Edema-corrected corticospinal tract integrity improves outcome prediction after minimally invasive intracerebral hemorrhage evacuation

Background As minimally invasive surgical evacuation of intracerebral hemorrhage (ICH) becomes increasingly adopted, long-term functional recovery remains heterogeneous and poorly explained. Injury to the corticospinal tract (CST) is central to motor recovery and can be quantified using diffusion tensor imaging (DTI). Because perihematomal edema and mechanical compression can confound DTI in ICH, we hypothesized that application of an edema-correction algorithm to postoperative imaging would reveal prognostic associations between CST injury and long-term outcome. Methods Patients undergoing minimally invasive surgery for ICH between 2017 and 2024 with postoperative diffusion MRI were retrospectively analyzed at a single center. CST injury was quantified using ipsilesional-to-contralesional fractional anisotropy ratios with free-water correction (rFA-wc) and without, derived from an automated tractography pipeline. Outcomes included 90-day modified Rankin Scale (mRS) and the National Institutes of Health Stroke Scale (NIHSS) motor subscores. Associations were assessed using Spearman correlations and multivariable proportional-odds regression. Results 60 patients were included, 52 with available 90-day NIHSS motor scores. Lower rFA-wc (decreased CST integrity) correlated with worse mRS (P=0.043) and worse NIHSS motor subscore (P=0.014) and independently increased the odds of worse functional outcome at 90 days (OR 1.41, 95% CI 1.02 to 2.02) and motor deficit (OR 1.78, 95% CI 1.21 to 2.78). Uncorrected metrics showed weaker and less consistent associations with outcome. Strongest associations occurred when postoperative MRI was obtained within 1 week, and incorporation of rFA-wc improved outcome prediction beyond the ICH score. Conclusion Edema-corrected assessment of postoperative CST integrity reveals a clinically meaningful prognostic signal after minimally invasive ICH evacuation.