Research supports adding computer-aided detection to virtual colonoscopy, but CMS says it may not add enough and denies coverage, at least for now.

A long-simmering debate over the efficiency and affordability of adding computer-aided detection (CAD) to virtual colonoscopy would have seemed settled based on research from a consortium of medical analysts in Italy and the United States. Their study, published in the February issue of Radiology, left no doubt that CAD added to CT colonography (CTC) equaled both lives and money saved.

At the same time, the Centers for Medicare and Medicaid Services (CMS) committee was doing some analysis of its own. The Medicare Evidence Development and Coverage Advisory Committee weighed comments from the Medical Imaging and Technology Alliance, the American College of Radiology, and other organizations and individuals. On May 12, the panel issued its final decision: For purposes of coverage, “the evidence is inadequate to conclude that CT colonography is an appropriate colorectal cancer screening test.”

The future of CTC is clearly in a bit of disarray for the moment, but over the next few years, the debate is expected to intensify and pressure from one source or another will ultimately lead to its widespread adoption, proponents say. And CAD may be the key variable.

“The CMS decision is only temporary,” said Perry J. Pickhardt, MD, associate professor of radiology at the University of Wisconsin Medical School and one of the authors of the multicenter study published in Radiology. “[CTC] is too good and is needed too much to keep down for much longer. Demonstrating the additional yield with CAD will certainly help to convince others and will provide more assurance for uniform performance.”

Colorectal cancer is the second leading cause of cancer-related deaths in the United States, according to the American Cancer Society. This year, more than 106,000 cases will be detected and nearly 50,000 deaths will be attributed to the disease.

Early screening and the subsequent removal of polyps before they can develop into cancers have hastened a steady decline in the number of deaths over the past 20 years. But the arduous bowel preparation and invasive procedure remain a deterrent for many people.

Optical colonoscopy remains the standard practice 40 years after it was introduced. But since 1994, when David Vining, MD, introduced virtual colonoscopy as a less-invasive test, a growing chorus has sung the praises of the new technology. But there is hardly harmony.

Pickhardt was joined in a model-based study of adding CAD to CT colonography by peers at the Institute for Cancer Research and Treatment in Candiolo, Italy; Nuovo Regina Margherita Hospital in Rome; and the University of Rome. Researchers looked at the cost-effectiveness of CTC—with and without computer-aided detection—and compared it with both optical colonoscopy and flexible sigmoidoscopy. A hypothetical population of 100,000 persons aged 50 years undergoing colorectal screening every 10 years was used as the model. The study also estimated the sensitivities for both experienced and inexperienced readers and the incremental accuracy when adding CAD.

“CAD supplements the radiologist interpretation and can increase overall detection of important precancerous lesions,” Pickhardt said. “The benefit is greatest for readers with less experience, but our current work has shown us that even experts can benefit. Using certain assumptions for the additional cost of CAD, we showed that the information gained for the inexperienced reader was clearly cost-effective.”

“Assuming a CAD cost of $50 per CT colonography, the overall program costs increased by only 3% to 5%, largely because of the substantial reduction in [colorectal cancer]-related costs,” the research team wrote in the journal article. “The incremental cost-effectiveness of CT colonography with CAD compared with CT colonography without CAD was $8,661 and $61,354 per life-year gained for inexperienced and experienced readers, respectively.”

Vining said that even in the best of hands, there is the potential to overlook things. “Much like in mammography, most experienced readers don’t need CAD, but I think in the colon, CAD might have a better opportunity to detect small lesions that may not be readily apparent—even to an experienced reader.”

Though he doesn’t support the CMS decision on CTC, Vining, who now works with the University of Texas MD Anderson Cancer Center, said it may be a blessing in disguise for CAD, a notion endorsed by Abraham Dachman, MD, a noted expert in abdominal radiology who teaches at the University of Chicago.

“I don’t agree with the CMS decision. I think it’s a bit short-sighted, and I think it’s probably misguided,” Dachman said. “The damage that’s been done by the CMS decision is really that people misread the decision. CMS didn’t criticize virtual colonography as a bad technique.”

One of the issues raised in the CMS decision was concern over radiation exposure. But the central question was whether outcomes could justify the costs.

“I’m not surprised,” Vining said. “It really comes down to two factors: Money and money. There’s a fear that virtual colonoscopy is going to cost the government a lot of money for screening and dealing with the extracolonic findings. The number two money issue is the turf battle with colonoscopy. Those guys with their scopes lobbied vehemently against virtual colonoscopy.”

Vining said there is a concern that because CTC can only detect polyps, patients with suspected lesions have to undergo the optical colonoscopy for removal, adding to the cost. “Most people won’t need that. Virtual has the ability to identify that 10% who do need it, so you’re actually saving money by not doing the more expensive optical on everybody. Virtual also has the ability to detect an issue outside the colon wall, so you get more for your money.”

The CMS panel appears to have recognized at least the potential by encouraging the publication of data on extracolonic findings from CTC compared to optical colonoscopy. And in that, proponents of virtual colonography find some hope.

That hope may be bolstered by computer-aided detection, Dachman said. “We need to see more data. But the data we have published would suggest that CAD is a good idea for CTC. It will increase sensitivity for everybody and have most of its impact on less-experienced readers with a minimum in reading time and reasonable minimal effect on sensitivity.”

Pickhardt agrees: “CAD is certainly part of the bigger CTC picture. If it can help ensure widespread quality performance, then it may play an important role in coverage decisions.” As CAD improves, he said, it will have a positive impact on both clinical efficacy and cost-effectiveness of CTC screening.

Over the last decade, Dachman, who wrote the second clinical research paper on CTC, has taught hundreds of people to read virtual colonographies through the University of Chicago as well as the American College of Radiology. He said any time a new technology comes along, there is a lag in widespread acceptance. “It’s a catch-22 because people don’t want to learn it until it’s reimbursed, and to get reimbursed, you have to have people who have mastered the technique.”

Many of the major training hospitals, including the Mayo Clinic, New York University, University of California at San Francisco, and the University of Chicago, have training in CTC, and several vendors have offered visualizing programs. As the level of expertise has grown, so have reimbursements—at least in the private sector.

In 2008, there were 43 states that had partial coverage for CTC, Dachman said. Vining added that Cigna, United Healthcare, and many of the Blue Cross/Blue Shield insurers have policies in place for virtual colonography.

Vining sees private insurance companies continuing to modify their coverage of CTC, and CAD may well be a part of that, if it is tied closely with efficiency and cost-effectiveness. CMS coverage will follow.

Vining said the American Cancer Society is lobbying Congress to push CMS for coverage of CTC. “Frankly, the fact that people aren’t lining up to get a colonoscopy is a pretty powerful tool. We’re not asking to abandon colonoscopy for screening, we’re just asking that [virtual colonography] be added as a viable, safe option.

“We’re making strides here,” he added. “If the CAD companies can sustain their existence, there will be a role for them. I think the biggest hurdle they all face is just hanging in there until the reimbursements show up.”

Dan Anderson is a contributing writer for Axis Imaging News.