BACKGROUND: Magnetic resonance imaging is useful in evaluating acute spinal cord injury. Apparent diffusion coefficient (ADC) values obtained by diffusion-weighted imaging can differentiate cytotoxic edema from vasogenic edema through microscopic motion of water protons.
OBJECTIVE: To determine whether ADC values in the cervical spinal cord match neurological grades and thus predict functional recovery in patients suffering from cervical spinal cord injury.
METHODS: Diffusion-weighted images were obtained using 15 axial slices covering the cervical spinal cord from 16 consecutive patients. ADC values were determined for both gray and white matter. All patients were treated surgically. Patient neurological status was evaluated preoperatively and postoperatively with the Frankel classification and neurosurgical cervical spine scale. One patient had complete spinal cord injury and showed no recovery. Using 15 patients with incomplete injury, we analyzed correlations between preoperative ADC values and neurological grading, degree of postoperative recovery, or cavity formation in follow-up magnetic resonance images. For comparison, ADC values of 11 healthy volunteers were also calculated.
RESULTS: There was significant correlation between ADC values and degree of postoperative recovery (P = .02). ADC values of patients showing cavity formation were significantly lower than those of patients without cavity formation (0.70 vs 0.96 × 10−3 mm2/s; P = .01). The cutoff ADC value of 0.80 × 10−3 mm2/s resulted in 75% sensitivity and 81.8% specificity for predicting cavity formation.
CONCLUSION: Low ADC values in acute spinal cord injury may indicate postoperative cavity formation in the injured spinal cord and predict poor functional recovery.
*Department of Neurosurgery, School of Medicine, Tohoku University, Sendai, Japan; †Department of Neurosurgery, Sendai Medical Center, Sendai, Japan
Received, October 26, 2009.
Accepted, June 10, 2010.
Correspondence: Toshiki Endo, MD, PhD, Department of Neurosurgery, School of Medicine, Tohoku University, Seiryo Aoba, Sendai, Japan 980-8574. E-mail: email@example.com