Can Computed Tomographic Colonography Become a “Good” Screening Test?

In the not-too-distant future, I'll have to decide about colorectal cancer screening—as a patient. Despite my first-hand witness to the benefits of colonoscopy, I am seriously considering computed tomographic colonography (CTC) (“virtual colonoscopy”). A recent study of CTC without bowel preparation reported a sensitivity of 90% and specificity of 92% for polyps 8 mm or larger (1). This test appeals to me because I can avoid the preparation, conscious sedation, discomfort, and risk for perforation. These inconveniences collectively cross my threshold for what a “good” screening test should be. A recent multicenter study of CTC reporting a sensitivity of 59% for polyps 9 mm or larger in diameter may cause me to reconsider, however (2).

When I apply the characteristics of a “good” screening test to the currently recommended tests for colorectal cancer screening, all fall short on more than one desirable feature: high sensitivity, high specificity, near-zero risk, patient acceptance (ease, convenience), and cost-effectiveness. Even colonoscopy, the “gold standard” for accuracy, is inconvenient, risky, costly, and, according to many patients, embarrassing. We really need better tests, as evidenced by screening rates below 50% (3).

Computed tomographic colonography may be getting closer to becoming a “good” screening test. The technology has evolved rapidly from single-slice scanners, prolonged breath holding, and unsophisticated software capable of displaying only 2-dimensional (2-D) images to multislice scanners (which allow for faster imaging and thinner sections) and software that provides 3-dimensional (3-D) “flythrough” endoluminal (the so-called virtual) views that simulate colonoscopy. However, several issues remain unsettled. Among issues directly related to CTC, the most important is the variation in sensitivity among studies. How and what can we learn from the published literature about why sensitivity varies …