Automatic identification of atherosclerosis subjects in a heterogeneous MR brain imaging data set

Carotid-artery atherosclerosis (CA) contributes significantly to overall morbidity and mortality in ischemic stroke. We propose a machine learning technique to automatically identify subjects with CA from a heterogeneous cohort of magnetic resonance brain images. The cohort includes 190 subjects with CA, white mater hyperintensites of presumed vascular origin or multiple sclerosis, as well as 211 presumed healthy subjects. We determined a set of handcrafted and convolutional discriminant features to perform this task. A support vector machine (SVM) was used to perform this four-class classification task. Our approach had an accuracy rate of 97.5% (higher than chance accuracy of 52.6% for guessing majority class), sensitivity of 96.4% and specificity of 97.9% in identifying subjects with CA, suggesting that the proposed combination of features may be used as an imaging biomarker for characterizing atherosclerotic disease on brain imaging621827CONSELHO NACIONAL DE DESENVOLVIMENTO CIENTÍFICO E TECNOLÓGICO - CNPQCOORDENAÇÃO DE APERFEIÇOAMENTO DE PESSOAL DE NÍVEL SUPERIOR - CAPES308311/2016-7PVE-88881.062158/2014-01We would like to thank various groups for the secondary use of their data sets: 1) The Canadian Atherosclerosis Imaging Network (CAIN Principal Investigator: Jean-Claude Tardif, MD; and CAIN Project 1 lead: Alan Moody, MD). CAIN was funded by the Canadian Institutes for Health Research (CIHR), the Canada Foundation for Innovation (CFI) and the Governments of Alberta, Ontario and Québec. 2) The MS Society of Canada for funding the MS image acquisition (Principal Investigator: Lenora Brown, PhD) and the Calgary MS Clinic for supporting the study. 3) The organizers of the WMH Segmentation Challenge (2017 International Conference on Medical Image Computing and Computer Assisted Intervention [27]; 4) The NC data was drawn from the Calgary Normative Study and this analysis was supported by the CIHR (MOP-333931; Principal Investigator: Richard Frayne, PhD). Mariana Bento, PhD was supported by the Hotchkiss Brain Institute (HBI) Burns Fellowship. Roberto Souza, PhD was supported by the T Chen Fong Postdoctoral Fellowship. Leticia Rittner, PhD thanks the National Council for Scientific and Technological Development (CNPq, Brazil; 308311/2016-7). Richard Frayne, PhD is supported by the Hopewell Professorship in Brain Imaging. The Calgary-Campinas collaboration was supported in part by an award from the Coordination for the Improvement of Higher Education Personnel (CAPES, Brazil), Special Visiting Professor Program (PVE-88881.062158/2014-01