Unsupervised Abnormality Detection In Neonatal Mri Brain Scans Using Deep Learning

This article presents a framework designed to support the analysis and assessment of neonatal MRI brain scans, which can be used as an aid to neuroradiologists. The framework was tested on the developing human connectome project (dHCP) dataset with 97 patients that were previously categorized by severity. Results showed that the model was able to identify and distinguish subtle morphological signatures present in brain structures, correctly categorizing them in up to 83% of cases. Additionally, new brain anomalies originally missed during the radiological reading were identified and corroborated by a neuroradiologist. This framework and modeling approach demonstrate an ability to improve prognostication of neonatal brain conditions and are able to localize new anomalies.