• Dong-Eog Kim, Jinseong Jang, Dawid Schellingerhout, Wi-Sun Ryu, Jong-Ho Park, Su-Kyoung Lee, Dongmin Kim, and Hee-Joon Bae.
• Korean Brain MRI Data Center & Department of Neurology, Dongguk University Ilsan Hospital, Goyang, Korea. kdongeog@duih.org.
• Sci Rep. 2019 Aug 12; 9 (1): 11708.

AbstractWe recently generated a high-resolution supratentorial vascular topographic atlas using diffusion-weighed MRI in a population of large artery infarcts. These MRI-based topographic maps are not easily applicable to CT scans, because the standard-reference-lines for axial image orientation (i.e., anterior-posterior commissure line versus orbito-meatal line, respectively) are 'not parallel' to each other. Moreover, current, widely-used CT-based vascular topographic diagrams omit demarcation of the inter-territorial border-zones. Thus, we aimed to generate a CT-specific high-resolution atlas, showing the supratentorial cerebrovascular territories and the inter-territorial border-zones in a statistically rigorous way. The diffusion-weighted MRI lesion atlas is based on 1160 patients (67.0 ± 13.3 years old, 53.7% men) with acute (<1-week) cerebral infarction due to significant (>50%) stenosis or occlusion of a single large cerebral artery: anterior, middle, or posterior cerebral artery. We developed a software package enabling the transformation of our MR-based atlas into a re-oriented CT space corresponding to the axial slice orientations used in clinical practice. Infarct volumes are individually mapped to the three vascular territories on the CT template-set, generating brain maps showing the voxelwise frequency of infarct by the affected parent vessel. We then mapped the three vascular territories collectively, generating a dataset of Certainty-Index (CI) maps to reflect the likelihood of a voxel being a member of a specific vascular territory. Border-zones could be defined by using either relative infarct frequencies or CI differences. The topographic vascular territory atlas, revised for CT, will allow for easier and more accurate delineation of arterial territories and borders on CT images.

30 keywords...

hide…