Real-time 'functional' DTI aids cerebral palsy patients

Diffusion tensor imaging, an MR technique that produces images based on the orientation of water molecules, has mainly been used to evaluate brain white matter diseases. Now, Stanford University researchers have combined DTI with computer-generated musculoskeletal models to plan gait-correcting surgeries in children with cerebral palsy.

Current diagnostic methods, including clinical exam, ultrasound, CT, video, and electromyography, can only partly assess the causes of abnormal movement. Generally, these techniques tell specialists that the muscles and tendons on the medial side of the thigh contribute to excessive knee-hip rotation. That is not the story told by cine phase-contrast MRI and DTI tractography.

"We have found that these suspected muscles are not major players. Most of them rotate the hip in the opposite direction. So to expect medial rotation to improve by surgically lengthening them or injecting them with botulinum toxin is not consistent with what we find in the MR-based models," said coauthor Scott L. Delp, Ph.D., chair of Stanford's bioengineering department.

Led by Silvia Salinas Blemker, Ph.D., the researchers developed gait models from static MR images of four children with cerebral palsy. They then used cine phase-contrast MRI to analyze muscle motion in another group of pediatric patients who had undergone surgical transfer of the distal tendon of the rectus femoris. And in a separate, ongoing study, they are using that experience to develop the DTI tractography technique to characterize muscle architecture in vivo.

Although they have used the technique only with healthy adult subjects to date, the researchers will soon apply it in children with cerebral palsy. Delp presented the group's findings at the 2005 International Society for Magnetic Resonance in Medicine meeting.

"Fiber direction and orientation are key parts of a complex muscle model," said senior investigator Dr. Garry Gold, an assistant professor of radiology at Stanford. "DTI can accurately track muscle fibers, which previously had not been possible to do noninvasively."

Muscle models are key to understanding how humans walk. Until now, however, most muscle architecture modeling has been based on austere 2D samples from cadavers.

Real-time DTI tractography has the potential to preoperatively review possible treatment strategies and define the one that will result in improved gait, Gold said.

Although it is preliminary, the work at Stanford demonstrates DTI's potential utility outside neuroradiology, said Dr. Bruce Forster, an associate professor of radiology at the University of British Columbia in Vancouver. Most research using these sequences concentrates on mapping white matter tracts of the brain. Muscle fascicles, on the other hand, show a similar organization.

"A better understanding of muscle function using DTI would be yet another example of radiology extending its application to the functional organization of an anatomic region," Forster said.

The combined DTI/biomechanical modeling approach may be valuable in assessing treatment responses to conditions such as eosinophilia-myalgia syndrome or polymyositis. It could also be used in the assessment and management of muscle injury in athletes. The physiology and function of muscle contraction in humans has not been effectively studied in vivo, and DTI tractography has the potential to allow functional assessments of entire muscles or muscle groups, said Dr. Douglas P. Beall, chief of musculoskeletal imaging at the University of Oklahoma Health Sciences Center.

"The clinical application of this technique could have many uses, including the assessment of both normal muscles and muscles that have been damaged from a biomechanical injury or a particular disease process. Whatever the application, this new technique has the potential to establish a new era in the evaluation of muscle function," Beall said.