Since hypoxia makes tumors more aggressive and therapy resistant¹, characterizing blood supply and oxygen consumption may be a crucial information for therapy planning and outcome prediction. While the mapping of blood perfusion by arterial spin labeling (ASL) is an established and proven method, the mapping of oxygen consumption by MRI remains challenging. However, a new promising method: QUantitative Imaging of eXtraction of oxygen and TIssue consumption² (QUIXOTIC) may be able to allow for robust and reliable oxygen consumption mapping. Therefore, the purpose of this study was to test the ability of the QUIXOTIC method to map the tumor oxygen consumption in conjunction with ASL in glioma patients.

14 patients, 6 with high grade and 8 with low grade preoperative gliomas, were scanned at 3T. Fluid attenuated inversion recovery (FLAIR) images were used to define the tumor and gray matter control region of interest (ROI). Pseudo continuous arterial spin labeling³ (pCASL) was used to map cerebral blood flow (CBF) and the QUIXOTIC method to map the oxygen extraction fraction (OEF). After co-registering FLAIR, pCASL and QUIXOTIC images using MeVisLab (MeVis Medical Solutions AG, Bremen, Germany) maps of cerebral metabolic rate of oxygen (CMRO2) were calculated simply by $$$CMRO_2 = CBF · OEF · Hct$$$, with Hct denoting the hematocrit. The hematocrit was determined for each subject by a blood sample prior to MRI. Further post-processing and statistical analysis was done using in house written software in Matlab and R. To account for tumor heterogeneity, only voxels with the highest 10% of the measured (high-)OEF/CBF values and their corresponding (corr-)CBF/OEF values were averaged and analyzed, respectively. All subjects gave written informed consent to the study, which was approved by the local ethics committee.

Parameter maps reveal strong tumor heterogeneity with high variation in OEF, CBF and CMRO2 (Fig.1). OEF measurements may be hampered and were excluded from analysis in regions with low perfusion signal, i.e. white matter, tumor edema and necrosis, since the QUIXOTIC method needs a sufficient amount of spins flowing from the capillary bed into veins. Higher high-OEF was found for high and low grade gliomas compared to GM control with the corresponding corr-CBF similar for both glioma grades and GM control (Fig.2). High-CBF was lower for high grade gliomas compared to low grade gliomas and GM control, whereas corr-OEF was lower for both glioma grades compared to GM control (Fig.3). All parameters were reasonably stable for GM control and showed lower variability as compared to the gliomas. Due to the low number of subjects and high inter subject variability of OEF and CBF values, no observed trend was statistically significant. P-values were about 0.15 using the Mann-Whitney U test.The observed trends of lower corr-CBF with high-OEF and lower corr-OEF with high-CBF are physiologically reasonable, especially for the GM control where a stable and homogeneous CMRO2 is a valid assumption. The high high-OEF also found for low grade gliomas is very interesting, since it suggests, that even low grade gliomas may possess hypoxic regions, which are difficult to detected with other blood oxygenation sensitive MRI methods⁴. Furthermore, the lower high-CBF of high grade gliomas suggests, that the proliferating tumor angiogenesis may cause hindered tumor perfusion. Additionally, the lower corr-OEF of both tumor grades as compared to GM control suggests that tumor cells are not able to utilize the oxygen oversupply in high perfusion areas. In summary, all these findings suggest that the QUIXOTIC method, in conjunction with pCASL perfusion measurements, is able to map tumor oxygen metabolism. We will continue to enroll patients into the study to reach statistical significance of the observed trends and we will include more parameters, i.e. contrast enhancement, blood volume and diffusion, in our future analysis to improve tumor characterization and to better cross validate the observed effects.