CT Colonography gets ready for its close-up

After twice considering and rejecting endorsement of CT colonography as a cancer screening tool, the American Cancer Society is taking another look. Prompted by advances in reporting standards, technique, and training, the ACS may include the virtual colonoscopy technique as an optional test in its colon cancer screening guidelines by year's end. Such a move, coupled by favorable results from a national clinical trial, could push CTC into the practice mainstream, bolstering adoption, furthering development of computer-aided detection, and most important, fostering reimbursement.

The ACS first reviewed CTC in 1999 and again in 2003. While the technology improved in the interim, concerns about training, standardization, and accuracy lingered. Upgrades in all of the above have spurred reconsideration, according to Dr. Durado Brooks, director of prostate and colorectal cancers at the ACS.

The association has created a panel of colorectal cancer experts, gastrointestinal specialists, radiologists, and surgeons to review its guidelines, Brooks said at the Sixth International Virtual Colonoscopy Symposium in Boston last November. The target date for a potential endorsement is this fall.

Equally important to CTC's future are results of the national trial of CT colonography conducted by the American College of Radiology Imaging Network. In fact, findings from the ACRIN 6664 are expected to be released in the same timeframe as the revised ACS guidelines.

But some radiologists don't want to wait for the ACRIN to file its findings. Enough data are already available, and waiting will only freeze progress until mid- to late 2007, said Dr. Joseph Ferrucci, chair of radiology at Boston Medical Center.

"There are several things we can do while we wait for these ACRIN results. First, we're obtaining reimbursement here and there by local Medicare and private insurers for diagnostic symptomatic indications. Second, the community is working feverishly on so-called prepless exams and the companion shoe, CAD. We would like to find a way to push the ACR into talking to the Centers for Medicare and Medicaid Services and the ACS in detail now. We would like to propose that the ACR drop its support for barium enema screening," Ferrucci said.

REIMBURSEMENT STEPS

While national reimbursement for CTC is expected to remain on hold until the ACRIN results are finalized, users want to start the push for local reimbursement now. The first task is to obtain reimbursement for the procedure as an alternative to a failed optical colonoscopy, and for patients in whom a traditional exam is contraindicated.

"There are local care decisions by health insurance providers that allow reimbursement for CT colonography, typically in the setting of failed colonoscopy," said Dr. Judy Yee, vice chair of radiology at the University of California, San Francisco. "In the future, there should be reimbursement for other indications as well."

To date, 29 states have passed local Medicare coverage decisions for CT colonography for patients with incomplete colonoscopies, according to Dr. Beth McFarland, director of CT colonography at the Center for Diagnostic Imaging in Chesterfield, MO. Six more states may have similar coverage within the year.

Some local payers are already providing reimbursement for CT screening. Since April 2004, major third-party payers have offered initial coverage for CT colonography screening in the Madison, WI, area, according to Dr. Perry J. Pickhardt, an associate professor of radiology at the University of Wisconsin.

When the facility began its CTC screening program, it charged $1600 for the study. The charge is now about $1100, and about 40% of that cost is reimbursed by payers, Pickhardt said.

In 2003, Pickhardt authored a pioneering study, sponsored by the Department of Defense, of 3D primary reads and found that CTC compares favorably with optical colonoscopy for detecting significant polyps (NEJM 2003; 349[23]:2191-2200).

"Judging from our experience, we know that CT colonography is a very robust screening tool," he said. "For widespread acceptance to take hold, we need to convince primary-care providers and third-party payers that this is a generalizable, clinically-effective, cost-effective approach."

Pickhardt outlined his team's screening experience at the University of Wisconsin in a study presented at the 2005 RSNA meeting. The study included 1110 adults, few of whom paid out of pocket, he said.

The facility uses either eight- or 16-slice CT machines with a 1.25-mm collimation at 120 kVp. Virtual colonoscopy software from Viatronix is used to provide 3D primary reads with 2D confirmation. For polyps l0 mm and larger, patients are referred for an immediate colonoscopy. For medium-sized polyps measuring 5 to 9 mm, patients are given a choice of immediate optical colonoscopy or a CT colonography follow-up within one to two years.

With no polyps or in the case of a polyp less than 5 mm, Pickhardt recommends screening in five years. Results indicate a positive predictive value of 92.6%, he said.

The overall referral rate in the study for endoscopic procedures averaged 6%, which is lower than expected, he said. There were no complications following CTC.

Simply adding the virtual colonoscopy technique to the colon cancer screening program at the facility has attracted more patients, Pickhardt said. The team has now performed CTC screening exams on more than 2000 patients with no complications.

"CT colonography using a proven technique is a robust primary colorectal screening tool when covered by third-party payers," Pickhardt said.

This is the kind of study that deserves more attention, said Dr. Abraham Dachman, a professor of radiology at the University of Chicago.

"This is a model for a high-volume screening program. What we're seeing in this program are high positive predictive and negative predictive values. We need to see larger papers like this that will give data to generate more optimistic cost-effectiveness studies," Dachman said.

STANDARDS AND TRAINING

In addition to establishing a screening track record and attracting local reimbursement, CTC practitioners are also honing standards and training. In 2005, one important achievement was the release of ACR practice guidelines for CT colonography. The document provides guidelines for indications, qualifications and specifications for personnel and equipment, and specifications for exam interpretation and CT colonography reporting. It can be found on the ACR Web site.

In addition to these guidelines, an international virtual colonoscopy working group has developed a standardized reporting system for CT colonography. Dubbed C-RADS, for colonography reporting and data system, it helps describe CT colonography findings both in and outside the colon. It is akin to the breast imaging reporting and data system (BI-RADS) used in mammography. C-RADS was published in the July 2005 edition of Radiology.

A slew of commercial 3D imaging tools present physicians with myriad ways to conduct their studies. The same working group is developing practice guidelines for CT colonography with a goal of completing them this year.

"Although there continue to be advances in CT acquisition and 3D postprocessing, we will go nowhere if we cannot demonstrate standardization of best practice guidelines," McFarland said.

When it comes to training, CTC has a steep learning curve and is not the easiest exam to interpret. Recommendations for training include attendance at a training course and review of at least 50 CT colonography cases with colonoscopy correlation, Yee said.

Researchers at the University of Chicago are in the process of evaluating ways to improve training.

"We are studying the learning curve for virtual colonoscopy and how to train novice readers in efficient interpretation," Dachman said. "We should have that data ready by the next RSNA meeting, if not earlier."

Continuing medical education is a must for radiologists planning to perform CT colonography, as indicated by the results of two studies presented by U.K. researchers at the 2005 RSNA meeting.

One study revealed that CT colonography reading skills among 13 nonacademic radiologists at seven British health facilities are highly variable and tied to the clinician's experience with the procedure. The study was conducted by researchers in the department of specialist radiology at University College Hospital in London.

Individual accuracy of the physicians studied, all of whom subspecialize in abdominal radiology, ranged from 53% to 93% in the interpretation of 15 cases validated by conventional colonoscopy. The mean accuracy of the physicians was 75%. But the accuracy rate among experienced readers, defined as those who have read as many as 350 cases, was 88%. Among nonexperts, who have read as few as five cases, the accuracy rate fell to 71%.

In a second study, Dr. Emily Tam of University College Hospital-University Park Clinic found that general radiology experience does not guarantee proficiency when physicians begin performing CTC.

Her test of reading performance found that even without specific training in CTC, an inexperienced resident can outperform a board-certified radiologist with subspecialty training in gastrointestinal radiology and a radiologist with more general training. For 1 to 5-mm polyps, the resident's detection rate was 14%, the same rate as the specialist and double the 7% rate of the generalist.

For polyps ranging in size from 6 mm to 9 mm, the detection rates of the resident, specialist, and generalist were 42%, 25%, and 25%, respectively. For lesions larger than 10 mm, detection rates were 83%, 17%, and 67%, respectively. Tam attributed the unexpected results to a lack of CTC training.

Computer-aided detection could bolster training as well as accuracy rates for virtual colonoscopy. Papers presented at the 2005 RSNA meeting highlighted CAD's potential for picking up polyps.

One study using CAD software developed by the National Institutes of Health examined data from 1186 colon cancer screening patients at three medical centers. The population contained two cancers and 180 adenomas 8 mm or larger, according to lead author Dr. Ronald Summers, chief of clinical image processing, virtual endoscopy, and computer-aided diagnosis at the NIH.

CAD achieved per-polyp sensitivities of 76.4% for polyps 8 mm or larger and 96.2% for polyps 10 mm or larger. CAD detected both cancers and exhibited a 0.7 false-positive rate per patient.

"Our CAD systems had high performance comparable to optical colonoscopy for polyps greater than or equal to 1 cm," Summers said.

The NIH CAD program was compared with the Siemens Medical Systems polyp-enhanced viewing software in a separate study presented by Dr. Joel Fletcher, an assistant professor of radiology at the Mayo Clinic in Rochester, MN.

Using a library of 65 clinical CT colonography data sets, Fletcher compared the two proprietary CAD systems. All cases included colonoscopy correlation.

The Siemens CAD system exhibited a 61% sensitivity for detecting lesions 1 cm or larger compared with a 96% sensitivity for the NIH system. However, the Siemens CAD showed a better specificity, with one false positive per case compared with 4.6 per case for the NIH CAD.

The performance characteristics of a CAD system will usually depend on whether the radiologist uses it as a first-read tool or to check a primary read, Fletcher said.

The Siemens system underwent a second workout in a study led by Dr. Anno Graser from the University of Munich, Grosshadern. Graser tested it in 105 cases. CAD demonstrated a 2.2 false-positive rate per data set. It yielded a 94% sensitivity for medium polyps (6 to 9 mm) and 87% for polyps 10 mm or larger, compared with 94% and 93%, respectively, for expert readers.

"CAD shows promising sensitivity in the detection of clinically significant polyps but does not outperform expert readers," Graser said.

THE TARGET LESION

As CTC continues to advance, physicians are better able to view and detect smaller lesions. But controversy rages over what to do with these smaller polyps.

"Determination of what polyp size is clinically important will have a tremendous influence on both further validation studies and clinical implementation in everyday practice," McFarland said. "CT technology now allows us to see small lesions, but to be cost-effective, consistent reporting of lesions that are clinically significant will be essential."

Polyps can be separated into three size categories:

- small: smaller than 5 mm;

- medium: between 6- and 9 mm; and

- large: 10 mm and larger.

There isn't much question in the CT colonography community about what to do with the large polyps-they are clinically significant and should be sent off for a colonoscopy follow-up. Many would note that the small lesions can be safely ignored.

It is the medium-sized polyps that provoke debate. A small number of these polyps will have problematic pathology.

"Within the inherent error of the measurement of these lesions, hopefully we can attain some level of multidisciplinary agreement on how to manage such lesions in the clinical context of patient age, comorbidity, and colorectal symptoms," McFarland said.

The C-RADS consensus addresses the topic and offers preliminary recommendations of surveillance for these polyps, according to McFarland.

At the University of Wisconsin, Pickhardt and colleagues follow the surveillance protocol for these mid-sized polyps because of information gleaned from a study on the natural history of polyps by Hofstad et al (Gastrointest Endosc Clin N Am 1997 Jul;7(3):345-363). That study found that only 0.5% of subcentimeter polyps grew larger than the 10 mm threshold, Pickhardt said. Researchers found that most of the medium-sized polyps actually regressed in size.

WILL TRIALS QUELL DEBATE?

The landmark CT colonography trial published by Pickhardt in 2003 gave CTC high marks, with a sensitivity in polyps 6 to 9 mm of 83.6% and in polyps 9 mm or larger of 92.2%. But on the heels of that study, several others have reported much lower sensitivities.

Media coverage has focused on several recent CTC studies:

- Rockey (2005). Found sensitivity in polyps 6 to 9 mm to be 60%; in polyps greater than 9 mm, 64%;

- Cotton (2004). Reported sensitivity in polyps 6 to 9 mm of 22.7%; and in polyps greater than 9 mm, 51.9%; and

- Johnson (2003). Found sensitivity in polyps 6 to 9 mm to be 47.1%; and in polyps greater than 9 mm, 46.3%.

With such a large discrepancy in sensitivities, some have said that more multicenter trials are warranted. Three such trials are in development.

The ACRIN trial will look at 2D versus 3D interpretation, patient prep issues, and cost-effectiveness. The $7 million trial is in the accrual phase, with a goal of recruiting 2600 subjects.

In the U.K., recruitment for the Special Interest Group in Gastrointestinal and Abdominal Radiology 1 Trial is under way. The trial will compare virtual colonoscopy to both optical colonoscopy and barium enema. The patient population is symptomatic, and investigators have set a recruitment target of 4500 patients, with current recruitment at about 800.

Elsewhere, 14 sites are already online for the Italian Multicenter Study on Accuracy of CT Colonography Trial. Investigators will assess sensitivity and specificity for virtual colonoscopy in detecting advanced adenomas and cancers. In order to include 150 patients with positive findings at optical colonoscopy, investigators plan to enroll between 350 and 3550 patients.

Some radiologists question whether these large trials are necessary. But it is unlikely that virtual colonoscopy will achieve a national coverage decision, or widespread adoption, without replication of Pickhardt's good results in other multicenter studies.

"Assuming these trials can demonstrate that other CT colonography platforms can provide acceptable results, then reimbursement would be the next logical step, particularly given the need for increased screening," Pickhardt said.

Others say enough data exist to prove that virtual colonoscopy is a legitimate addition to the colon cancer screening arsenal.

"Although the ACRIN results will have an important influence on screening patients, a recent meta-analysis (Ann Intern Med 2005 Apr 19;142(8):635-650) demonstrates a remarkably positive summary of results, including all large trials to date," said McFarland.

The meta-analysis reported average sensitivities of 70% for 6 to 9-mm polyps and 85% for polyps greater than or equal to 10 mm, according to McFarland.

"Those results would demonstrate CT colonography to be a strong second to traditional colonoscopy, compared to all colorectal screening tests in current practice," McFarland said.

James Brice contributed to this article.

Ms. Trevino is an assistant editor of Diagnostic Imaging.