A five years longitudinal magnetic resonance imaging study of severe traumatic brain injury patients. Correlation with functional outcome
Résumé
Introduction
Traumatic brain injury (TBI) doesn't seem to be a single insult with a
monophasic resolution. Recently, degenerative mechanisms have
been suggested to occur in the chronic phase and could constituted
"tertiary" lesions [1]. These degenerative phenomena can potentially
have a worsening impact on the long-term functional prognosis.
Objectives The objective of this prospective study was to longitudinally evaluate (1) white and grey matter structures volumes measured from Tl three-dimensional (30) and (2) white matter integrity assessed from diffusion tensor imaging (DTI) in severe TBI.
Method 20 severe TBI (37 ± 16 yrs) and 12 healthy volunteers (HV; 42 ± 6 yrs) underwent multimodal magnetic resonance imaging in the subacute phase (within 21 ± 8 days after injury). A longitudinal follow-up was obtained for all of them at the chronic phase of injury (median 64 ± 16 months after injury) together with neuropsychological assessments. Longitudinal imaging changes were assessed using cortical volumetric reconstruction and segmentation of white and deep grey matter structures with Freesurfer [2]; cortical sulci were automatically reconstructed and identified with Brainvisa software, and a voxelbased DTI analysis was performed with Comasoft. The Extended GOS (GOSE) was used to classify at 5 years the TBI subjects into " good" (GOSE 6-7; n = 11) and "intermediate" (GOSE 3-5; n = 9) recovery. Cortical morphometry and fractional anisotropy (FA) derived from DTI were used with linear mixed effects models to link changes to behaviour status.
Results At baseline, there were no volumetric differences between the 3 groups (GOSE 3-5; GOSE 6-7; HV). At 5 years, patients with TBI demonstrated a significant volumetric reduction of the whole white matter (-10±4 %; P< 0.01), and of the deep grey matter structures (-13 ± 10 %; P < 0.03). In contrast, HV did not present any significant change over the same period. Specifically, direct comparisons be tween patient groups revealed that over time GOSE 3-5 showed greater atrophy than GOSE 6-7 in the parietal lobe (-5 ± 2 vs. -3 ± 5 % respectively; P < 0.001), brain stem (-12±6 vs. -6 ± 7 %; P < 0.006), corpus callosum (-1 9 ± 19 vs. -10 ± 1 5 %; P < 0.01 ), and cingulate (-7 ± 5 vs. -2 ± 6 %; P < 0.023). This was associated with higher depth mean on sulci data. Furthermore, FA was lower at the first MRI in GOSE 3-5 group in the same regions.Finally, neuropsychological score (Z-score) correlated s ignificantly with the volume loss in these anatomical regions.
Conclusions: We observed a strong correlation between neuropsychological scores and morphometric changes over time suggesting (1) occurrence of tertiary lesions and (2) that lesions location influence functional outcome. These data provide further insight into early and late pathophysiology of cognitive dysfunctions after TBI.
[no other pdf]
Objectives The objective of this prospective study was to longitudinally evaluate (1) white and grey matter structures volumes measured from Tl three-dimensional (30) and (2) white matter integrity assessed from diffusion tensor imaging (DTI) in severe TBI.
Method 20 severe TBI (37 ± 16 yrs) and 12 healthy volunteers (HV; 42 ± 6 yrs) underwent multimodal magnetic resonance imaging in the subacute phase (within 21 ± 8 days after injury). A longitudinal follow-up was obtained for all of them at the chronic phase of injury (median 64 ± 16 months after injury) together with neuropsychological assessments. Longitudinal imaging changes were assessed using cortical volumetric reconstruction and segmentation of white and deep grey matter structures with Freesurfer [2]; cortical sulci were automatically reconstructed and identified with Brainvisa software, and a voxelbased DTI analysis was performed with Comasoft. The Extended GOS (GOSE) was used to classify at 5 years the TBI subjects into " good" (GOSE 6-7; n = 11) and "intermediate" (GOSE 3-5; n = 9) recovery. Cortical morphometry and fractional anisotropy (FA) derived from DTI were used with linear mixed effects models to link changes to behaviour status.
Results At baseline, there were no volumetric differences between the 3 groups (GOSE 3-5; GOSE 6-7; HV). At 5 years, patients with TBI demonstrated a significant volumetric reduction of the whole white matter (-10±4 %; P< 0.01), and of the deep grey matter structures (-13 ± 10 %; P < 0.03). In contrast, HV did not present any significant change over the same period. Specifically, direct comparisons be tween patient groups revealed that over time GOSE 3-5 showed greater atrophy than GOSE 6-7 in the parietal lobe (-5 ± 2 vs. -3 ± 5 % respectively; P < 0.001), brain stem (-12±6 vs. -6 ± 7 %; P < 0.006), corpus callosum (-1 9 ± 19 vs. -10 ± 1 5 %; P < 0.01 ), and cingulate (-7 ± 5 vs. -2 ± 6 %; P < 0.023). This was associated with higher depth mean on sulci data. Furthermore, FA was lower at the first MRI in GOSE 3-5 group in the same regions.Finally, neuropsychological score (Z-score) correlated s ignificantly with the volume loss in these anatomical regions.
Conclusions: We observed a strong correlation between neuropsychological scores and morphometric changes over time suggesting (1) occurrence of tertiary lesions and (2) that lesions location influence functional outcome. These data provide further insight into early and late pathophysiology of cognitive dysfunctions after TBI.
[no other pdf]