INTRODUCTION

The brain has high Adenosine Triphosphate(ATP) demand, even in the resting state. Oxidative phosphorylation of glucose provides the great majority(~87%) of brain ATP.¹ However, low levels of non-oxidative glucose metabolism (glycolysis) do occur, even in the presence of adequate blood oxygen, a phenomenon termed “aerobic glycolysis”.² PET studies using ¹⁵O and ¹⁸F- fluorodeoxyglucose (¹⁸FDG) to measure cerebral metabolic rate of oxygen (CMRO₂) and glucose (CMRGlu) have reported that relative rates of oxidative and non-oxidative glucose consumption are not uniform through the brain ². Aerobic glycolysis rate is increased in default mode network (DMN) regions, which also exhibit age-related atrophy, amyloid deposition and decreased CMRGlu. Aerobic glycolysis rate is decreased in the thalamus, insular cortex and cerebellum³. Aerobic glycolysis rates decrease with age both regionally and whole-brain ⁴. Many tumors exhibit elevated aerobic glycolysis, termed the Warburg effect ⁵. Collectively, these observations argue that aerobic glycolysis may provide a quantitative index of brain pathophysiology, both in age-related disorders and in malignancies. ¹⁵O PET measurements of CMRO₂ require an on-site cyclotron for production of ¹⁵O (T_½ = 123 sec) and are logistical challenging, limiting its availability. To address this limitation, T2-relaxation-under-spin-tagging (TRUST) MRI, together with phase-contrast MRI, can be an alternative means for measuring whole-brain CMRO₂ and, in combination with ¹⁸FDG PET (T_½ = 110 min), of measuring aerobic glycolysis. The goal of this study was to test this TRUST-phase contrast-quantitative FDG strategy, via within-subject quantitative comparisons.

METHODS

Participants
Seven healthy volunteers (4M/3F; 27.5±4.4years fasting state) underwent sequential MRI-CMRO2, PET-CMRO2 and PET-CMRGlu measurements on the same morning. Hematocrit, pulse oximetry, and fasting plasma glucose level were measured right before the scans.
MR Measurement
MRI was performed on a 3T Siemens Tim Trio with a 12-channel head coil. MRI measurements were made of: brain volume/weight (T1w MPRAGE), whole-brain blood flow (phase-contrast MRI), and venous oxygenation (TRUST), and were used to compute whole-brain MRI-CMRO₂ ⁶.
PET Measurements
PET-CMRO2 measurements used the classical three-tracer (O¹⁵O, C¹⁵O, H₂¹⁵O) method on an ECAT HR+ scanner, with the arterial input function (AIF) sampled from a radial artery catheter ⁷. ¹⁸FDG PET has performed dynamically(1 min frame up to 70 minutes), also with a radial-artery AIF. PET images were reconstructed with filtered back projection (¹⁵O) or maximum-likelihood estimation (¹⁸FDG) with 5 mm FWHM Gaussian smoothing. PET-CMRO₂ and PET-CMRGlu images were computed in native space prior to spatial normalization (MNI space) for overlay. Oxygen-Glucose Index (OGI) was calculated as CMRO₂/CMRGlu for both PET-CMRO₂ and MRI-CMRO₂ ⁸. PET/PET OGI was computed in both voxel-wise and whole-brain manners. MRI/PET OGI was computed for whole-brain only, as an illustration for the feasibility of T2 based MR oxygenation methods for cerebral aerobic glycolysis measurements.

RESULTS

The PET/PET voxel-wise computation of OGI (Figure 1) replicated the previously reported spatial distribution ², showing higher aerobic glycolysis (low OGI) in cingulate, parietal and frontal cortical regions and lower aerobic glycolysis (high OGI) in cerebellum, thalamus and insula. Whole-brain values for ¹⁸F PET CMRGlu, ¹⁵O PET CMRO₂, TRUST MRI - CMRO₂, PET/PET OGI, and MRI/PET OGI are shown in Figure 2. Values are within established ranges⁸, with the exception of MRI/PET OGI in subject 7. PET/PET OGI and MRI/PET OGI are relatively correlated (r = 0.67, bland-altman Bias and agreement analysis bias 0.76±0.98), shown in Figure 3.

DISCUSSION

These preliminary results suggest that accurate, quantitative, whole-brain measurements of OGI can be made using the combination of TRUST MRI and ¹⁸FDG PET.
There are two substantial advantages of the proposed method for assessing glucose pathway preference, relative to the traditional ¹⁵O PET/¹⁸F PET method: availability and convenience. As regards availability, MRI and ¹⁸F PET are widely available; ¹⁵O PET has limited availability. As regards convenience, ¹⁵O PET requires arterial catheterization (for AIF) and rapid (q 2-4 sec) sampling; TRUST MRI is entirely non-invasive. There is one substantial limitation of the MRI/PET approach: lack of resolution. TRUST MRI is a whole-brain measurement, which may limit its clinical utility that largely relies on demonstrating changes in regional patterns of aerobic glycolysis. PET/PET measurements, by contrast, allow voxel-wise computation of OGI and related metrics quantifying aerobic glycolysis.

Acknowledgements

No acknowledgement found.