Abstract #1976

Improving CBF Image Contrast with Frequency Extrapolation for DSC-MRI During Acute Stroke

Matthew Ethan MacDonald^1,2, Micheal Richard Smith^1,3, Richard Frayne^2,3

¹Departments of Electrical & Biomedical Engineering, University of Calgary, Calgary, AB, Canada; ²Seaman Family MR Research Centre, Foothills Medical Centre, Calgary, AB, Canada; ³Departments of Radiology & Clinical Neurosciences, University of Calgary, Calgary, AB, Canada

DSC-MRI is can provide estimates of the tissue perfusion in the early onset of stroke. CBF is a perfusion parameter that has been suggested for detecting regions of tissue that are affected by ischemia and may still be salvageable with treatment. The calculation of CBF from raw perfusion data requires deconvolution, which has been shown to have MTT-dependent bias. In this work, frequency domain extrapolation is introduced during the deconvolution process to correct for this bias. The algorithm is tested across ten ischemic stroke patient data sets, and compared with follow up imaging, ROC analysis is used to determine detection.