Improved image guidance is needed for neurosurgeons to reduce the residual remaining clot levels during minimally invasive evacuation of intracerebral hemorrhage (ICH) while not causing rebleeds. Neurosurgeons would benefit from a means to periodically render the clot volume against surrounding normal tissue during mechanical evacuation or pharmaceutical-based clot-busting. Using convolutional neural networks (CNN), we created machine learning models to automatically segment the constituent clot and edema induced by ICH cases using T2-weighted MR images. The CCN’s output results were found to be in agreement with manual segmentations of the same ICH cases.
This abstract and the presentation materials are available to members only; a login is required.