This study aims to diagnose objectively term neonate hypoxic ischemic encephalopathy (HIE) by using deep learning network to extract deep information from multiple modalities of conventional magnetic resonance (MR) images. Neonate HIE diagnosis accuracy is restricted to lesion diversity, MR images quality, high interobserver variability. The network got high diagnosis accuracy in the ROC curve. The network could detect severe neonate HIE with characteristic appearance such as basal ganglia injury and periventricular leukomalacia. The network can help diagnosis neonate HIE objectively without the effect of different radiologist experience and contribute to risk stratification and clinical decision making.
This abstract and the presentation materials are available to members only; a login is required.