Dr. Tamar Stella of Hadassah Medical Center in Jerusalem found that only 12 of 28 studies published since 1980 fulfilled the researchers' inclusion criteria, and none of the studies examined enough patients to produce statistically valid results. The largest study, performed in 1995, included 1831 patients but found only 211 polyps.

"So there was no single study that had large enough numbers to come up with an accurate assessment," Stella said.

Results from the 12 studies were pooled and a meta-analysis was performed to produce viable data. Eight of the 12 involved a high-risk population. Seven studies compared the double-contracts barium enema results with conventional colonoscopy, and five compared the results with flexible sigmoidoscopy.

Double-contrast barium enema exams identified 105 of 134 patients in the seven studies that provided cancer data. Overall, 990 polyps were identified, but only 50% were diagnosed with the barium enema technique. The sensitivity of barium enema increased when only polyps of 1 cm were considered. Of 237 polyps in this category, 155 were identified with double-contrast barium enema.

The overall pooled sensitivity of the technique was 79%, and the sensitivity for 10-mm or larger polyps was 65%. Even with pooled data, the researchers could not measure the test's sensitivity for polyps small than 10 mm because only four of the 10 studies included such data.

A meta-analysis of conventional CT, published in the American Journal of Roentgenology in 2003, found that the pooled sensitivity of CT colonoscopy in the evaluation of 1400 polyps was 81%.