Choice of Screening Procedure
Various procedures (fecal occult blood test [FOBT]), sigmoidoscopy, colonoscopy, CT colonography) are available for screening for colorectal cancer. These methods differ in terms of sensitivity, specificity, and cost. Complication rates are low overall. Colonoscopy has the highest complication rate, but does offer a therapeutic option in polypectomy.
Selecting the best screening strategy is very difficult. The chosen method should be simple, safe, and accurate. High sensitivity (i.e., identification of as many patients with adenomas as possible) and high specificity (i.e., as few false-positive diagnoses as possible) are both desirable, since false-positive results are unsettling for the patient and result in expensive follow-up procedures and a potentially increased risk of complications. The most commonly used screening procedures at this time are the FOBT and colonoscopy. Sigmoidoscopy is rarely used in Europe, but is a frequent procedure in the US. Recent studies have shown that CT colonography could also play an important role in screening for colorectal carcinoma (Fig. 6.3).
Fecal Occult Blood Test
Guaiac-based test. Fecal occult blood tests (the most widely used is Hemoccult; Beckman Coulter Inc., Brea, California, USA) use guaiac compounds. Guaiac changes color upon contact with hematin which has entered the bowel lumen as a result of bleeding. The sensitivity of this procedure for colorectal carcinoma is about 60%, but for adenomas it is only around 20%. The low sensitivity rates are due to the fact that the only adenomas or carcinomas that can be detected are those that cause significant amounts of bleeding into the bowel lumen. However, colorectal carcinomas bleed only intermittently, and adenomas often do not cause any bleeding at all. The specificity of the guaiac-based FOBT for detecting occult blood is 85%–90%.
Fecal immunochemical tests. More recent FOBTs are based on immunochemical principles and can offer better results than the standard tests that use guaiac (Levi et al. 2007). The sensitivity of these tests has been reported to be up to 94% for carcinomas and 67% for advanced neoplasias (adenomas ≥1 cm, adenomas with high-grade dysplasia, and carcinomas).
Results. Despite the low sensitivity of guaiac-based stool tests, two meta-analyses of large randomized studies reported a relative reduction in colorectal carcinoma mortality of 14%–16%. In these studies, screening did not lead to a significant decrease in the incidence of carcinoma. This apparent contradiction may be explained by the fact that the guaiac-based test detects only a small number of adenomas, and only a relatively low percentage of carcinomas at an early stage when they are still treatable. Modern stool tests using immunochemical procedures, as well as tests for pyruvate kinase M2, have been shown in preliminary studies to have higher sensitivities than the standard Hemoccult test for carcinomas and adenomas. These newer methods are currently being tested. A disadvantage of the new stool tests, however, is that they are much more costly than a guaiac-based FOBT.
Sigmoidoscopy is a more expensive and more invasive technique than a FOBT, but it also has several advantages over traditional stool tests: in the segments of the colon that it can reach—the rectum and sigmoid colon—it can reliably detect even small adenomas, and these can be removed in the same session (therapeutic option). A disadvantage compared with a FOBT is the necessity of complete bowel cleansing, albeit a reduced form usually only performed with an enema. This limits patient compliance. In addition, proximal colonic lesions cannot be diagnosed by sigmoidoscopy. If a polyp is detected in the distal colon (“index polyp”), a colonoscopy must be performed to rule out the presence of synchronous lesions in the proximal colon. In such a case the patient has to make an additional appointment for a follow-up examination, for which thorough bowel cleansing must be performed.
Results. In terms of detecting advanced colonic neoplasia (adenomas >1 cm, adenomas of any size with high-grade dysplasia, or carcinomas), the sensitivity of sigmoidoscopy compared to that of colonoscopy is relatively low, ranging from 35% to 70%, because of the significant number of right-sided lesions. This is one of the reasons why sigmoidoscopy is not widely used for screening in Europe. In the United States, it is more widespread partly because it can be performed by specially trained paramedical personnel, not just by gastroenterologists.
Gold standard. Colonoscopy is the standard procedure for screening for adenomas and carcinomas in the colon. In Germany, patients aged 55 with no known risk factors are eligible to undergo screening colonoscopy as a preventive measure which is covered by statutory health care. Colonoscopy is currently the gold standard for detection of colonic adenomas and carcinomas, but it is not a perfect gold standard. In a study of “back-to-back colonoscopies,” Rex et al. (1997) showed that the examination missed 6% of adenomas measuring 10 mm or larger, 13% of those measuring 6–9 mm, and 27% of those that were 5 mm in size or smaller. The high efficacy of colonoscopy as a screening tool, despite such results, is due to the fact that polypectomy can reduce the incidence of carcinoma and thus carcinoma-related mortality by 76%–90% (Winawer et al. 1993). This means that there is a high level of evidence for the use of colonoscopy as a screening method. Nevertheless, the potential risks associated with this screening method must be set against its positive effects.
Complications and compliance. In one colonoscopy screening study with more than 3000 patients, 0.3% of patients experienced severe complications (usually hemorrhage), although none of these included perforation. Other studies have reported a perforation risk of 0.03%–0.1% for diagnostic colonoscopy, with a fatality rate of 0.01% (1:10 000). Nevertheless, the general public is largely unwilling to take part in colonoscopy screening programs, mainly because of the unpleasantness associated with bowel preparation and also the examination itself. Compared with optical colonoscopy, CT colonography could be more readily accepted by patients, especially when performed with reduced bowel preparation.
CT colonography has the potential to become the equivalent of conventional colonoscopy for colorectal carcinoma screening. It is virtually the equal of conventional colonoscopy in the detection of advanced neoplasia, while being far less invasive than conventional colonoscopy (Kim et al. 2007b).
Complications and compliance. State-of-the-art CT colonography screening requires a full bowel preparation including dietary restrictions, laxatives, and tagging agents—one factor that, as in conventional colonoscopy, may limit patients’ willingness to take part in CT colonography screening programs. However, CT colonography is less invasive than conventional colonoscopy: At around 1:10 000, the risk of perforation is far lower than in diagnostic colonoscopy (where it is up to 1:1000). Unlike in symptomatic patients, in whom CT colonography is performed with intravenous contrast administration (with a corresponding risk of contrast-agent-induced nephropathy and anaphylactic reactions), CT colonography screening studies are performed without the use of intravenous contrast agents, minimizing the related risks and costs (Fig. 6.4).
Target Lesion: Advanced Adenoma
Advanced neoplasia of the colon include adenocarcinomas and advanced adenomas, which are premalignant neoplasms. Advanced adenomas are the ideal target lesion for colorectal cancer prevention: Although these lesions are still benign, the risk of malignant transformation is high. An advanced adenoma is defined by at least one of the following three criteria: lesion size 10 mm or greater, presence of a substantial villous component (>25%), or the presence of high-grade dysplasia. The removal of an advanced adenoma by colonoscopy disrupts the adenoma–carcinoma sequence—the transformation of the benign neoplasm into colorectal cancer.
The prevalence of advanced adenomas in a screening population is very low, around 3%–4%. The classification of an adenoma as “advanced” is usually determined by its size: 90%–96% of all advanced adenomas are larger than 10 mm. Only about 4% of adenomas that are 6–9mm in size are advanced adenomas. CT colonography does not permit histological analysis of detected polyps, so lesions can only be classified as suggestive of malignancy on the basis of their size on CT. Based on this distribution frequency, to be effective for cancer prevention CT colonography needs to have high detection rates for adenomas measuring 10 mm or more, which has been confirmed in the recent literature.
Results of Studies on CT Colonography as a Screening Procedure
Because CT colonography is a minimally invasive procedure, it has the potential for greater patient compliance than conventional colonoscopy. In a prospective study by Pickhardt et al. published in 2003, the reported sensitivity of CT colonography for polyps measuring 10 mm or larger in asymptomatic adults was 93.8%. In later studies, Cotton et al. (2004) and Rockey et al. (2005) reported sensitivity rates between 34% and 53%. The large difference between sensitivity rates is primarily due to methodological issues, and can partly be explained by differences in examination techniques, data analysis, and varying experience levels of individual readers. In the studies by Rockey et al. and Cotton et al., either thicker slices were used in the CT protocol or fecal tagging was not included in the bowel preparation. More recent results also suggest that a lack of reader expertise may have contributed to the suboptimal results: In a re-evaluation of the study published by Rockey et al., for example, a working group led by Doshi et al. retrospectively identified up to 20% of the polyps missed during the original study.
Recent literature results. On the basis of the results reported above, there were insufficient data to support the routine use of CT colonography for screening of an asymptomatic population. Both the American Cancer Society and the American Gastrointestinal Association recommended that CT colonography not yet be used for screening for colon cancer because reliable data on its performance in screening populations were still lacking. For this reason, large randomized multi- and single-center studies have been undertaken in Europe and in the United States, and early results have already been published. Unlike the early studies, the results of the newer generation of studies have shown high sensitivity rates in asymptomatic patient populations. The results of the multicenter ACRIN study 6664 (Johnson et al. 2007), in which 15 study centers in the United States participated, showed that in 2600 asymptomatic screening patients the sensitivity of CT colonography was 90% for the detection of patients with adenomas measuring 9 mm or more, with a specificity of 86%.
The results of the Munich Colon Cancer Prevention Trial (Graser et al. 2009), a prospective single-center study evaluating CT colonography in a screening population, showed the following sensitivity rates in 307 patients: 94% for lesions measuring 10 mm or more, 92% for lesions measuring 6–9 mm, and 72% for lesions measuring 5 mm or less. The sensitivity of CT colonography for advanced colonic neoplasia (advanced adenomas and invasive carcinomas) was 93%.
In a study published in 2007 in the New England Journal of Medicine, Kim et al. reported similar detection rates for CT colonography and optical colonoscopy for advanced adenomas in two separate, but demographically similar, groups of patients. In this study, 3120 patients underwent initial CT colonography and 3163 patients conventional colonoscopy with polypectomy. If the results of CT colonography were positive, a follow-up colonoscopy was performed. Advanced colonic neoplasia was detected in 3.2% in the CT colonography group and in 3.4% in the optical colonoscopy group. The difference in results between the two groups was not statistically significant. The rate of referral for optical colonoscopy among patients who first underwent CT colonography was 7.9%. The total number of polypectomies in the CT colonography group was 561 (patients with polyps measuring 6–9 mm will undergo interval follow-up) and in the optical colonoscopy group it was 2434. In the first group of patients, no colon perforation occurred, while in the second group there were seven perforations.
In a study by Stoop et al. published in Lancet Oncology in 2012, CT colonography showed a significant 55% improvement in screening participation for CT colonography over colonoscopy; both methods showed a similar diagnostic yield for the detection of advanced neoplasia.
Conclusions. These results demonstrate that with an adequate examination technique and a sufficiently experienced radiologist, CT colonography can attain high sensitivities for clinically significant polyps, even in asymptomatic patients. The detection rates for advanced adenomas achieved with primary CT colonography do not differ significantly from those achieved with primary optical colonoscopy. However, colonoscopy is still more accurate in the detection of smaller polyps (≤5 mm), although the clinical relevance of these lesions is questionable since less than 0.01% are malignant. In addition, the rate of complications associated with CT colonography would be expected to be lower than with optical colonoscopy.
These results support the use of CT colonography as a primary screening method before referral for therapeutic optic colonoscopy. CT colonography is an accurate and perhaps more cost-efficient “filter” than optical colonoscopy. Given the overall minimal invasiveness of the procedure, compliance with screening of the general population could potentially be increased and the risk of associated complications reduced if CT colonography were to be introduced into mass screening programs.