Professor McMillan was a co-author on a study published in the journal “Magnetic Resonance Imaging” entitled “Neural network for autonomous segmentation and volumetric assessment of clot and edema in acute and subacute intracerebral hemorrhages.” A link to the full paper is available here: https://doi.org/10.1016/j.mri.2023.07.015
Co-authors of the study included: Dr. Azam Ahmed, Associate Professor of Neurological Surgery and Radiology; Walter Block, PhD, Professor of Biomedical Engineering, Medical Physics & Radiology; Matthew Henningsen, MS, Department of Electrical & Computer Engineering; Matthew Larson, MD, PhD, fellow, Department of Radiology; Thomas Lilieholm, PhD student, Department of Medical Physics
The study focused on enhancing minimally-invasive surgical techniques for intracerebral hemorrhage (ICH) evacuation by using a Convolutional Neural Network (CNN) for autonomously segmenting clots and peripheral edema in MRI brain images, to assist in estimating the remaining clot volume. The study used a retrospective dataset of ICH patient scans from 3 T MRI scanners. Accuracy was assessed through comparisons with manual segmentations, using the Dice coefficient (DC). Results showed the model effectively segmented clot core and edema, though it slightly underestimated clot volumes by about 17%. The study underscores the potential of this CNN in guiding surgical interventions for ICH.
Of note, University of Wisconsin Football player Matthew Henningsen, M.S., was involved in the study. This garnered additional news coverage: here and here.