Looking for colorectal cancers in adults with symptomsBMJ 2013; 346 doi: http://dx.doi.org/10.1136/bmj.f1124 (Published 21 February 2013) Cite this as: BMJ 2013;346:f1124
Twin trials from the UK confirm that computed tomography (CT) of the colon is a sensitive test for colorectal cancer in adults with suspicious symptoms. Computed tomography found more cancers and large polyps than a barium enema in the first trial (7.3% (93/1277) v 5.6% (141/2527); relative risk 1.31, 95% CI 1.01 to 1.68) and diagnosed as many as a colonoscopy in the second trial (10.7% (57/533) v 11.4% (119/1047); 0.94, 0.7 to 1.27).
Barium enema is clearly the least accurate of the three options, says a linked comment (doi:10.1016/S0140-6736(13)60056-2). The choice now is between computed tomography of the colon and colonoscopy for adults with symptoms, such as change of bowel habit or rectal bleeding. Both tests are good at finding cancers and large polyps, so they are unlikely to miss serious colonic disease. In this head to head trial, colonoscopy missed none of the 55 colorectal cancers in the group tested, and computed tomography missed one of 28. Both groups were tracked for three years after their initial investigation.
Patients might prefer computed tomography, but the less invasive option has clear drawbacks, says the comment. Those with positive findings need further investigation—usually a colonoscopy and biopsy—to confirm or rule out a cancer or a potentially premalignant polyp. Almost a third of patients assigned computed tomography needed further tests in this trial (30%, 160/533), a higher proportion than expected and significantly higher than the proportion of patients needing further investigation after colonoscopy (8.2%, 86/1047).
These extra investigations add to the cost of detecting colorectal cancers and large polyps with CT colonography, says the comment. Investigating incidental findings outside the colon will cost even more. Ultimately, the cost effectiveness of computed tomographic colonography will depend on the nature of the initial findings that trigger further investigation. Those details have yet to be worked out.
Cite this as: BMJ 2013;346:f1124