MR diffusion tensor imaging identifies and categorizes cervical spine trauma

In a novel approach to cervical spine trauma, researchers have found that changes in diffusion tensor imaging parameters are most marked at the spinal cord injury site and also reflect the severity of cord injury.

In the U.S., nearly 12,000 patients annually suffer spinal cord injuries. CT imaging nicely depicts fractures, dislocations, or normal musculoskeletal anatomy. MRI reveals soft tissue edema, ligament injury, and cord contusion with or without hemorrhage. Generally, patients with hemorrhage have a worse prognosis, according to lead investigator Dr. Sendhil Cheran, a radiology resident at the University of Maryland Medical Center in Baltimore.

Researchers have found, however, a subset of patients with unremarkable CT and conventional MR scans who show differences in diffusion tensor imaging (DTI).

Cheran and colleagues retrospectively reviewed the records of 50 symptomatic patients with neck trauma (average age 45 years) and 11 controls (average age 34 years). All patients and volunteers had undergone DTI.

In damaged white matter fibers, water molecules flow isotropically, rather than in a unified direction along a predetermined path. The DTI metric for this is fractional and/or relative anisotropy.

Additionally, damaged white matter fibers affect the diffusivity of extracellular water molecules by leaving them less room for movement. This measurement is called the apparent diffusion coefficient (ADC).

Researchers divided patients into four groups by their findings on conventional MR imaging:

• No cord contusion, normal CT, 12 patients

• No cord contusion, abnormal CT, 13 patients

• Cord contusion, no hemorrhage, 12 patients

• Cord contusion with hemorrhage, 13 patients

They found significantly different regional ADC values between the upper and mid and upper and lower cervical cord sections, whereas the fractional anisotropy was significantly different between the upper and lower sections.

Compared to volunteers, whole-cord ADC values were significantly lower in patients and in a subset of patients with soft tissue injury but no cord contusion. Whole-spine fractional anisotropy was not significantly decreased in patients.

The ADC and fractional anisotropy values of patients were significantly decreased at sites of cord injury when compared to volunteers. The greatest decrease in whole-cord ADC, fractional anisotropy, relative anisotropy, and volume ratio was observed in patients with hemorrhagic cord contusions when compared with the normal group. All reached significance.

The decrease in ADC paralleled the severity of the cervical spine injury, Cheran said.

"The really interesting changes we saw occurred in patients who didn't have cord contusion on the conventional MRI sequence," Cheran told Diagnostic Imaging. "Even in that subset of patients, we saw changes in the fractional anisotropy and ADC that approached significance. Those are the patients where we may be able show changes on DTI that aren't apparent on any other imaging modality that we are currently using clinically and correlate them with clinical findings."

Researchers did not determine individual DTI value thresholds. They looked at the mean for the entire group.

"That is something to do next," Cheran said.