Meta-analysis: computed tomographic colonography

Abstract

Background: Computed tomographic (CT) colonography, also called virtual colonoscopy, is an evolving technology under evaluation as a new method of screening for colorectal cancer. However, its performance as a test has varied widely across studies, and the reasons for these discrepancies are poorly defined.

Purpose: To systematically review the test performance of CT colonography compared to colonoscopy or surgery and to assess variables that may affect test performance.

Data sources: The PubMed, MEDLINE, and EMBASE databases and the Cochrane Controlled Trials Register were searched for English-language articles published between January 1975 and February 2005.

Study selection: Prospective studies of adults undergoing CT colonography after full bowel preparation, with colonoscopy or surgery as the gold standard, were selected. Studies had to have used state-of-the-art technology, including at least a single-detector CT scanner with supine and prone positioning, insufflation of the colon with air or carbon dioxide, collimation smaller than 5 mm, and both 2-dimensional and 3-dimensional views during scan interpretation. The evaluators of the colonogram had to be unaware of the findings from use of the gold standard test. Data on sensitivity and specificity overall and for detection of polyps less than 6 mm, 6 to 9 mm, and greater than 9 mm in size were abstracted. Sensitivities and specificities weighted by sample size were calculated, and heterogeneity was explored by using stratified analyses and meta-regression.

Data synthesis: 33 studies provided data on 6393 patients. The sensitivity of CT colonography was heterogeneous but improved as polyp size increased (48% [95% CI, 25% to 70%] for detection of polyps <6 mm, 70% [CI, 55% to 84%] for polyps 6 to 9 mm, and 85% [CI, 79% to 91%] for polyps >9 mm). Characteristics of the CT colonography scanner, including width of collimation, type of detector, and mode of imaging, explained some of this heterogeneity. In contrast, specificity was homogenous (92% [CI, 89% to 96%] for detection of polyps <6 mm, 93% [CI, 91% to 95%] for polyps 6 to 9 mm, and 97% [CI, 96% to 97%] for polyps >9 mm).

Limitations: The studies differed widely, and the extractable variables explained only a small amount of the heterogeneity. In addition, only a few studies examined the newest CT colonography technology.

Conclusions: Computed tomographic colonography is highly specific, but the range of reported sensitivities is wide. Patient or scanner characteristics do not fully account for this variability, but collimation, type of scanner, and mode of imaging explain some of the discrepancy. This heterogeneity raises concerns about consistency of performance and about technical variability. These issues must be resolved before CT colonography can be advocated for generalized screening for colorectal cancer.