HDO Imaging as a Quantitative Marker of Cerebral Glucose Oxidation

HDO Imaging as a Quantitative Marker of Cerebral Glucose Oxidation HDO Imaging as a Quantitative Marker of Cerebral Glucose Oxidation Metabolic imaging of the brain has found numerous applications ranging from the neurosciences to cancer detection. Upregulated glucose uptake in cancer provides the mechanistic underpinning for the success of fluoro-deoxyglucose-positron emission tomography (FDG-PET) as one of the most widely used metabolic imaging methods. Magnetic resonance (MR) based methods are not as easily generalized for metabolic imaging of the brain, though recent results using Deuterium MR Imaging (DMRI) demonstrate that glucose can serve as a metabolic contrast agent for glycolysis and tricarboxylic acid (TCA) cycle activity. This new technique faces manifold challenges in the form of limited signal-to-noise ratio (SNR) for 2H detected experiments, and in the physiological implications of the large dose of glucose needed to produce the images. Essential steps for developing deuterated glucose as an imaging agent must address both SNR issues and the complex metabolism associated with an administered glucose load, which is similar to a glucose tolerance test. Recently, the PI demonstrated that simple gradient-echo methods detecting HDO generated from metabolism of [2H7]glucose can report on cerebral metabolic turnover. We found that a pseudo-steady state is reached within 10 minutes of glucose injection where HDO production is linearly correlated with glx and lactate production. After this initial period, cerebral HDO levels continue to increase compared to glx, most likely because of inflowing HDO produced by extra-cerebral tissues. If this is the case, diffusion weighting should be a simple paradigm for recovering intracellular HDO generation, and hence metabolic activity at extended times after perdeuterated glucose administration. These results suggest HDO may be a more viable imaging target than carbon containing metabolites, but the metabolic complexity of the experiment must be more formally addressed and MR methods optimized for HDO detection. We will use compressed sensing, a new, but extensively validated approach for reducing the number of samples needed in k-space thereby significantly shortening total experimental time. To validate the new method, we will use standard 13C MR based approaches for estimating total TCA cycle turnover for comparison to the HDO method. Furthermore, we will compare our new paradigm to the FDG-PET approach directly at two sites within our multi-PI research team as part of a 2nd validation step. Finally, we will also use a pediatric cancer model to determine the suitability of the HDO detection approach compared to more standard lactate detection strategies. Hypothesis. HDO production after [2H7]glucose administration faithfully reports on metabolic turnover. Aim 1. Develop methods for accelerating, and increasing the specificity of, 2H imaging. Goal: Establish a method that produces the highest HDO signal-to-noise ratio efficiency (SNRE) with suitable in plane imaging resolution and explore methods for localizing signal to intracellular spaces. CSI has traditionally been used for spectroscopic imaging, but it is slow relative to multi-echo chemical shift encoding methods. We will use compressed sensing-based pseudo random under-sampling and iterative reconstruction to accelerate both CSI and multi-echo methods. Results achieved with 2H diffusion weighted imaging (DWI) will be compared to imaging at early and late (> 1 hour) times after glucose injection. If diffusion weighting produces a linear correlation between HDO production and/or glucose consumption/glx production, future imaging paradigms would be simplified as preconditions associated with timing of imaging after glucose administration would be relaxed. Aim 2. Confirm the HDO based metric of cerebral metabolic turnover agrees with gold standard 13C based spectroscopic approaches. Goal: Fully assess the viability of glycolytically derived HDO as a metric of CMRglu. Sprague-Dawley rats will be imaged using both 2H and 13C tracers. Peripheral production of HDO will be assessed. Tracer dosages will be carefully calibrated to produce metabolically equivalent estimates of glucose consumption. Data will be modeled to produce quantitative estimates of glucose oxidation. The inhibitor 2-deoxyglucose will serve as a negative control for HDO production. Aim 3. Compare HDO imaging methods to FDG-PET to identify possibilities for synergistic applications. Goal: Establish strengths and weaknesses of 2H imaging versus gold standard FDG-PET. FDG-PET is extremely sensitive, but only estimates glucose uptake. We will compare achievable in plane resolutions and the relative contrasts available via both methods in a pediatric cancer model. We will determine if a hybrid 2H/PET approach is a suitable target for future research. We will also study the effects of 2-deoxyglucose treatment of the tumors as a model of chemotherapeutic intervention. Impact. At the conclusion of this project, a wholly new in vivo method for measuring glucose utilization via MR will be established. Such a technique would find wide application across brain health as a whole.