Related Experiment Video
Updated: Feb 25, 2026
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
Association of Computed Tomography Ischemic Lesion Location With Functional Outcome in Acute Large Vessel Occlusion
Marielle Ernst1, Anna M M Boers2, Annette Aigner2
1From the Department of Diagnostic and Interventional Neuroradiology (M.E., J.F.) and Institute of Medical Biometry and Epidemiology (A.A.), University Medical Center Hamburg-Eppendorf, Germany; Departments of Radiology (A.M.M.B., O.A.B., H.A.M., C.B.L.M.M.), Biomedical Engineering and Physics (A.M.M.B., H.A.M.), and Neurology (Y.B.R.), Academic Medical Center, Amsterdam, the Netherlands; Department of Robotics and Mechatronics, University of Twente, Enschede, the Netherlands (A.M.M.B.); Division of Neurointervention, Texas Stroke Institute, Dallas-Fort Worth (A.J.Y.); Departments of Neurology (O.A.B., D.W.J.D.) and Radiology (A.v.d.L.), Erasmus MC University Medical Center, Rotterdam, the Netherlands; Departments of Neurology (R.J.v.O.) and Radiology (O.A.B., W.H.v.Z), Maastricht University Medical Center, the Netherlands; and Cardiovascular Research Institute Maastricht, the Netherlands (R.J.v.O.). m.ernst@uke.de.
Background And Purpose:
Ischemic lesion volume (ILV) assessed by follow-up noncontrast computed tomography correlates only moderately with clinical end points, such as the modified Rankin Scale (mRS). We hypothesized that the association between follow-up noncontrast computed tomography ILV and outcome as assessed with mRS 3 months after stroke is strengthened when taking the mRS relevance of the infarct location into account.
Methods:
An anatomic atlas with 66 areas was registered to the follow-up noncontrast computed tomographic images of 254 patients from the MR CLEAN trial (Multicenter Randomized Clinical Trial of Endovascular Treatment of Acute Ischemic Stroke in the Netherlands). The anatomic brain areas were divided into brain areas of high, moderate, and low mRS relevance as reported in the literature. Based on this distinction, the ILV in brain areas of high, moderate, and low mRS relevance was assessed for each patient. Binary and ordinal logistic regression analyses with and without adjustment for known confounders were performed to assess the association between the ILVs of different mRS relevance and outcome.
Results:
The odds for a worse outcome (higher mRS) were markedly higher given an increase of ILV in brain areas of high mRS relevance (odds ratio, 1.42; 95% confidence interval, 1.31-1.55 per 10 mL) compared with an increase in total ILV (odds ratios, 1.16; 95% confidence interval, 1.12-1.19 per 10 mL). Regression models using ILV in brain areas of high mRS relevance instead of total ILV showed a higher quality.
Conclusions:
The association between follow-up noncontrast computed tomography ILV and outcome as assessed with mRS 3 months after stroke is strengthened by accounting for the mRS relevance of the affected brain areas. Future prediction models should account for the ILV in brain areas of high mRS relevance.
