In Review - News from the 2004 Meeting of the RSNA

Researchers joust over CT colonography's value

Perforation rates and new sensitivity data that don't match earlier results cause lingering debate

By: Merlina Trevino

Low sensitivity values, higher than expected perforation rates, and a continuing debate over 2D or 3D reading technique had many CT colonography proponents in McCormick Place gnashing their teeth.

Ever since the landmark Pickhardt trial indicated that CT colonography had sensitivity numbers as good as those of optical colonoscopy, researchers have been trying to replicate and explain the good results (NEJM 2003;349[23]:2191-2200). But a wave of studies showing mediocre sensitivity values prompted session moderator Dr. C. Daniel Johnson to call it a jousting match.

"We don't know if CT colonography is going to be our Sir Galahad and win the battle for us," said Johnson, a professor of radiology at the Mayo Clinic in Rochester, MN.

The Duke University multicenter prospective comparison of optical colonoscopy, CT colonography, and air contrast barium enema delivered the first blow to the "virtual" side. In 614 patients, the sensitivity for 1-cm polyps was 48% for air contrast barium enema, 59% for CT colonography, and 98% for optical colonoscopy. For polyps greater than 6 mm, sensitivity was 41%, 55%, and 99%, respectively. Despite conventional colonoscopy's good results, this was the first study in which CT colonography surpassed air contrast barium enema for polyps greater than 6 mm, said Dr. Erik K. Paulson, director of abdominal imaging at Duke.

A panel of experts suggested that CT colonography should be compared to barium enema and not to optical colonoscopy.

"CT colonography is just not there yet," Johnson said.

Researchers also compared the perforation rates of the colon screening methods. A major deterrent to widespread implementation of CT colonography has been the fear of perforation, said Dr. Jacob Sosna, a physician at Hadassah Hebrew University Hospital in Jerusalem. Sosna and colleagues retrospectively reviewed three medical centers' data on perforation, as well as subsequent management and outcomes, between January 2001 and April 2004. Of 9120 CT colonography exams, six resulted in perforations, yielding a 0.066% perforation rate. Four patients required surgical intervention, and no exams resulted in deaths. Older males with underlying colonic pathologies were at a higher risk for perforations, Sosna said.

Sixteen published studies have placed optical colonoscopy perforation rates at 0.03% to 0.61% for diagnostic procedures and 0.07% to 0.72% for therapeutic procedures. Perforation rates for barium enema are closer to one in 10,000, according to Johnson. Barium contrast may not require as much air insufflation as CT colonography, and physicians can actually see what they're doing, said Dr. Martina Morrin, an assistant professor of radiology at Harvard University. Problems after CT colonography may not be apparent until the patient reports pain. And the perforation rate should be placed in the perspective of the 59,000 deaths attributed to colorectal cancer every year, she said.

Dr. Carl Jaffe, chief of the National Cancer Institute's Diagnostic Imaging Branch, was not surprised at these findings. Cases in which patients have diverticulitis and air leakage through the diverticulum are often counted as perforations, even though some do not require surgery, he said.

"Perforation with optical colonoscopy, especially during polypectomies, is more likely to lead to bleeding. The Sosna study is an interesting case report, but it's not something that sets off alarm bells," he said.

Radiologists expecting guidance on CT colonography interpretation probably left McCormick Place disappointed; a series of four presentations provided no consensus on which method would increase sensitivity for polyp detection. Two studies did not support Pickhardt's advocacy of 3D primary reads with 2D problem solving. Both indicated no significant difference in sensitivity between the two techniques, and 3D primary reads took significantly longer to perform.

New York University Medical Center researchers, led by Dr. Michael Macari, an assistant professor of radiology, recorded an average time of 16.4 minutes for 3D primary reads compared with 10.9 minutes for 2D. Differences in overall sensitivity were not statistically significantly for polyps larger than 6 mm. Dr. Riccardo Iannaccone and colleagues from the University of Rome, Italy, reported similar results using the same software as Pickhardt: 18.8 minutes to read 3D exams versus 11.2 minutes for 2D exams, with no statistically significant difference in sensitivity (76.6% for 3D and 73.3% for 2D).

Dr. Rogier E. van Gelder and colleagues from the University of Amsterdam, however, arrived at a different conclusion. For polyps greater than 10 mm, 3D primary reads had an average 83% sensitivity compared with 72% for 2D. Three-D led to only one perception error (defined as a polyp or patient identified by at least one but not all reviewers), while 2D led to six errors.

Dr. Andrew Lee and colleagues at the University of Wisconsin, Madison theorized that the longer time and distance for 3D reads in their study indicated a greater opportunity for polyp detection. Not everyone agreed, however.

Answers to the many questions surrounding virtual colonoscopy may come soon from the recently approved National CT Colonography Trial (ACRIN protocol 6664). Study researchers expect to accrue 2289 participants from 15 medical centers for this three-year study.