CT colonography: unraveling the twists and turns

The performance of CT colonography has varied widely among published studies to date. This paper reviews differences in indications, populations, technique, and radiologist visualization, to elucidate the key factors that may give rise to disparate results.

The role of CT colonography in screening asymptomatic patients is controversial. Studies employing subjects with known neoplasms generally report higher accuracy, while studies employing surveillance subjects report lower accuracy. Technical factors that appear to be associated with higher accuracy include meticulous bowel preparation and inflation, multidetector CT, combined two- and three-dimensional visualization, and radiologist experience and proclivity. Interobserver variability and practice guidelines remain significant issues for this developing technique. Interobserver variability may be reduced in the future by computer-aided detection algorithms or the routine use of skilled second readers. CT colonography is performed routinely for some indications, as it has demonstrated superior performance to noncolonoscopic screening alternatives.

The disparity in results of reported large-scale CT colonography studies in asymptomatic subjects may be explained in part by differences in patient population characteristics, CT technique, and interobserver variability. CT colonography exceeds the performance of nonendoscopic approaches to colorectal cancer screening. Continuing innovation will seek to improve endoluminal visualization, reduce interobserver variability, and improve patient acceptance.