Automated segmentation of pediatric brain tumors based on multi-parametric MRI and deep learning
In person: 23 February 2022 • 5:30 PM - 7:00 PM PST
Tumor segmentation is essential for surgical and treatment planning and radiomics studies, but manual segmentation is time-consuming and has high interoperator variability. This study presents a deep learning-based method for automated segmentation of pediatric brain tumors based on multi-parametric MRI scans (T1, T1w-Gd, T2, and FLAIR). DeepMedic, three-dimensional convolutional neural network, was trained on a training set (n=67), and then it was evaluated on an independent test set (n=30). The model displayed strong performance on segmentation of the whole tumor region (mean+/-SD Dice was 0.82+/-0.18), indicating that it can facilitate detection of abnormal region for further clinical measurements.
Univ. of Pennsylvania (United States), Children’s Hospital of Philadelphia (United States)
Rachel Madhogarhia is a Masters student in Bioengineering at the University of Pennsylvania, where she earned her Bachelor of Science in Engineering in Bioengineering with a double minor in Mathematics and Philosophy. Before joining the research team at Penn, she was a software development engineer at Amazon Alexa. The goal of Rachel’s research is to create an auto-segmentation module for pediatric brain tumors that can be implemented into standard, everyday clinical practice across institutions and into a pipeline to help streamline and improve future radiomics studies.