Abstract
Objective: To compare the CT colonography performance of a trained radiographer with the CT colonography performance of a skilled physician.
Subjects and Methods: 2,342 asymptomatic average risk participants were referred for screening CT colonography. With 3D primary reading, two readers performed an endoluminal fly-through from anus to cecum and back again for supine and prone data sets. Transverse 2D and MPR views were used for problem solving. Results: In double-blind reading, the positive rate of CT colonography was 4.4% (103/2,342) and the positive predictive value was 89.3% (92/103) for lesions 6 mm or larger. Misdiagnoses were caused by careless positional shift in four participants and by misevaluating of lesion's CT value in one participant. The overall agreement between two readers was 99.8%.
Conclusions: By trained and experienced readers, CT colonography interpretation in double-read paradigm is not necessary for further improvement of accuracy. In CT colonography interpretation, positional shift of the lesion and air inclusion is important for efficient differentiation between polyp and faecal residues.
