Development of multinuclear MRI for image guided therapy of glioma patients

点击翻译按钮获取中文摘要

7. Project Summary/Abstract Gliomas have a low survival rate of 36% at five-years, half the average survival rate across all cancers according to the last SEER and CBTRUS reports. Malignant brain tumors cause on average 20 years of potential life lost (YPLL), exceeding most cancers. Survival and YPLL have not improved for gliomas as in other cancers and urgent progress is needed. Recent studies showed maximal surgical resection extends survival of glioma patients. However, glioma margins are ambiguous on conventional anatomical imaging, and neurosurgeons are in great need for more specific imaging to map tumor boundary. Gliomas have large metabolic alterations, most notably D-2-hydroxyglutarate (2HG), cystathionine and nicotinamide adenine dinucleotide in isocitrate dehydrogenase (IDH) mutations. Imaging metabolism using proton (1H) and phosphorus (31P) magnetic resonance spectroscopic imaging (MRSI) has high specificity for glioma. Mapping tumor margins requires high resolution, which is challenging for metabolic imaging. Our project addresses neurosurgeons’ need for high resolution metabolic imaging. Recent discoveries highlighted metabolism role in modulating epigenetic factors, gene expression and immune system responsible for tumor progression. Hence, new treatment strategies target metabolic pathways to increase the effect of chemoradiation and immunotherapy to prevent tumor resistance and relapse. Metabolic imaging can accelerate clinical trials of novel therapies, and its development has been mandated by the last neuro-oncology consensus (SNO, EANO). The long-term goal of our research is to create non-invasive metabolic cancer imaging tools for personalized precision oncology. The objective of this application is to develop fast high-resolution whole-brain quantitative metabolic multinuclear MR imaging for image guided therapy of glioma patients. The central hypothesis of our proposal is that multinuclear MR imaging will enable clinicians deliver precision oncology treatments with improved outcomes tailored to individual glioma patients. We will perform three specific aims: 1) develop 1H-MRSI for intra-operative image guided surgical removal of gliomas, 2) develop 31P-MRSI at ultra-high field to guide pharmacologic interventions in glioma, 3) clinical translation of multinuclear MRI for treatment of glioma patients. There are strong rationales for the proposed research: 1) there is no alternative in vivo imaging that is specific for IDH mutations, 2) MRSI is non-invasive and safe to scan without radiation and tracer, 3) fast, 4) cost effective, 5) probes entire tumor and healthy brain without sampling bias of biopsies, 6) can be performed pre-surgically and in inoperable patients. The approach is innovative because it employs the first available whole-brain 2HG imaging at high resolution accelerated by deep learning compressed sensing, novel receive-shim array hardware for 31P-MRSI to improve data quality, and automated high-throughput processing by deep learning. The contribution of the proposed research will be significant because it will provide clinicians a validated precision oncology imaging tool to select optimal treatment in glioma patients.