MR colonography evolves to meet screening needs

MR colonography has yet to capture radiologists' imagination to the same extent as CT. But the radiation-free exam has a bright future, especially if stool-tagging techniques can avoid the need for bowel cleansing, according to speakers at ECR 2006.

Patients with incomplete endoscopy results and individuals requiring an assessment of inflammatory bowel disease can already benefit from MR colonography, said Dr. Thomas Lauenstein, formerly of the University of Essen, Germany, and now an assistant professor of radiology at Emory University in Atlanta, Georgia. The technique may also play a role in colorectal cancer screening if it can demonstrate sufficient accuracy and patient acceptance.

Published figures put the sensitivity of MR colonography at about 85% for polyps 5 to 10 mm in diameter and 100% for polyps greater than 10 mm. These studies were based on cancer screening in a high-risk population, however, so results may have been influenced by radiologists' expectations of finding pathology, he said.

While working at Essen, Lauenstein and colleagues conducted a trial designed to reflect a general screening population using 294 subjects over age 50. MR colonoscopy showed a sensitivity of 74% for the 5 to 10-mm polyps and a specificity of 97% compared with optical colonoscopy.

"We are not perfect, but we are quite good," he said. "But we also have to get away from the idea that colonoscopy itself has a sensitivity of 100%. That's not true."

One surprise for researchers was patient feedback, which showed equal acceptance for MR colonography and endoscopy. Patients' top dislikes included insertion of a rectal tube during MR colonography and the need for bowel cleansing with optical colonoscopy. These could both be avoided in future MR colonography protocols by switching the tube for a catheter and finding a robust, reliable method for fecal tagging, Lauenstein said.

Radiologists trained in MR colonography interpretation should be able to review scan results within 10 minutes, he said. His preferred routine is to scroll through contrast-enhanced T1-weighted data sets in multiplanar reformat (MPR) mode. Identification of suspect lesions prompts a review of the relevant precontrast images. Residual fecal matter will appear the same on both sets of images, whereas lesions will show a change in enhancement. The evaluation finishes with a review of T2-weighted images.

Wider public awareness of risks posed by medical radiation exposure could drive the popularity of MR colonography, he said, predicting that the relative positions of CT and MR colonography could be quite different in five years.

Speaking at the same session, Dr. Nikolaus Papanikolaou from the radiology department at the University of Crete showed how alternative sequence options and choice of contrast agent can influence the clarity of MR-based bowel exams.

Spoiled gradient-echo sequences, such as FLASH, offer high contrast between the lumen and colonic wall, aiding detection of extraluminal abnormalities. Three-D FLASH provides near-isotropic resolution, making it suitable for MPRs and generation of endoscopic views. The drawback of this sequence is that it is more prone to artifacts than some others.

One option for T2-weighted imaging is the HASTE sequence, Papanikolaou said. This ultrafast pulse sequence is insensitive to motion and susceptibility artifacts. It is unable to demonstrate extraluminal disease, however. Single-shot turbo spin-echo MRI is insensitive to artifacts, too. But increased attenuation of signal from soft tissue restricts its use to luminal imaging.

Steady-state free precession sequences are important for bowel imaging as well, he said. Two-D TrueFISP, for instance, is extremely sensitive to extraluminal disease and has no problems with motion artifacts. But susceptibility artifacts can be an issue. Efforts have also been made to use 3D TrueFISP clinically. While this has the added advantage of enabling MPRs, it too suffers from susceptibility artifacts.

Agents used to distend the colonic lumen prior to imaging include barium sulfate, pure water, water spiked with paramagnetic contrast, air, and carbon dioxide. Water is cheap and safe, but it can sometimes cause patient discomfort, Papanikolaou said. Adding gadolinium improves the signal-to-noise ratio significantly but also increases cost. Both air and carbon dioxide are inexpensive and risk-free and are better tolerated by patients. Top-end scanners are required to avoid susceptibility artifacts when using these agents, however.

Patients undergoing MR colonography at the University of Crete are imaged first with a T1-weighted gradient-echo sequence to monitor contrast administration. Three-D T1-weighted spoiled gradient-echo imaging is then performed with patients prone and supine. The positional change avoids any possible misinterpretation of trapped air as pathology.

The greatest influence on MR colonography's future development will be hardware advances, however.

"Further technical improvements may increase the likely role of MR colonography in colon cancer screening," he said. "I believe that the combination of parallel acquisition techniques and imaging at 3T could be the future of MR colonography."