-Jeghers syndrome (PJS) and juvenile polyposis. Both are extremely rare, accounting for only .01% of all colorectal cancers. The lifetime risk for colorectal cancer is approximately 10% to 20% with PJS and as high as 40% with juvenile polyposis.35 Both syndromes also increase the risk for upper gastrointestinal tract malignancies.
PJS patients should undergo colonoscopy every 3 years beginning at age 18 and upper endoscopy every 3 years beginning at age 25.35 Because patients with juvenile polyposis are at higher risk for malignancy, surveillance every 2 years beginning at age 15 to 18 (or earlier if symptomatic) for colonoscopy and age 25 for upper endoscopy is recommended.35 Prophylactic subtotal or total colectomy with ileoanal anastomosis has not been shown to decrease mortality due to colorectal cancer. Therefore, any consideration of prophylactic colectomy should be based on the assessment of each individual patient’s risk.
PERSONAL HISTORY OF COLORECTAL CANCER
An efficient and reliable surveillance program is perhaps of greatest importance in patients previously diagnosed with and treated for colorectal cancer, because this population has a substantial risk of developing recurrent disease. Approximately 30% of stage I/II and 60% of stage III colorectal cancer patients will experience a recurrence of their disease, usually within 2 years of surgical resection.37 Additionally, these patients have an 8% incidence of metachronous lesions.
Several authors have noted a correlation between the adequacy of the primary operation and the rate of disease-free and overall survival. The most commonly used determinant of surgical adequacy has been the number of lymph nodes retrieved at the time of resection. Law et al.38 reported 5-year overall survival rates of 62% vs. 86% when the number of recovered lymph nodes was ≤6 vs. >6. Similarly, Tepper et al.39 noted a 5-year overall survival rate of 68% vs. 82% for examination of ≤4 vs. ≥14 nodes. Both the American Joint Committee on Cancer and the International Union Against Cancer recommend the examination of at least 12 lymph nodes per specimen.40,41 These data support the adoption of more aggressive surveillance measures in patients deemed at high risk as a result of inadequate surgical resection. Prandi et al.42 have suggested that examination of fewer than seven nodes is likely to understage Dukes’ B colorectal cancer; these patients should be considered for adjuvant chemotherapy.
The goals of surveillance are to identify and characterize asymptomatic resectable local recurrence, liver and lung metastases, metachronous lesions, and symptomatic disease requiring palliation. There is no debate about the need for surveillance, but the extent and frequency of monitoring are controversial. Unfortunately, studies randomizing patients to receive intensive follow-up versus simple screening tests have produced conflicting results, with several demonstrating no survival advantage.43 As a result, there is no consensus regarding the optimal mode and frequency of colorectal cancer surveillance.
In an effort to establish definitive recommendations for colorectal cancer surveillance, in 1998 the Gruppo Italiano di Lavoro per la Diagnosi Anticipata initiated a prospective, multicenter randomized trial.44 The trial compares low-intensity (office visit, carcinoembryonic antigen, colonoscopy, liver ultrasonography, and abdominopelvic computed tomography, as indicated) with high-intensity (low-intensity studies plus complete blood count, CA 19-9, and chest radiograph) follow-up in colorectal cancer patients who have undergone resection with curative intent. The primary end points are overall survival and tumor-specific mortality. Thus far the trial has randomized 985 patients; target accrual is 1500. Results should provide substantial data to support definitive recommendations for an effective surveillance regimen.
The National Comprehensive Cancer Network provides annually updated recommendations for colorectal cancer screening and surveillance based on expert opinions from multiple oncological specialties. The current postresection recommendations are as follows: (1) a history and physical examination every 3 months for 2 years and then every 6 months for 5 years after surgery and (2) carcinoembryonic antigen testing every 3 months for 2 years and then every 6 months through the fifth year (T2 and greater lesions).45 The first postoperative colonoscopy should be performed 1 year after surgical resection and every 3 years thereafter, provided that there are no significant findings. Routine computed tomographic scans are not currently recommended but may be considered for patients at high risk for recurrence.
Future approaches to colorectal cancer screening will likely emphasize a molecular approach. Molecular markers of proliferation and apoptosis, the detection of DNA mismatch-repair genes, tumor-suppressor gene hypermethylation, and microsatellite instability hold promise as prognostic indicators and predictors of recurrence.46–48 The role of these tools as adjuncts to current surveillance methods remains unknown.
Because of its prevalence in Western countries, its prolonged period of oncogenesis, and its high rate of response to treatment, colorectal cancer is an ideal target of screening and surveillance efforts. Screening and surveillance protocols should consider a patient’s personal and family medical history as an indicator of overall risk. Although molecular approaches to colorectal cancer diagnosis are promising, the sensitivity, specificity, and efficiency of these techniques must be confirmed in a prospective randomized fashion before they become routine for colorectal cancer screening and surveillance.
Supported in part by funding from the Harold J. McAlister Charitable Foundation (Los Angeles, CA).