41 White matter hyperintensity shape and long-term progression of cerebrovascular disease 3 imaging (GRE-EPI) (repetition time = 3050 ms; time to echo = 50 ms; field of view: 220 mm; matrix = 256 x 256); a proton density/T2-weighted fast spin-echo (repetition time = 3220 ms; time to first echo = 22 ms; time to second echo = 90 ms; echo train length = 8; field of view = 220 mm; matrix = 256 × 256). The FLAIR, T2*, and proton density/T2 scans were acquired with 3 mm thick interleaved slices with voxel sizes of 0.86 mm x 0.86 mm x 3.00 mm. 3.3.3 WMH shape WMH shape markers were calculated using an in-house developed pipeline, as described in detail previously.4 In short, WMH were segmented automatically on the registered FLAIR images using the LST toolbox7 in SPM12. Lateral ventricles were segmented from the T1 scans and the ventricle masks were inflated with 3 and 10 mm. The inflated ventricle masks aided WMH classification into periventricular/ confluent, and deep WMH (figure 3.2). Periventricular and confluent WMH were merged into one category, because of spatial overlap in these lesion types preventing separated shape analyses. WMH need to be at least 3 mm distant from the ventricular wall to be classified as deep WMH. WMH shape markers were calculated based on the WMH segmentations. Convexity, solidity, concavity index, and fractal dimensions were determined for periventricular/confluent WMHs.8 A lower convexity and solidity, and higher concavity index and fractal dimension indicate more irregularly shaped periventricular/confluent WMH. Fractal dimensions and eccentricity were calculated for deep WMH.8 A higher eccentricity indicates a more elongated shape, while a higher fractal dimensions indicates a more irregular shape of deep WMH. The formulas used to calculate the WMH shape markers are shown in figure 3.2. Average shape markers were calculated per participant. Each WMH shape marker captures a slightly different variation of WMH shape. In supplementary figure S.3.9.1 examples of shapes with high or low values of different shape markers are shown. 3.3.4 Other cerebrovascular markers Gray matter, white matter, cerebrospinal fluid and WMH volume were segmented automatically with a modified algorithm based on the Montreal Neurological Institute pipeline.9 Intracranial volume resulted from the addition of the volumes of gray matter, white matter, cerebrospinal fluid and WMH.10 The same processing-pipeline was used at baseline and follow-up. The WMH volume change was calculated by substracting the baseline WMH volumes from the follow-up WMH volumes.
RkJQdWJsaXNoZXIy MTk4NDMw