Report from ISMRM: Ultrahigh-field MRI improves characterization of multiple sclerosis

May 9, 2008

Diagnostic power and field strength go hand in hand when it comes to evaluating multiple sclerosis.

Diagnostic power and field strength go hand in hand when it comes to evaluating multiple sclerosis.

It is no secret among board-certified radiologists that 1.5T MRI misses many MS-related white-matter lesions found during postmortem histology. It is unable to characterize much gray matter pathology at all.

This problem stems from limitations of spatial resolution and contrast that are solved by raising the field strength to 4.7T, according to Dr. Amir Eissa of the University of Alberta.

Eissa had excellent 4.7T images to show at the International Society for Magnetic Resonance in Medicine meeting, including 2D fast spin-echo and 3D susceptibility-weighted images acquired with a flow-compensated T2*-weighted gradient-echo sequence.

A sample minimum intensity projection from a susceptibility-weighted image showing a coronal view of the midbrain revealed deep penetrating veins and deep brain structures, both showing a sensitivity to iron.

Based on experience with seven consecutive patients with chronic MS, Eissa correctly characterized 82 of 99 known lesions with the FSE sequence and 26 of 99 with phase imaging. Phase imaging alone uncovered 17 lesions not visible on FSE. The lesions were hyperintense with FSE and hypointense on the phase images.

"The combination of FSE and SWI at 4.7T provided two unique forms of image contrast for MS lesion visualization," Eissa said. "Together, they can achieve a higher level of lesion conspicuity than imaging at a lower field strength."

Dr. Kathyrn Hammond from the University of California, San Francisco presented preliminary results from 7T MRIs of 18 patients with relapsing-remitting MS and 13 age- and gender-matched controls to explore the influence of magnetic-susceptibility-shifted iron accumulations on the disease. The accumulation of iron over time in the brains of MS patients creates an incentive for phase imaging, to capitalize on its effects, she said.

Imaging in Hammond's trial focused on the basal ganglia, where she compared the magnetic field distortion among MS patients with that of controls.

Anatomical magnitude images were acquired with gradient-echo imaging. They then underwent postprocessing to produce phase images of the B0 field.

Though phase imaging missed some lesions that were identified with magnitude imaging, it proved superior on all other counts.


As expected, the magnetic field in the basal ganglia was more paramagnetic in the MS patients than in the healthy controls. Hammond identified a significant correlation between the amount of magnetism in the caudate and the duration of the disease. A similar pattern was not identified in the putamen.

About 67% of the lesions observed with phase imaging were accompanied by penetrating veins. About 83% were surrounded by phase contrast that is thought to be associated with iron-rich macrophages. Fewer penetrating veins were seen with magnitude imaging. Their edges tended to be affected by blooming artifact.

The next step is a follow-up trial involving histology to confirm these findings, she said.

Standing in for first author Dr. Caterina Mainero, a radiology instructor at Massachusetts General Hospital, Dr. Bruce Rosen, associate director of the Athinoula A. Martinos Biomedical Imaging Center at MGH, described results of a trial assessing the ability of 7T MRI to characterize the three different types of cortical MS lesions that have not been consistently observed at lower field strengths:


  • Type I: leukocortical lesions that involve the gray/white matter boundary but do not extend to the pial surface


  • Type II: intracortical lesions located within the cerebral cortex but not extending to the surface of the brain or to the subcortical white matter


  • Type III: subpial, wedge-shaped lesions that extend from the pial surface to cortical layers 3 and 4


  • Type IV: lesions that affect the entire width of the cortex from pial surface to white matter.

"If one can see them in vivo, it would give use a better sense of their natural history, their pathogenic evolutions, and allow us another biomarker to see the effect of treatment," Rosen said.

The trial, involving 14 MS patients and six age-matched healthy control subjects, indicated that 7T handles the task well.

Two-D Flash T2* spoiled gradient echo-weighted images and T2-weighted turbo spin-echo with the same resolution and orientation as the Flash T2*-weighted scans were acquired to allow whole-brain coverage from the level of the junction of the brain stem and cerebellum. MPRAGE was also performed to improve gray/white matter boundary identification; 1-mm slices were acquired. Real-time shimming was used to minimize magnetic field B0 field inhomogeneities.

Of 91 lesions, 29 (32%) were leukocortical, 14 (15%) were intracortical, and 47 (52%) were subpial or type IV. One lesion was classified as undefined.

Nearly all of the lesions were identified on the T2- and T2*-weighted images, though conspicuity was greatest with the T2*-weighted Flash sequence.

There was no significant difference in the number of lesions found in the secondary-progressive and relapsing-remitting patients.

The MGH researchers also examined disseminated disease after identifying broad areas of signal hyperintensity in the brains of the MS patients that were not observed in the normal controls. Subsequent evaluations found a pronounced right shift in the MS patients compared with controls, suggesting a global increase in signal in the subpial regions. Most of the cortical signal abnormalities were localized in the prefrontal regions of the cingulate cortex but not found in the occipital cortex, Rosen said.

For more online information, visit Diagnostic Imaging's ISMRM Webcast.