Background and purpose: The diagnosis of external borderzone infarction is made when the stroke is located at the border between the arterial territories. Recent studies have raised questions regarding the location of this borderzone given the variability in the arterial territories. We examined the location of this region using a digital approach and its correspondence with the 'traditional' template.
Methods: Infarcts resulting from occlusion of the middle cerebral artery (MCA) or posterior cerebral artery (PCA) trunk or branches were segmented from T(2)-weighted MR images and linearly registered into a common stereotaxic coordinate space. For MCA infarcts and PCA infarcts, maps of voxels in a rim surrounding the infarct were created. The maps of individual rims were averaged to create images of the probability of each voxel lying in the MCA and PCA rims. The MCA and PCA rims were used to create a digital atlas of the probability of each voxel lying concurrently in both rims.
Results: The MCA group consisted of 36 patients (16 males) with a median age of 73 (range 25-87) years. The PCA territory group consisted of 30 patients (24 males) with a median age of 61 (range 22-86) years. The probability of involvement in the digital atlas was higher in the posterior putamen (probability 0.12-0.29) and optic tract (probability 0.13-0.0.20) than in the angular gyrus (probability 0.01).
Conclusion: We have created a digital model of the border region between the MCA and PCA territories. This approach may be useful for evaluating the likelihood of a stroke mechanism from topography.