Projects

Our team is working to improve the predictive accuracy of pathology for pLGG through the application of deep learning methods. Specifically, this project seeks to integrate radiological reports and image data to enhance the predictive outcome. As the project progresses, there is a possibility that it may also be expanded to improve the model’s anomaly detection ability beyond the initial stage.

This project aims to determine if Freesurfer could be an optimal segmentation method to automatically obtain choroid plexus volume from MR imaging and to better understand the normal size of the choroid plexus in a healthy pediatric population. 777 children completed MRI and obtained choroid plexus volumes by Freesurfer. 40 of them also obtained volumes via manual segmentations on 3D Slicer. The results show that Freesurfer is equivalent to manual segmentation that is accurate and more rapid to acquire choroid plexus volumes in children.

This project seeks to establish the role of ADC (Apparent diffusion coefficient) in the Differentiation of BRAF- Mutated and BRAF-Fused Tumors. We aim to develop and validate a radiomics-based ADC signature predictive of the molecular status of PLGNTs.