Abstract #2828

Machine Learning Approach Assisted MRI Characterization for Diagnosis of Neonatal Bilirubin Encephalopathy

Zhou Liu^1,2, Bing Ji², Ge Cui³, Ling Ding⁴, Xiaofeng Yang³, Liya Wang^2,4, and Hui Mao²

¹Graduate School, Medical College of Nanchang University, Nanchang, China, ²Radiology and Imaging Sciences, Emory University School of Medicine, Atlanta, GA, United States, ³Radiation Oncology, Winship Cancer Institute, Emory University, Atlanta, GA, United States, ⁴Radiology, People's Hospital of Longhua, Shenzhen, China

Diagnosis of acute bilirubin encephalopathy (ABE) in newborns based on T1-weighted spin echo images in clinical routine is challenging due to subtle signal intensity change in the basal ganglia caused by ABE often overlapping with the presence of signal enhancement from the normal myelination in the infant brain. We used a feature-extraction-based machine learning approach to identify ABE associated morphological features from T1-weighted images of 34 ABE neonates. We found textural features along with intensity-based features added specific information for distinguishing ABE from normal myelination, demonstrating the feasibility of using this approach to assist MRI diagnosis of neonatal ABE.

This abstract and the presentation materials are available to members only; a login is required.

Join Here