Quantitative proton MRS predicts outcome after traumatic brain injury

Abstract

OBJECTIVE: To determine whether proton MRS (1H-MRS) neurochemical measurements predict neuropsychological outcome of patients with traumatic brain injury (TBI). BACKGROUND: Although clinical indices and conventional imaging techniques provide critical information for TBI patient triage and acute care, none accurately predicts individual patient outcome. METHODS: The authors studied 14 patients with TBI soon after injury (45+/-21 days postinjury) and again at 6 months (172+/-43 days) and 14 age-, sex-, and education-matched control subjects. N-acetylaspartate (NAA), creatine, and choline were measured in normal-appearing occipitoparietal white and gray matter using quantitative 1H-MRS. Outcome was assessed with the Glasgow Outcome Scale (GOS) and a battery of neuropsychological tests. A composite measure of neuropsychological function was calculated from individual test z-scores probing the major functional domains commonly impaired after head trauma. RESULTS: Early NAA concentrations in gray matter predicted overall neuropsychological performance (r = 0.74, p = 0.01) and GOS (F = 11.93, p = 0.007). Other metabolite measures were not related to behavioral function at outcome. CONCLUSION: 1H-MRS provides a rapid, noninvasive tool to assess the extent of diffuse injury after head trauma, a component of injury that may be the most critical factor in evaluating resultant neuropsychological dysfunction. 1H-MRS can be added to conventional MR examinations with minimal additional time, and may prove useful in assessing injury severity, guiding patient care, and predicting patient outcome.