Background: Mortality can be predicted by intracerebral hemorrhage (ICH) volume, but motor recovery in survivors is variable. Motor impairment is likely related to the spatial relationship between the hematoma and corticospinal tract (CST). Diffusion Tensor Imaging (DTI) tractography can be used to visualize white matter tracts in three dimensions. We hypothesized that the interaction between the hematoma and CST would predict motor impairment in ICH patients.
Methods: ICH patients with small-moderate hematomas were prospectively imaged with CT and DTI within 14 days of onset. Hematoma volume was assessed on CT using planimetric techniques. Three-dimensional recreations of the ipsilateral CST and the hematoma were made for each patient. The CST was categorized by interaction with the ICH as CST: Unaffected, Displaced, Partially Severed, Completely Severed, and Splitting the ICH. Motor function was classified as 'good' (NIHSS motor subscale 0-2) or 'poor' (3-8).
Results: Thirty patients (mean age 68±13) underwent CT at a median (IQR) of 2.3 (3.5)h and DTI at 2.0 (3.6, range 0.6-13) days. Median hematoma volume was 8.2 (23) ml.
Lesion distribution was: lobar 11 (37%), basal ganglia 18 (60%), brainstem 1 (3%). CSTs were primarily Displaced (n=9) or Unaffected (8), with the remainder being Partially Severed (4), Completely Severed (5), and Splitting the ICH (4). The latter 4 (13%) patients had small (<6ml, median 2.5 [3.0] ml) basal ganglia bleeds which enfolded the intact CST.
Motor score at Day 7 was good in 50% of patients. Good outcome was seen in 8 (100%) Unaffected, 4 (44%) Displaced, 1 (25%) Partially Severed, 0 (0%) Severed and 2 (50%) Splitting the ICH patients. Logistic regression indicated that good motor score was predicted by CST category (r=2.3, p=0.016).
Conclusion: CST integrity can be maintained when enfolded by small basal ganglia bleeds. Diffusion tractography patterns may be useful for predicting motor scores in small to moderate-sized hematomas.