• Barbara A Holshouser, Karen A Tong, Stephen Ashwal, Udochukwu Oyoyo, Mark Ghamsary, David Saunders, and Lori Shutter.
• Department of Radiology, Loma Linda University School of Medicine, Loma Linda, California, USA. bholshouser@ahs.llumc.edu
• J Magn Reson Imaging. 2006 Jul 1; 24 (1): 33-40.

PurposeTo investigate whether longitudinal magnetic resonance proton spectroscopic imaging (MRSI) demonstrates regional metabolite abnormalities after traumatic brain injury (TBI) that predict long-term neurologic outcome.Materials And MethodsTwo-dimensional-MRSI (point resolved spectroscopy sequence [PRESS]; TR/TE = 3000/144 msec; 10 mm) was acquired prospectively in 42 adults with severe TBI through the level of the corpus callosum 7 +/- 4 days after injury. Measurements were repeated in 31 patients six to 12 months after injury. Regional and pooled (all regions combined) mean ratios were compared with control values and then used to predict long-term (six- to -12-month) neurologic outcome (good vs. poor) using a logistic regression model.ResultsInitial pooled mean N-acetylaspartate (NAA) ratios were lower (P < 0.01) and choline (Cho)/creatine (Cr) ratios higher (P < 0.01) in all TBI patients compared to controls. Ratios from the corpus callosum region were affected most and predicted long-term dichotomized outcome with 83% accuracy. When repeated at six to 12 months after injury, pooled mean NAA/Cr remained lower (P = 0.03) and Cho/Cr remained higher (P = 0.01) in patients with poor outcomes.ConclusionThe NAA/Cr ratio from the corpus callosum was most useful for outcome prediction. Chronic alterations of metabolite ratios are likely due to neuronal loss and glial proliferation long after injury.(c) 2006 Wiley-Liss, Inc.

21 keywords...

hide…