segcsvdWMH: A convolutional neural network-based tool for quantifying white matter hyperintensities in heterogeneous patient cohorts Link to paper on MedRxivAuthorsErin Gibson, Joel Ramirez, Lauren Abby Woods, Julie Ottoy, Stephanie Berberian, Christopher J.M. Scott, Vanessa Yhap, Fuqiang Gao, Roberto Duarte Coello, Maria Valdes Hernandez, Anthony E. Lang, Carmela M. Tartaglia, Sanjeev Kumar, Malcolm A. Binns, Robert Bartha, Sean Symons, Richard H. Swartz, Mario Masellis, Navneet Singh, Alan Moody, Bradley J. MacIntosh, Joanna M. Wardlaw, Sandra E. Black, ONDRI Investigators, ADNI, CAIN Investigators, colleagues from the Foundation Leducq Transatlantic Network of Excellence, Andrew SP Lim, Maged GoubranAbstractWhite matter hyperintensities (WMH) of presumed vascular origin are an MRI-based biomarker of cerebral small vessel disease (CSVD). WMH are associated with accelerated cognitive decline and increased risk of stroke and dementia, and are commonly observed in aging, vascular cognitive impairment, Alzheimer’s and Parkinson’s disease, and related dementias. The accurate, reliable, and rapid measurement of WMH in large-scale multi-site clinical studies with heterogeneous patient populations remains challenging. The diversity of MRI protocols and image characteristics across different studies as well as the diverse nature of WMH, in terms of their highly variable shape, size, distribution, and underlying pathology, adds additional complexity to this task. Here, we present segcsvdWMH, a novel convolutional neural network-based tool for quantifying WMH. segcsvdWMH is specifically designed for accurate and robust performance when applied to diverse clinical patient datasets. Central to the development of this tool is the curation of a large patient dataset (>700 scans) sourced from seven multi-site studies, encompassing a wide range of clinical populations, WMH burden, and imaging parameters. The performance of segcsvdWMH is evaluated against three widely used WMH segmentation tools, where we demonstrate significantly enhanced accuracy and robustness across a range of challenging conditions and datasets.Related linksLink to paper on MedRxivRoberto Duarte CoelloMaria Valdes HernandezFoundation Leducq Transatlantic Network of ExcellenceSocial media tags and titlessegcsvdWMH: A convolutional neural network-based tool for quantifying white matter hyperintensities in heterogeneous patient cohorts @EdinUniImaging @SVDResearch This article was published on 2024-12-23
