Quantitative MRSI for the prediction of response to chemoradiation therapy in GBM

Emory University, Atlanta GA

Investigators

Linked publications, trials & patents

Abstract

DESCRIPTION (provided by applicant): The current standard for monitoring glioblastoma multiforme (GBM) progression relies heavily on the changes of enhancing tumor regions on contrast-enhanced, T1-weighted MRI using the Macdonald Criteria (MC). However, the inherent heterogeneity of GBM along with the volume and physiological changes associated with surgical resection make this two-dimensional assessment strategy problematic. Moreover, the addition of temozolomide (TMZ) to radiation therapy significantly increases the incidence of pseudo progression, further complicating response determination. It has become apparent that in the present era of molecularly-targeted therapy, when cytotoxicity may not be the primary therapeutic effect, contrast- enhanced MRI is inadequate for monitoring response and progression. More sophisticated and reliable techniques are desperately needed. Proton magnetic resonance spectroscopic imaging (1H-MRSI) is a promising technique that offers a non-invasive means to differentiate tumor progression from post-treatment changes based on the unique magnetic properties of molecular species within tissues. Furthermore, 1H-MRSI can be used to monitor response to therapies that cause widespread metabolic alterations -like suberoylanilide hydroxamic acid (SAHA), a histone deacetylase inhibitor -if 1H-MRS biomarkers are quantified and standardized. As such, this work proposes to implement state-of-the-art 1H-MRSI technology to generate three-dimensional metabolite maps of the entire brain that can be co-registered with other imaging studies in a clinically useful fashion. By analyzing numerous data sets from clinical trials at Emory University processed in this fashion, 1) a segmentation algorithm for determining tumor volume and intracellular signal density will be developed, and 2) the spatial coherence of 1H-MRS-visible metabolites with standard MR image volumes will be determined. These standardized algorithms will then be used 3) to produce response vectors that will be tested against currently used response criteria and observed clinical outcomes to evaluate their effectiveness in determining tumor response to TMZ, radiation, and SAHA therapy. The intended outcome of this work is to make 1H-MRSI and the associated bioinformatic techniques practical for clinical use and a part of the standard assessment of brain tumor patients. Lastly, the proposed project will serve as a framework for the applicant's training plan, which is specifically designed to integrate basic analytical techniques and tumor imaging technology to assist the applicant in achieving the career goal of becoming a physician-scientist with a focus in image-guided therapy planning for malignant disease.