Saumya Gurbani

Biomedical Engineering Ph.D. Thesis Defense

Date: Tuesday, March 12, 2019

Time:  3-4pm

Location: Kauffman Auditorium, 5th floor Winship Cancer Institute, Emory University

Advisors:        

Hyunsuk Shim, PhD 

 Lee Cooper, PhD

Committee Members:

Melissa Kemp, PhD

Peng Qiu, PhD

Hui-Kuo Shu, MD, PhD

David Yu, MD, PhD

Title:                      Machine Learning Enables the use of Spectroscopic MRI to Guide Radiation Therapy in Patients with Glioblastoma

Abstract:             

Glioblastoma is the most common adult primary brain tumor and is highly aggressive due to its diffusely infiltrative nature - median survival is just 15 months despite resection, radiation, and chemotherapy. Radiation therapy has been shown to be the best single treatment for improving prognosis but requires accurate pre-therapy imaging for proper radiation dosage planning. Proton spectroscopic magnetic resonance imaging (sMRI) is an advanced imaging modality that measures specific in vivo metabolite levels within the brain and has shown to be highly sensitive and specific in the detection of proliferative pathology. Clinical application of sMRI has been extremely limited due to computational challenges in sMRI data analysis. Current analysis pipelines require skilled user intervention at multiple points, there are no consistent models for spectral quality assessment and artifact removal, and existing computational techniques are inadequate for reliable tumor segmentation. In this work, we utilize novel machine learning architectures to develop a software framework to close the gap for clinical utilization of sMRI in radiation therapy planning. First, we develop convolutional neural network that is able to identify and remove spectral artifacts that lead to erroneous measurement. Next, we develop an algorithm for internally normalizing sMRI volumes, enabling voxel-to-voxel comparison across subjects and allowing threshold-based techniques to be used for target delineation. Third, we create a novel unsupervised learning framework to perform accelerated spectral quantitation, reducing the computational time and power needed to utilize sMRI. Finally, we develop a web-based software framework that bridges the gap between sMRI and its clinical use, and demonstrate the feasibility of using this software in a multi-site clinical study to guide a radiation boost to regions of metabolic abnormality in patients with glioblastoma.