CT vs. MR: Virtual Colonography Making A Mark on Diagnosis & Care

Virtual colonography (VC) stands poised on the brink of acceptance as a screening tool for colon cancer, tempered by the realities of reimbursement issues, while a whirlwind of research activity on numerous fronts excels. While optical colonoscopy remains the gold standard, advances in the virtual model show promise for future screening activities. Currently CT colonography is used in certain specific instances, such as when optical colonoscopy has failed or with patients for which anesthesia poses a risk.

In the United States, most of the activity focuses on the use of multidetector CT scanners for colonography exams, while in Europe, researchers pursue feasibility studies and technique refinement for using magnetic resonance imaging for the same purpose. CT research encompasses perfecting preparation and imaging techniques to obtain reliable results, development of computer assisted detection capabilities, and advances in software-based computer algorithms to improve analysis of 2D and produce 3D image data sets.

The objective of MR and CT colonography is the same - finding colorectal cancer early as well as adenomatous polyps that can lead to cancer over a decade or more. Colorectal cancer strikes 130,000 Americans each year and kills about 56,000. And interestingly, about 75 percent of all new cases of colorectal cancer occur in people who have no known risk factors other than age - with research indicating that the risk for colorectal cancer increases substantially after age 50. Further, the American Cancer Society estimates that only one-third of at-risk Americans are being screened each year for colon cancer via colonoscopy, whether invasive or virtual - so more patient-friendly testing is needed to increase the numbers of Americans being screened.


MR colonography

MR colonography is in the earliest stages of development, with the prospect of several years before large scale clinical trials will be conducted to study diagnostic efficacy, according to D.J. Lomas, MD, professor of clinical MRI at the University of Cambridge and Addenbrookes Hospital in Cambridge, U.K., who has been involved in a pilot study to determine the feasibility of using MR imaging for VC.

A 1.5T MRI scanner is a prerequisite to MR colonography, Lomas has found in his research using a GE 1.5T Signa Excite HD scanner with multichannel array coils that cover the abdomen and pelvis.

One area of his research involves the use of MR fluoroscopy with very rapid MR imaging with frame rates varying from very slow to one per second to as many as 10 to 15 frames per second. As these techniques become refined, he sees promise in future applications.

"To use an analogy: Contrast-enhanced MRA [magnetic resonance angiography] when it was being developed 10 years ago was considered a clumsy and lengthy procedure," says Lomas. "As technique applications and technology were integrated and improved, it has now become a routine examination that is performed simply."

Judy Yee, MD, associate professor and vice chair of radiology for the University of California at San Francisco and chief of radiology at the San Francisco Veterans Administration Hospital, explains an obvious advantage to MR in that there is no ionizing radiation. This means that for certain patients, such as young women, the use of MR would be preferable.

However, Yee notes, CT is more readily available and less costly than MR. "The other advantage for CT is that the procedural time is less than MR - on the order of less than half the time for MR colonoscopy," she says. "For CT, the scan time is so fast that you can scan the entire abdomen and pelvis in both supine and prone positions, and the actual scan time in each position is about 10 seconds. With MRI, there are additional sequences required, and each of the sequences takes longer."

Multislice CT may not be essential for colonography, but it is certainly preferred, Lee says. "The advantage is that you are able to scan using thinner collimation, which improves your resolution on the axial and 3D images."

Finally, Yee explains that although clinical studies suggest that MR may be comparable to CT in performance when looking for large polyps (10mm or greater), there would need to be other advantages of MR to propel it into wide acceptance for colonography screening. There are fewer research centers in the U.S. studying the benefits of this approach, and CT scanners are more widely available here.

Over the past few years in the U.S., research activity devoted to the use of CT for colonography studies has explored a number of different aspects of the field including technique, technology and important reimbursement issues.


CT colonography

Perry J. Pickhardt, MD, associate professor of radiology at the University of Wisconsin Medical School in Madison has developed a vigorous, and well-known program of CT colonography screening that receives third-party reimbursement from local HMOs. With approximately 2,000 screening exams performed since April 2004, the program has a proven track record.

To reach this level of achievement, Pickhardt and his colleagues have developed a specific protocol for colon preparation (patent pending), imaging and analysis of the data and published the details of their approach in the April issue of Radiology (235:17-20).

All VC studies are performed in the morning, so that when a finding requires optical colonoscopy for polyp removal, the patient can be scheduled for that procedure immediately following the morning scan.

The first step involves the use of a low-volume preparation protocol for colon cleansing with addition of diluted 2 percent barium for solid stool tagging and an ionic water-soluble contrast agent for fluid opacification. This technique brings dramatically increased accuracy in polyp differentiation from remnants of fecal material or fluids, Pickhardt says.

Next, the scan is performed using either their GE LightSpeed 8-slice or LightSpeed 16-slice scanner. The data are analyzed using Viatronix (V3D-Colon) software, with review of 3D data to detect potential lesions that are confirmed with 2D slices.

"Using 2D is very difficult for searching patterns," explains Pickhardt. "If you flip-flop it and use 3D to find a lesion and 2D to confirm it, that makes it much easier." He suggests that everyone will use this methodology once their software has evolved to facilitate these activities. "We're able to read them in 10 to 15 minutes, but the study has to be sent to the workstation where they build the 3D model first."

The protocol produces a win/win for both radiology and gastroenterology because only patients with known findings are sent for optical colonoscopy exams. That way the GI team doesn't waste their time performing negative colonoscopy procedures. Besides, although complications with traditional colonoscopy studies are rare, this non-invasive technique renders them practically non-existent.

Abraham H. Dachman, MD, FACR, professor of radiology at the University of Chicago and director of Computed Tomography, says they have experienced an increase in VC patients who are at high risk of complications from the conventional approach due to their status as cardiac transplant candidates.


Ongoing research

Exciting directions in clinical research hold great promise for the future of VC, too. Dachman's group at the University of Chicago, for one, has developed both a research and non-research clinical program. Collaborating with their computer division, Rossmann Laboratories, they've built a program for CAD which they have begun to market. They are participants in the large ACRIN (American College of Radiology Imaging Network) prospective trial in VC, and they're studying the learning curve for teaching VC to non-experts. They've designed an intensive two-day course to impart necessary knowledge.

Dachman notes that several image software companies offer programs that "cut the colon open and lay it flat in a filet" style that he believes may prove valuable for non-experts to speed up their reading times. His group uses scanners from GE and Philips Medical Systems, and advanced visualization software from Vital Images.

"I think the next major advance will be a high-quality CAD program…that is easy to use with a low number of false positives … and high sensitivity and specificity," says Dachman.

He also strongly urges anyone starting a VC program to devote significant time and energy to appropriate training. "If someone has the time to devote to proper training, I think [virtual colonoscopy] is ready for the right radiologist."

Michael Macari, MD, associate professor of radiology at NYU School of Medicine echoes the importance of proper education in these techniques. His team provides five training programs a year where participants learn how to interpret the data, differentiate residual fecal material from polyps and how to identify polyps from the data. Their students sit in front of a workstation and review 50 proven cases where the outcome is known, followed by a second read with an expert.

"I think in the end, VC will have a major impact on colon cancer screening because it will be able to detect the vast majority of significant polyps," says Macari.

His practice performs many VC scans following an incomplete traditional colonoscopy where the gastroenterologist is unable to advance the scope. That could include a patient who has obstructive cancer and the surgeon wants to know if there is a second lesion proximal to the original site, with about 5 percent of patients falling into that category. They also do VC for older, frail patients or those with cardiac conditions who are not candidates for sedation.

"Virtual colonography is a much safer examination for a number of reasons, including that it is non-invasive, and also because the patient is awake," Macari concludes.

Markus Juchems, MD, radiologist at the University Hospital in Ulm (Germany) is using the Philips Brilliance 40 CT scanner to research a low-dose protocol.

"We reduce the mAs [milliampere-second] from 150 to 50, and with doing that, we can reduce the dose to one-third of what is normal," says Juchems. "We get to 2.8 milliSievert (mSv), which is quite good for a screening setting."

The design of this study with 200 subjects, includes a VC in the morning followed by a colonoscopy in the gastroenterology department. They "unblind" the endoscopist segment by segment of the colon. In other words, when the endoscopist examines the transverse colon and does not find a polyp, and Juchems tells him there is something, the GI physician has a second chance to look for it. If he still doesn't see it, it is a false negative. Thus far, after 50 patients, they are finding the lower dose to be an effective means of screening.

Prior to this study, Juchems was involved in beta testing the "filet view" VC interpretation software that "dissects the colon and unfolds it" from the Philips Virtual Colonoscopy tool (version 2.x) on the Brilliance Workspace.

UCSF's Yee has a 3D imaging lab with products from all of the major vendors where she conducts research to help develop software applications. She has been involved in designing a variety of 3D applications including the endoluminal fly-through, or one where a small volume of the colon is visualized in cube form that can be rotated in any direction for review of any area, as well as the "filet view" where the colon is divided along its long axis and opened for display.

She says that people always ask her which vendor's approach she prefers, and she tells them it really varies. "It depends on a particular lesion which may look better on one system than another, or it may be easier to manipulate on one system than another." Additionally, user preference enters into the decision-making process.

Returning to the issue of MR versus CT for virtual colonography, Yee explains that she thinks there will have to be numerous advantages of MR over CT before it becomes widely adopted. Since more people are working in the field of CT virtual colonography, it is unlikely that MR will eclipse CT in this arena, despite its obvious advantages in soft-tissue visualization.

Multidetector CT use for colonography is sure to be a hot topic at RSNA this month, too. At least two relevant abstracts should stimulate discussion, according to Jay P. Heiken, MD, chair of the subcommittee on gastrointestinal radiology. One abstract involves using multidetector colonography without bowel preparation, but with nonionic isomolar contrast medium for fecal tagging that yielded a diagnostic accuracy comparable to conventional optical colonoscopy for the detection of polyps of 10mm in diameter or larger, while another found multidetector VC without cathartic bowel preparation compared favorably with conventional colonoscopy for the detection of polyps 8 mm in diameter or larger.