Study Ranks CT Colonography Less Cost Effective Than Stool Test, Colonoscopy

CT colonography is better than not screening,
but is the most costly, least effective of the
screening methods available (Photo used
by permission of the RSNA)

Screening people for colorectal cancer using CT colonography is more cost-effective than not screening people at all, but not as effective as every other screening method currently available, a new cost-effectiveness study shows.

Led by Dr. David Vanness, assistant professor at the University of Wisconsin School of Medicine and Public Health, the researchers compared the cost-effectiveness of CT colonography for colorectal cancer screening in average-risk asymptomatic subjects in the United States aged 50 years. (Link to published site)

Vanness told the Hub in a phone interview that the study used data from the ACRIN 6664 protocol of the National CT Colonography Trial (NCTCT) the largest study to date to look at CT Colonography performance. The study was published this week online in the journal Radiology.

“We re-analyzed the data to be sure that it is applicable to patient decision-making,” Vanness said. “The ACRIN 6664 was originally designed to determine clinical accuracy of detection of the various screening methods, how accurate the methods are in detecting individual lesions. What really matters at the patient level, however, is whether a virtual colonoscopy results in referring patients with precancerous adenomas down the right path to get them addressed.”

The other first for the study is that they used the three statistical models that have been validated for comparing effectiveness of screening methods for a variety of cancers. Vanness said there are very few long-term studies of screening because it takes 10 to 20 years to complete such studies. As a consequence, he says it is unknown what the impact is of finding those adenomas, especially the small ones.

“It is unknown whether doctors are removing lesions that are dangerous or whether they’re removing lesions that never would have become cancerous,” he said.

In an effort to develop the means to efficiently do these types of cost-effectiveness studies, the National Cancer Institute sponsored the Cancer Intervention and Surveillance Modeling Network (CISNET). The CISNET researchers have developed three validated statistical models for simulating the natural course of various cancers. Each model simulates large cohorts representative of 50-year-old subjects in the United States in 2007.

“Each model has different modeling approaches and we wanted to see if using different models produced different results or whether they were basically in agreement,” Vanness said. “We largely found that they are (in agreement), even though they show different absolute values for each of the screening methods, they were very comparable in terms of relative ranking of the screening methods. So it validates the models.”

The NCTCT study enrolled 2,600 asymptomatic people aged 50 years or older. The individuals were scheduled to undergo computed tomographic (CT) colonography followed by colonoscopy, which for trial purposes served as the reference standard.
In the current study, Vanness’ team ran each simulation using assumptions of 100 percent adherence to screening guidelines and 50 percent adherence to guidelines.

When they assumed 100 percent of the population would adhere to screening guidelines all models found that colonoscopy, the fecal immunochemical test plus flexible sigmoidoscopy, and fecal occult blood test plus flexible sigmoidoscopy increased life expectancy and reduced lifetime colorectal cancer risk more than CT colonographic screening at 5- or 10-year intervals.

When they assumed 50 percent of the population would adhere to screening guidelines, performing CT colonography every 5 years reduced colorectal cancer risk more than colonoscopy in two of the three models. In the MISCAN model CTC reduced the risk by 39.1 percent compared to 38.7 percent for colonoscopy. In the SimCRC model the risk reduction was 61.6 percent compared to 59.5 percent.

When they looked at cost, they found that the predicted cost and effectiveness were similar across the models. With 100 percent adherence, all models predicted total costs to be lowest with no screening, ranging from $1,982 to $2,459. The highest cost was for CT colonography every 5 years ranging from $2,900 to $3,854. With 50 percent screening adherence, all simulations found CT colonography every 5 years to be the most costly.

Two models found that screening with conventional methods reduced total costs relative to no screening. The CRC-SPIN model predicted a cost of $1,978 for FOBT plus flexible sigmoidoscopy compared to $1,982 for no screening. The SimCRC model predicted an average cost of $2,188 for FOBT plus flexible sigmoidoscopy, $2,239 for FIT plus flexible sigmoidoscopy, and $2,447 for colonoscopy compared to $2,459 for no screening.

The researchers concluded that all three CISNET microsimulation models predict CT colonography to be more costly and less effective than non-CT colonographic screening but that CTC was better than no screening. Although absolute costs and outcomes differed slightly, the models were largely consistent in the relative costs, outcomes, and ranking by net benefits among the strategies.

“Our conclusion is a little more guarded,” Vanness said. “There are more effective uses of resources (than screening with CTC), but we also know that there is a segment of the population that wants a noninvasive test that is better than fecal occult test. It may not be the first, best option for everyone, but it is better than not doing anything.”

By Michael O’Leary, contributing writer, Health Imaging Hub