One mild day in San Francisco in late September, a few dozen people from all over the globe passed around sample rectal tubes, boxes of various bowel preparations, and other clinical paraphernalia. The props were part of a hands-on virtual colonoscopy workshop with instructor Dr. Judy Yee, an expert in the blossoming technique.
Virtual colonoscopy is about to take off, said Yee, vice chair of radiology at the University of California, San Francisco. She has cause for optimism. UCSF was part of a groundbreaking national CT colonography trial sponsored by the American College of Radiology Imaging Network.
The same day the UCSF workshop was held, glowing results from the ACRIN 6664 study were announced at a meeting in Washington, DC: CTC was comparable to the gold standard colonoscopy for screening intermediate and large-sized polyps. And since prevalence of lesions sized 6 mm or larger was low at 8.3%, most patients under-going CTC would not need to undergo therapeutic colonoscopy. Yee told her workshop attendees that based on the results, CTC was validated as a screening test for colorectal cancer and detection of large and medium-sized polyps.
The results signify the end of the technique's long journey toward validation and a "stunning achievement" for ACRIN, as well as principal investigator Dr. C. Daniel Johnson, said Dr. Joseph Ferrucci, chair of radiology at the University of Massachusetts.
According to another prominent study from the University of Wisconsin, CTC achieves similar outcomes with a fraction of the number of polypectomies. Results were published in The New England Journal of Medicine in October. Results from other major trials showing positive results for CTC were released at a virtual colonoscopy symposium in Boston in mid-October (see table).
The data suggest CTC is a noninvasive "disruptive technology," with potential to displace the invasive colonoscopy standard as a primary screening tool, Ferrucci said. Increasingly, the exam will be viewed as a patient-friendly filter to pick out those few patients who actually need a colonoscopy.
Until now, take-up of virtual colonoscopy has been limited, despite some very promising single-center studies. Coverage for screening is almost unheard of. Local Medicare carriers cover CTC for diagnostic purposes but vary in terms and conditions. Most cover the exam in case of a failed colonoscopy, and in about 20 states coverage is granted for symptomatic patients if colonoscopy is contraindicated. Private payers have offered more limited coverage for diagnostic CTC.
Inclusion in the American Cancer Society colorectal cancer screening guidelines, which were expected to be revised at the end of 2007, would give the procedure a big boost in the reimbursement arena.
If the ACS endorses CTC, other organizations will follow suit and there will be a much more rapid progression toward widespread reimbursement, said Dr. Mark Klein, a radiologist at Washington Radiology Associates in the nation's capital.
"Once reimbursement is widespread, we will see adoption by radiologists and gastroenterologists. We are hoping to see major growth in this area over the next 24 months," Klein said.
Yee points out that screening participation remains relatively low despite heavy media coverage about colorectal cancer. Only 37% of cases are diagnosed when disease is localized, which has a negative impact on treatment options and survival rates.
The available screening tests have disadvantages that might be putting people off screening. The fecal occult blood test has low positive predictive value, and patients have reported pain during the double-contrast barium enema. Optical colonoscopy is invasive and carries the risk of bowel perforation (one in 1000 for diagnostic and one in 500 for therapeutic procedures).
"The nice thing about virtual colonoscopy is that you almost never have a bowel perforation and don't need to sedate the patient. The risk is close to zero, which is really what you want in a screening exam," Klein said.
OKLAHOMA LAND RUSH
In addition to the 75 million people currently eligible for screening, another four million will become eligible for screening every year, and more providers will be needed to screen an aging population. The ACR is planning to roll out a training center that can accommodate large groups of people in 2008.
Increased reimbursement could trigger a kind of "Oklahoma land rush," as greater numbers of gastroenterologists and radiologists seek to break into CTC. But the costs of acquiring a CT scanner and difficulties in securing interpretations of extracolonic findings may deter nonradiologists, Ferrucci said.
There will be some turf issues initially, but those physicians who do the best job will get the patients, Klein said. Radiologists should also bear in mind that tens of millions of scans need to be done.
CTC providers will need to be accountable for performance. The ACR CTC practice guideline, published in October 2005, recommends that radiologists read a minimum of 50 cases. This recommendation can be achieved in three ways: formal hands-on training, supervision with a trained physician acting as a double reader, and correlation of CTC and endoscopy in patients who have undergone both procedures.
The American Gastroenterological Association published CTC performance and interpretation standards for members in September. The organization recommends review and correlation of 75 cases with endoscopic correlation, followed by six weeks of mentored training.
Studies show there is a lot of variability in performance for detection of polyps, regardless of the reader's experience, said Dr. Joel Fletcher, an associate professor of radiology at the Mayo Clinic, during a presentation at the Eighth Annual Symposium on Virtual Colonoscopy in Boston in October.
Research also suggests that 50 cases may be inadequate for radiologists and that optimal sensitivity is achieved after 75 studies-and nonradiologists need even more training, Fletcher said. According to the literature, radiologists with extensive experience in nonvalidated cases but without formal training may still have suboptimal sensitivity.
It's easy to train individuals about indications and performance. Instruction in use of the workstation, optimization of 2D and 3D search criteria, and problem-solving techniques like decubitus imaging, is more challenging.
Following training, a period of second reading by a mentor helps practitioners identify strengths and weaknesses. At the Mayo Clinic, new GI radiologists read 50 to 60 data sets and then their studies are double-read for four to six weeks, Fletcher said.
In the future, CAD could play a role as a second reader to improve sensitivity, because most of the errors in interpretation are errors of detection. Formalized testing could also help trainees focus on idiosyncratic weaknesses and educate them in challenging cases, Fletcher said.
ACRIN 6664 researchers have attributed the trial's success in part to reader training and certification. Participant radiologists had to either read 500 studies or undergo one and a half days of training, including at least 50 cases. They also had to pass a certifying exam, which required detection of 90% of abnormal lesions 1 cm or larger. More than half of readers did not initially pass and needed additional training before they could become certified. Significant differences in performance were not demonstrated in the study.
Testing could become more feasible with the development of large consolidated training courses in virtual colonoscopy, Fletcher said. Furthermore, the ACR has also established a task force to establish metrics for virtual colonoscopy and to develop a national database for performance and safety standards in CTC.
HONING TECHNIQUES
The consensus is that both 2D and 3D techniques are necessary and both are effective for making the primary interpretation, with minor trade-offs, Ferrucci said. The ACRIN 6664 trial showed equal performance for 2D and 3D methods in primary reading, but 3D required almost six minutes longer for primary review (25.5 minutes versus 19.4 minutes).
However, Klein said that the best 3D software packages were not used across the participating centers. Using optimal software, 3D reading would require half the time reported in the ACRIN study.
New and innovative techniques for image display, such as the ability to flatten out the colon and read it as a filet, could speed up reading time and boost efficiency further, Ferruci said.
In terms of bowel prep, there is growing consensus that prep can be done with barium, iodine(Drug information on iodine), or both. Minimal-prep and prepless techniques are being investigated and could become available in a few years, increasing appeal for patients.
HARD SELL FOR PUBLIC?
From the patient point of view, cost and primary-care recommendation are important factors that influence screening choice. But the biggest issue is convenience. Patients typically ask themselves how many hoops they have to go through for screening, said Dr. Ronald Myers, director of cancer prevention, control outcomes, and research at the Kimmel Cancer Center at Thomas Jefferson University.
With CTC, they might calculate that the screening process could require double the number of steps, because if the first exam is positive, they then proceed to therapeutic colonoscopy. In terms of scheduling their day, CTC is not making that much of a difference, because patients will still need to arrange for someone to pick them up in case they need the follow-up colonoscopy.
The benefits of CTC compared with colonoscopy are, to some extent, lost in the mind of the public, and CTC could actually prove to be a hard sell, Myers said.
However, CTC could be a good option for patients who do not want to undergo colonoscopy, with great potential to increase screening rates and further decrease mortality, he said.
Experience at the Colon Health Initiative by the National Naval Medical Center in Bethesda, MD, indicates that patients respond well to CTC readings and same-day therapeutic colonoscopy when necessary, said Dr. Duncan Barlow, senior radiologist.
Ferrucci believes that radiologists will need to reeducate the public about the nature of polyps. With traditional colonoscopy, all polyps-including diminutive ones-are removed, yet a very low percentage of polyps are actually life-threatening.
For example, in the new NEJM study from the University of Wisconsin, the number of polypectomies was much larger in the optical colonoscopy group compared with the CTC group (2434 versus 561), but outcomes were similar. And of 2006 polypectomies performed for diminutive polyps during optical colonoscopy, only four turned out to be advanced adenomas.
"It will be interesting to see how that spin on the significance of polyps plays out," Ferrucci said. "Radiologists have one version of the story. Colonoscopists have the other. Those in the middle will have to duck, because there will be a war of words."
Emily Hayes is feature editor of Diagnostic Imaging.
