Performance of five automated white matter hyperintensity segmentation methods in a multicenter dataset

• Published in: Scientific reports 2019-11, Vol.9 (1), p.16742-12, Article 16742
• Main Authors: Heinen, Rutger, Steenwijk, Martijn D., Barkhof, Frederik, Biesbroek, J. Matthijs, van der Flier, Wiesje M., Kuijf, Hugo J., Prins, Niels D., Vrenken, Hugo, Biessels, Geert Jan, de Bresser, Jeroen
• Format: Article
• Language: English
• Subjects: 59/57; 692/617/375/1370/534; 692/617/375/534; Aged; Algorithms; Automation; Automation, Laboratory; Correlation coefficient; Female; Humanities and Social Sciences; Humans; Image Interpretation, Computer-Assisted - methods; Image processing; Lesions; Magnetic Resonance Imaging; Male; Middle Aged; Multicenter Studies as Topic; multidisciplinary; NMR; Nuclear magnetic resonance; Scanners; Science; Science (multidisciplinary); Segmentation; Substantia alba; Toads; Vascular diseases; White Matter - diagnostic imaging
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-019-52966-0

White matter hyperintensities (WMHs) are a common manifestation of cerebral small vessel disease, that is increasingly studied with large, pooled multicenter datasets. This data pooling increases statistical power, but poses challenges for automated WMH segmentation. Although there is extensive literature on the evaluation of automated WMH segmentation methods, such evaluations in a multicenter setting are lacking. We performed WMH segmentations in sixty patients scanned on six different magnetic resonance imaging (MRI) scanners (10 patients per scanner) using five freely available and fully-automated WMH segmentation methods (Cascade, kNN-TTP, Lesion-TOADS, LST-LGA and LST-LPA). Different MRI scanner vendors and field strengths were included. We compared these automated WMH segmentations with manual WMH segmentations as a reference. Performance of each method both within and across scanners was assessed using spatial and volumetric correspondence with the reference segmentations by Dice’s similarity coefficient (DSC) and intra-class correlation coefficient (ICC) respectively. We found the best performance, both within and across scanners, for kNN-TTP, followed by LST-LPA and LST-LGA, with worse performance for Lesion-TOADS and Cascade. Our findings can serve as a guide for choosing a method and also highlight the importance to further improve and evaluate consistency of methods in a multicenter setting.