Traumatic hemorrhagic brain injury: impact of location and resorption on cognitive outcome.

ABSTRACT

OBJECTIVE Hemorrhagic contusions are often the most visible lesions following traumatic brain injury. However, the incidence, location, and natural history of traumatic parenchymal hemorrhage and its impact on neurological outcome have been understudied. The authors sought to examine the location and longitudinal evolution of traumatic parenchymal hemorrhage and its association with cognitive outcome. METHODS Sixteen patients with hemorrhagic contusions due to acceleration-deceleration injuries underwent MRI in the acute (mean 6.3 days postinjury) and chronic (mean 192.9 days postinjury) phases. ImageJ was used to generate GRE and FLAIR volumes. To account for the effect of head-size variability across individuals, the authors calculated each patient's total brain tissue volume using SIENAX. GRE and FLAIR volumes were normalized to the total brain tissue volume, and values for absolute and percent lesion volume and total brain volume change were generated. Spearman's rank correlations were computed to determine associations between neuroimaging and 6-month postinjury neuropsychological testing of attention (Symbol Digit Modalities Test [SDMT], oral [O] and written [W] versions), memory (Selective Reminding Test, total learning and delayed recall), and executive function (Trail Making Test Part B [TMT-B]). RESULTS The patients' mean age was 31.4 ± 14.0 years and their mean Glasgow Coma Scale score at admission was 7.9 ± 2.8. Lesions were predominantly localized to the frontal (11 lesions) and temporal (9 lesions) lobes. The average percent reductions in GRE and FLAIR volumes were 44.2% ± 46.1% and 80.5% ± 26.3%, respectively. While total brain and frontal lesion volumes did not correlate with brain atrophy, larger temporal lobe GRE and FLAIR volumes were associated with larger volumes of atrophy (GRE acute, -0.87, p < 0.01, chronic, -0.78, p < 0.01; FLAIR acute, -0.81, p < 0.01, chronic, -0.88, p < 0.01). Total percent volume change of GRE lesions correlated with TMT-B (0.53, p < 0.05) and SDMT-O (0.62, p < 0.05) scores. Frontal lobe lesion volume did not correlate with neuropsychological outcome. However, robust relationships were seen in the temporal lobe, with larger acute temporal lobe GRE volumes were associated with worse scores on both oral and written versions of the SDMT (SDMT-W, -0.85, p < 0.01; SDMT-O, -0.73, p < 0.05). Larger absolute change in temporal GRE volume was strongly associated with worse SDMT scores (SDMT-W, 0.88, p < 0.01; SDMT-O, 0.75, p < 0.05). The same relationships were also seen between temporal FLAIR lesion volumes and neuropsychological outcome. CONCLUSIONS Traumatic parenchymal hemorrhages are largely clustered in the frontal and temporal lobes, and significant residual blood products are present at 6 months postinjury, a potential source of ongoing secondary brain injury. Neuropsychological outcome is closely tied to lesion volume size, particularly in the temporal lobe, where larger GRE and FLAIR volumes are associated with more brain atrophy and worse SDMT scores. Interestingly, larger volumes of hemorrhage resorption were associated with worse SDMT and TMT-B scores, suggesting that the initial tissue damage had a lasting impact on attention and executive function.