Public Release: 

Model predicting colorectal cancer screening suggests higher mortality for virtual colonoscopy

Greater risk of cancer deaths and procedure-related deaths

American College of Gastroenterology

ORLANDO (November 1, 2004) -- Analyses of virtual colonoscopy found both procedure-related mortality and cancer deaths were higher compared to complete (i.e., traditional) colonoscopy, according to findings presented today at the 69th Annual Scientific Meeting of the American College of Gastroenterology (ACG). The analyses, conducted by researchers from Oregon Health and Science University and the Portland VA Medical Center, found that virtual colonoscopy, a test for colorectal cancer that relies on a CT scan of the abdomen and pelvis, carries significant radiation exposure and is less likely to detect cancer or pre-cancerous growths in the colon.

Douglas Faigel, M.D. and colleagues Amnon Sonnenberg, M.D. and David A. Lieberman, M.D. looked at hypothetical cohorts of 100,000 patients undergoing either a single colonoscopy or a virtual colonoscopy at age 60. A key assumption of their decision analysis model is that one-third of patients who undergo virtual colonoscopy will need to have a follow-up colonoscopy.

The Oregon researchers examined complications and deaths per 100,000 patients. "With virtual colonoscopy, we don't avoid the risks of complications from colonoscopy such as bleeding or perforation," explains Dr. Faigel. Rather, the rate of complications for virtual colonoscopy is higher than expected because most of those follow-up colonoscopies are likely to involve the removal of polyps, which increases the likelihood of complications.

Radiation Exposure a Significant Risk

Dr. Faigel warns that virtual colonoscopy exposes normal, healthy people to significant radiation, "A CT scan of the abdomen and pelvis in virtual colonoscopy is a lot of radiation."

For this analysis he and his colleagues assumed a single CT scan of the abdomen and pelvis at age 60 carries the risk of cancer death of 1 in 4,000 and based this assumption on studies looking at cancer deaths among those exposed to radiation from atomic bombs in the 1940s. Using this assumption, the Oregon model predicts 25 radiation related deaths per 100,000 associated with virtual colonoscopy. Overall, the model attributes nine procedure-related deaths per 100,000 to traditional colonoscopy, compared to 30 deaths per 100,000 for virtual colonoscopy, including the 25 deaths related to radiation and five procedure-related deaths.

Deaths Attributed to Missed Cancers

The model determined there were 286 colorectal cancer deaths per 100,000 with virtual colonoscopy, compared to 220 per 100,000 with conventional colonoscopy. Dr. Faigel attributes this higher rate of mortality for virtual colonoscopy to declining sensitivity when the two tests are performed. In the model, the researchers assumed a rate of sensitivity for both virtual colonoscopy and traditional colonoscopy of 90 percent. Therefore, in an exam 10 percent of significant lesions (large polyps or cancers) will be missed. "These assumptions are a bit high for virtual colonoscopy and actually understate the sensitivity for traditional colonoscopy which usually would exceed 90 percent," explains Dr. Faigel. "In our model, if a patient had a virtual colonoscopy first and then a conventional colonoscopy, the sensitivity declines to about 81 percent. That's where we found greater colorectal cancer deaths for virtual colonoscopy than with conventional colonoscopy alone."

Information About Colonoscopy and Virtual Colonoscopy from ACG

The four largest studies of virtual colonoscopy have been published in the last year, and three of the four studies showed very poor results for virtual colonoscopy compared to colonoscopy. Only a study by Pickhardt et al. showed results for virtual colonoscopy comparable to complete colonoscopy, and the reasons for the discrepancies between studies remain uncertain. No guideline group has yet endorsed virtual colonoscopy as appropriate for colorectal cancer screening. Colonoscopy remains the most sensitive test available for detection of colorectal and cancer.

The questions raised by recent studies underscore the importance of colonoscopy being performed by skilled practitioners. According to the American College of Gastroenterology, patients undergoing colonoscopy should identify doctors who have substantial experience having performed large numbers of endoscopies, usually evidenced by their having received training in gastroenterology fellowship or surgical residency programs.


The American College of Gastroenterology has both published its own guideline recommendations on colorectal cancer screening which identified colonoscopy as the preferred colorectal cancer screening strategy (Rex DK, Johnson DA, Lieberman DA et al. AmJGastroenterol 2000;95:868-77) and the ACG endorses the recommendations of the U.S. Multi-Society Task Force on Colorectal Cancer with regard to quality in the technical performance of colonoscopy. (Rex DK, Bond JH, Winawer S, et al. Am J Gastroenterol. 2002;97:1296-308.) The message for patients remains unchanged. Everyone over 50 should have a colonoscopy every ten years by an experienced and qualified endoscopist.

About ACG
The ACG was formed in 1932 to advance the scientific study and medical treatment of disorders of the gastrointestinal (GI) tract. With more than 8,000 members, the College promotes the highest standards in medical education and is guided by its commitment to meeting the needs of clinical gastroenterology practitioners. Consumers can get more information on GI diseases through the following ACG-sponsored programs:

  • 1-800-978-7666 (free brochures on common GI disorders, including ulcer, colon cancer, gallstones, and liver disease)
  • 1-866-IBS-RELIEF and (free educational materials)
  • 1-800-HRT-BURN (free brochure and video on heartburn and GERD)
  • (ACG's Web site)

  • Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.