Clinical significance of virtual colonoscopy (CT colonography) with special reference to polyp morphology

BACKGROUND/AIMS: Computed tomographic colonography [virtual colonoscopy (VC)] is a new imaging method that may contribute to the detection of colorectal neoplasms. The aim of this study was to investigate the clinical significance of VC for the diagnosis of colorectal tumors with special reference to their morphology. METHODOLOGY: Sixteen patients with colorectal tumors were enrolled in this study. Colonic preparation was performed using magnesium sulfate, and scopolamine butylbromide was injected before air insufflation from the anus. VC was performed with an Aquilion multislice CT system (Toshiba) and Workstation M900 maximum (ZIO). The results were compared with those of conventional colonoscopy (CC). In six cases, total colonoscopy was not possible due to tumor stricture. RESULTS: The number of lesions detected by CC was 58; 48 early carcinomas/adenomas (group A) and 10 advanced carcinomas (group B). All group B lesions and 43.8% (21 lesions) of group A lesions were identified by VC. The specificity of VC for group A and B lesions was 45.8% and 100%, respectively. There was no significant difference in detection rate for group A lesions by subclassification according to their morphology: pedunculated; 22% (2/9), semipedunculated; 50% (6/12), sessile; 29% (4/14), superficially elevated; 56% (5/9), and superficially elevated morphology with central depression; 100% (4/4). The mean diameter of the lesions that were detected and not detected by VC was 7+/-8.0 mm (range: 3-30 mm) and 8+/-5.8 mm (range: 2-20 mm), respectively (P=0.90). CONCLUSIONS: Polyps with superficial morphology could be detected by VC as well as those with protuberant morphology. According to evidence that neither the morphology nor the size of a lesion was a significant factor for detection by VC, the authors consider that the diagnostic ability of VC may be improved by better preparation.