Computed Tomographic Colonography for Patients at High Risk of Colorectal Cancer: Title and subTitle BreakTrading Accuracy for Access and Compliance

With nearly 150 000 cases detected annually, colorectal cancer is the second most common cause of cancer-related death in the United States.¹ Over the past 2 decades, however, the annual death rate from colorectal cancer has been declining. This improvement has partially been attributed to an increase in colorectal cancer screening and surveillance nationally. Removal of polyps that have the potential for future malignant transformation and the detection of colorectal cancer at an earlier, more curable stage have contributed to the reduction of colorectal cancer deaths.² While colonoscopy remains the gold standard for detection of colorectal pathology, alternative modalities used for screening, including flexible sigmoidoscopy, double-contrast barium enema, and computed tomographic (CT) colonography, are acceptable choices for average-risk patients.³

For individuals at increased risk for the development of colorectal cancer—those with a family history of colorectal cancer in a first-degree relative or personal history of advanced adenomas—colonoscopy is currently the recommended surveillance modality. Current US Multi-Society Task Force guidelines recommend follow-up colonoscopy in 3 years after removal of an advanced adenoma.³ Whether alternative, less-invasive surveillance techniques such as CT colonography are accurate for detecting colorectal polyps and cancer in this high-risk population is unclear.

In this issue of JAMA, Regge et al⁴ assess the accuracy of CT colonography in a cohort of patients at increased risk of colorectal cancer. This multicenter study evaluated 937 patients with 1 of 3 risk factors for colorectal cancer—a family history of advanced neoplasia in a first-degree relative, a personal history of colorectal adenoma, or a positive result from a fecal occult blood test (FOBT). To measure the sensitivity and specificity of CT colonography in detecting advanced adenomas or colorectal cancer, the study recruited participants to undergo CT colonography interpreted by experienced radiologists followed by colonoscopy on the same day. Overall, compared with unblinded colonoscopy as the gold standard, CT colonography had 85% sensitivity and 88% specificity for detecting advanced neoplasia. Based on these findings, the authors conclude that CT colonography may be an acceptable alternative to colonoscopy for surveillance in many individuals at elevated risk for advanced colorectal neoplasia.

This carefully designed and executed validation study of CT colonography reports higher sensitivity than earlier studies examining its use in screening. Early studies reported sensitivity as low as 39% for polyps larger than 6 mm,⁵ and a 2005 meta-analysis of 33 studies reported sensitivity of 48%, although sensitivity data from individual studies varied widely.⁶ While the study by Regge et al adds important new information to the ongoing debate about optimal strategies for colorectal cancer screening and surveillance, its results should be interpreted in context and with caution. First, for several reasons, the results of this study may not be generalizable on a population level. The centers participating in the study were carefully selected. All endoscopists performing the colonoscopies were experienced and the facilities were required to have at least 1 radiologist who had extensive training with an expert in CT colonography. Such results may not be achievable in other settings. Second, the negative predictive value of CT colonography for detecting advanced cancers was 96.3% overall and for the group with positive FOBT results was 84.9%, raising the concern of whether false-negative rates of approximately 4% and 15%, respectively, are acceptable for surveillance of advanced neoplasia in a high-risk population.

Although controversy remains about the accuracy of CT colonography for surveillance in populations at high risk for colorectal neoplasms, it is important to understand the additional trade-offs between colonoscopy and CT colonography when evaluating a surveillance modality. Complications after colonoscopy are infrequent but costly and associated with significant morbidity. Although perforation rates are low in the general population (1 in 1000), they are as high as 1 in 500 in the elderly Medicare population.⁷ Perforation after CT colonography has been reported but is rare. Up to half of the adverse events after endoscopy are cardiovascular complications that result from sedation.⁸

This risk is avoided by CT colonography because it ordinarily requires no sedation. On the other hand, individuals undergoing CT colonography are exposed to radiation—albeit low doses. Individuals at high risk for colorectal pathology would likely undergo frequent imaging and, as a result, would accumulate radiation exposure over a lifetime of surveillance. Incidental extracolonic findings are both a risk and a benefit of CT colonography. Single-center studies have found that 6% to 10% of patients have unsuspected extracolonic findings.^{9 - 10} Medical and surgical workups for these findings were estimated to result in an average added cost per CT colonography of about $30, not to mention added patient anxiety.^{9 - 10} In most cases, CT findings were found to be benign, although in 2.5% of patients in 1 study, clinically significant new diagnoses were made.⁹

It is essential for CT colonography to be implemented carefully and in appropriate patient populations. First, there is a learning curve for accurate interpretation of CT colonography.¹¹ To ensure accuracy, CT colonography should be offered at centers that have the appropriate state-of-the-art equipment and are staffed by radiologists with training and experience in interpreting the images. Second, appropriate patient selection is important. Computed tomographic colonography is notoriously poor in detecting flat lesions.¹² For this reason, patients with a history of flat neoplasia or who are at risk for flat neoplasia (eg, those with inflammatory bowel disease) are not good candidates for CT colonography. As Regge et al confirmed, up to 50% of patients with positive FOBT results have underlying colorectal pathology that would necessitate a colonoscopy, making CT colonography a very cost-ineffective strategy. Moreover, the algorithms for referring patients with suspicious CT findings for colonoscopy need to be refined. Whether small polyps (<10 mm) detected on CT colonography need to be removed is a matter of debate. Because the risk of cancer in a small polyp is very low, the cost and risk of colonoscopy may not be warranted. However, leaving a polyp in situ may result in patient anxiety and will likely require follow-up CT colonography with its attendant cost and radiation exposure.

Even if CT colonography were widely adopted, it is unclear whether it would be an effective means of addressing the low rates of recommended screening and surveillance on a population level. Low rates of colonoscopy are not purely a result of inadequate capacity. In 2003, an estimated 14 million colonoscopies were performed in the United States when the estimated capacity nationally was 22 million.¹³ Access issues beyond actual capacity—lack of insurance coverage and limited availability in rural areas—clearly play a role in poor compliance with recommended screening and surveillance. Some evidence suggests that CT colonography may increase colonic evaluation overall. In a single-center analysis, investigators found no change in referrals for colonoscopy after the introduction of screening CT colonography.¹⁴ These findings suggest that the adoption of CT colonography may facilitate increased screening and surveillance rather than simply replace colonoscopy.

While the use of CT colonography as a screening and surveillance modality is still a matter of debate, the study by Regge et al⁴ suggests that CT colonography may be an acceptable alternative to colonoscopy in patients with a history of adenoma and those with a family history of colorectal neoplasm. The question remains whether clinicians are willing to accept a study with decreased sensitivity for the potential of increased adherence with recommended screening and surveillance guidelines. With the majority of individuals in the United States who meet criteria for colorectal cancer screening and surveillance not undergoing recommended procedures, an imperfect test that has a lower risk profile and greater acceptance among patients¹⁵ seems to be an appealing solution.