CT Colonographic Screening: Does the Technique Merit Widespread Adoption?

Abstract & Commentary

By Malcolm Robinson, MD, FACP, FACG, Emeritus Clinical Professor of Medicine, University of Oklahoma College of Medicine, Oklahoma City. Dr. Robinson reports no financial relationship to this field of study.

Synopsis: CT colonographic screening of asymptomatic adults identified 90% of polyps or cancers measuring 10 mm or more.

Source: Johnson CD, et al. Accuracy of CT colonography for detection of large adenomas and cancers. N Engl J Med 2008;359:1207-1217.

Colorectal cancer is the third most common malignancy in the United States, and it's the second leading cause of cancer deaths. Screening is believed to be very important for interrupting the natural history of colon cancer. Anything that improves early detection of adenomas and early carcinomas would be most welcome, and many radiologists have been very enthusiastic about the promises of CT colonography. This technique rapidly examines the entire colon in a minimally invasive way without any need for sedation.

In the present study, 2600 asymptomatic volunteers ≥ 50 years of age in need of screening colon examinations were recruited at 15 U.S. centers. Some radiologists used 2-dimensional software, others a 3-dimensional version. All selected radiologists were considered to be highly qualified in the interpretation of CT colonographic examinations and had been tested for such competence. A total of 2531 patients successfully completed CT colonography followed by colonoscopy (99% on the same day as the CT exam). The primary study endpoint was the detection by CT colonography of histologically determined adenomas and adenocarcinomas ≥ 10 mm in size. Results among radiologists were averaged. Calculations were made for sensitivity, specificity, false-positive rate, and positive predictive value. However, the bottom line result of the study was that CT colonography correctly identified 90% of polyps and cancers measuring 10 mm or more. Using this technology, 65% of 5 mm lesions were identified, and 78% of 6 mm lesions were correctly diagnosed with increasing accuracy as size approached 10 mm. Radiologists varied in their detection rates from 67% to 100%, with 47% of radiologists identifying all of the large (≥ 10 mm) lesions. By contrast, 128 large (≥ 10 mm) adenomas or carcinomas were found on colonoscopy in 109 of the 2531 patients (4%). A total of 547 lesions 5 mm or larger were found including 136 hyperplastic polyps (25%), 7 lipomas (1%), and 30 miscellaneous non-neoplastic lesions. One 10 mm rectal cancer was missed on CT colonography. A total of 30 lesions ≥ 10 mm were reported on CT colonography and not found on the initial colonoscopy. Only 15 of these 27 participants returned for a second colonoscopy. Five of 18 lesions seen on the CT exam could be confirmed (one turning out to be a 35 mm tubulovillous adenoma with dysplasia). For various reasons, colonoscopy was not repeated in the other 12 patients. Polyethylene glycol preparation was utilized in 40% of patients, sodium phosphate in 55%, and magnesium citrate in 4%. Barium sulfate was employed for fecal tagging for the CT, and iodinated contrast material was used for fluid tagging. Glucagon was administered pre-CT in 92% of participants. Adverse events included nausea and vomiting for less than 24 hours in one patient after CT colonography and E. coli bacteremia 24 hours after both procedures had been done. One patient was hospitalized for post-polypectomy bleeding. As in other studies of CT colonography, various extracolonic findings were identified at CT. The follow-up and relevance of these were not described.

This study indicated that 17% of this patient population would have been referred for colonoscopy for lesions of 5 mm or more had this not been an investigation already requiring colonoscopy. No differences were found in the software types utilized (i.e., 2-dimensional vs 3-dimensional). The authors propose that CT colonography might lead to an increased acceptance of colorectal screening by relevant populations.


This study has results similar to previous reports. Although the concept of a rapid radiographic evaluation of the colon seems attractive, the actualities are not encouraging. In the first place, this procedure is quite expensive. Second, the major obstacle to colonoscopy is the pre-procedure preparation, which is the same for CT colonography as for colonoscopy. Third, and most important, colonoscopy not only finds colon lesions, it also allows their removal. As a result, colonoscopy is both diagnostic and potentially curative. To me, the most valuable role of CT colonography would be the evaluation of patients in whom colonoscopy proves incomplete or otherwise unsuccessful. However, if this were the only indication, no radiologist would ever have enough cases to warrant full training to competence in this technology. Many patients do not like CT colonography, which involves insufflations of gas or air into the rectum during the procedure. In previous studies, patients have preferred colonoscopy to CT colonography. Colonoscopy is presumably better accepted because of the utilization of sedation for almost all of these procedures. Considerable scatter was observed in individual sensitivities of these trained radiologists in successful polyp identification. This suggests that wide adoption of this technique in communities would be unwise. I predict that there is a very limited future for CT colonography in its present configuration. Others strongly disagree. Time will tell.