3489 Virtual colography compared to endoscopic colonoscopy in detecting ascending colon and cecal lesions.

Abstract

Introduction: Virtual colography (VC) is being studied as a potential screening tool for colorectal polyps and cancers that is less invasive than endoscopic colonoscopy (EC). VC performance characteristics have been assessed only in small numbers of patients who had known lesions. The ascending colon and cecum are not visualized in up to 10% of EC.We therefore chose to evaluate the performance of VC, in compariosn with EC as a gold standard, in the ascending colon and cecum of all patients undergoing EC. Methods: Patients scheduled for EC for any reason were prospectively recruited for the study. 58 consecutive patients who agreed and gave informed consent were enrolled. EC was performed by a single gastroenterologist (12 yrs postgraduate) who was blinded as to VC results. All underwent VC immediately or within a week after EC, and the exams were done in separate areas of the hospital by different staff members. After standard EC prep, VC was performed with air insufflation to tolerance, and by single breath-hold thin section helical CT. GE Navigator software was used to generate interactive 3D images. Images were coded and in a blinded fashion. Results: 58 patients, mean age 58, range 31-83, and 72% were male. Reasons for EC were FOBT/rectal bleeding 49%, unexplained abdominal pain 11%, and others 40%. EC reached the cecum in 53/58 (91%). 26 polypoid lesions were detected, 12 were < 5mm, 8 were 5- 9 mm, 6 were >10 mm. 2 flat lesions were detected and confirmed on histology to be flat adenomas. Of the 26 polypoid lesions detected by EC, VC detected 11 (42%), 2/12 < 5mm (17%), 4/8 of 5-9 mm (50%), and 5/6 >10 mm (83%). VC detected 4 lesions not seen on EC, 0 < 5 mm, 2 of 5-9 mm, and 2 >10 mm. However, VC detected 0/2 of the flat adenomas seen on EC. Thus, overall sensitivity of VC was 42% (95% CI: 23%-63%) with 85% (95% CI: 66%- 96%) specificity. Conclusion: VC has comparable or higher sensitivity than EC for polypoid lesions >10mm, comparable sensitivity to EC for polypoid lesions of 5-9 mm, but has lower sensitivity for lesions < 5mm and for flat lesions. VC technology may need improvement for mucosal surface irregularity prior to wider acceptance. Nevertheless, the fact that malignant potential of colorectal polyps is clinically important generally only for polyps >10mm suggest that VC may have potential for routine clinical screening, especially in those patients in whom the ascending colon or cecum was not reached by EC.