Meeting Banner
Abstract #5523

ASL meets Machine Learning: Classification of type-2 diabetes and normoglycemia using ASL-derived cerebral hemodynamic features.

Maria-Eleni Dounavi1,2, Christopher Martin2,3, Dinesh Selvarajah4, Aneurin J. Kennerley2,3, Solomon Tesfaye4, Eleni Vasilaki5, and Iain D. Wilkinson1,2

1Academic Unit of Radiology, University of Sheffield, Sheffield, United Kingdom, 2Neuroimaging in Cardiovascular Disease (NICAD) Network, University of Sheffield, Sheffield, United Kingdom, 3Psychology Department, University of Sheffield, Sheffield, United Kingdom, 4Academic Unit of Diabetes and Endocrinology, University of Sheffield, Sheffield, United Kingdom, 5Department of Computer Science, University of Sheffield, Sheffield, United Kingdom

QUASAR ASL is an arterial transit time insensitive perfusion imaging technique which can be used to unravel hemodynamic patterns. This study evaluates cerebral perfusion hemodynamics using QUASAR in patients with type-2 diabetes mellitus (T2DM) and normoglycemic controls. In addition to standard perfusion parameters, multiple metrics were extracted from five QUASAR-derived curves pre and post acetazolamide injection both globally and locally, from regions adjacent to major vascular territories. Following feature reduction, a binary classification task was performed (normoglycemia vs. T2DM). Necessary steps were undertaken to reassure that the observed results were not due to overfitting. The achieved classification accuracy was 95%.

This abstract and the presentation materials are available to members only; a login is required.

Join Here