Arjen Lindenholz

138 CHAPTER 6 (clinoid, supra-clinoid and terminal segments). A vessel wall lesion was defined as a visually judged focal or diffuse vessel wall thickening of > 50% compared with the neighboring vessel wall (or contralateral vessel wall in case of diffuse thickening), with or without contrast enhancement. 11 Contrast enhancement was assessed by comparing pre- and postcontrast images including co-registered subtraction images using the Elastix toolbox in MeVisLab (version 2.7, MeVis Medical Solutions, Bremen, Germany). 26 Contrast enhancement of a vessel wall lesion was defined as lesion hyperintensity approximating the signal intensity of the pituitary stalk on the postcontrast MR images on at least two consecutive images using the precontrast MR images as a reference and the coregistered subtraction images as confirmation. Contrast enhancing foci at the location where the ICA pierces the dura mater were classified as vasa vasorum and were not considered vessel wall enhancement. The intracranial vessel wall lesion burden was used as a continuous variable for the primary analysis, and was defined as the total number of vessel wall lesions for each patient. The total number of enhancing vessel wall lesions was used as a continuous variable in the secondary analysis. Parenchyma assessment The DWI and FLAIR images and T 1 -weighted MPIR-TSE images were assessed for anterior circulation parenchymal changes by an expert neuroradiologist (TDW, 30 years of experience). The reader was blinded for patient details, clinical information and vessel wall lesion assessment. The following cerebral parenchymal changes were scored: cortical infarcts and (recent) small subcortical infarcts, deep grey matter infarcts and lacunes of presumed vascular origin, cortical microinfarcts and periventricular and deep white matter hyperintensities (WMH). A recent position paper has judged all of these MR findings except for cortical infarcts (and possibly cortical microinfarcts) to be related to SVD. 27 Cortical infarcts were defined as hyperintense lesions on FLAIR imaging located in the cerebral cortex and greater than 5 mm in diameter, with or without associated tissue loss or extension into the deep white matter. When a cortical infarct was multifocal or extending into the white matter, but with similar presumed origin, the infarct was scored as a single cortical infarct. Small subcortical infarcts and lacunes of presumed vascular origin were defined as hyperintense lesions of respectively 15-20 mm and 3-15 mm in diameter on FLAIR imaging with or without a hypointense center or cavity and local tissue loss, classified and scored according to the STRIVE criteria. 27 Deep grey matter infarcts were defined as hyperintense lesions on FLAIR imaging located in the basal ganglia or thalamus greater than 3 mm in diameter (not specifically defined in the STRIVE criteria but derived from the position paper. 27 Cortical microinfarcts were defined as hyperintense lesions on FLAIR imaging in the cerebral cortex smaller than 5 mm