PURPOSE

In the energy metabolism of most organisms molecular oxygen (O₂) plays a vital role. In various diseases such as cancer (‘Warburg effect’) [1], Parkinson’s [2], or Alzheimer’s disease [3] the functional parameter of cerebral metabolic rate of oxygen consumption (CMRO₂) contains information about tissue viability. The CMRO₂ parameter can be measured via dynamic ¹⁷O MRI of the stable oxygen isotope ¹⁷O employing an inhalation procedure where enriched ¹⁷O₂ gas is administered during continuous imaging [4]. Recently, the reproducibility and reliability of CMRO₂ determination via ¹⁷O MRI with additional partial volume (PV) correction was demonstrated in a small volunteer cohort [5]. The presence of strong PV effects in ¹⁷O-MRI is caused by low spatial resolution and rapid transverse relaxation (T₂~5ms), requiring an efficient correction procedure to separate signal contribution from different compartments. In the presented work the verified method from [5] was utilized in a clinical setting to measure two tumor patients and quantifying CMRO₂ values in various regions in healthy tissue and tumor regions.

METHODS

A three-phase inhalation experiment [4] (t₁ baseline-phase; t₂ ¹⁷O-inhalation-phase; t₃ decay-phase) was performed with a MR‑compatible breathing system in which a ¹⁷O₂-bolus is administered in a closed circuit. In this ongoing IRB approved study one untreated glioma-patient (male, 26y.-o. astrocytoma WHO grade II) and one untreated glioblastoma-patient (male, 34y.-o., WHO grade IV) were included. ¹⁷O MRI was conducted at a 7 Tesla whole-body scanner (Siemens Healthineers, Erlangen, Germany) with a nominal spatial resolution of (7.5mm)³ employing a density-adapted radial sequence [6] with Golden Angle acquisition scheme [7] (TR/TE=20/0.56ms, total acquisition time 30:00min). Data were reconstructed with a temporal resolution of Δt=1:00min. Per experiment 3.8±0.1L of 70%-enriched ¹⁷O₂ gas (NUKEM Isotopes Imaging GmbH, Alzenau, Germany) were administered. High-resolution T₁ MPRAGE (0.6mm)³ and T₂ TSE (0.4x0.4x0.5mm)³ sequences were additionally acquired and used for tumor and normal brain segmentation for the employed PV correction algorithm [8,9]. Prior to patient measurements the imaging protocol was optimized utilizing dynamic numerical signal simulation of the human brain: the impact of shortening of the baseline- and decay-phases on the accuracy of quantification was investigated and the final patient protocol was adjusted accordingly.

RESULTS

The available fit information in individual breathing phases (t₁, t₃) was systematically reduced in two considered compartments (gray matter (GM), white matter (WM)). The deviation from the simulation ground truth is shown in Fig.1. Only a minor dependence is seen for variation of t₁ and a deviation >5% for t₃<15min. With this information the patient inhalation protocol was set to: t₁=5:00min, t2≤10:00min (variable until ¹⁷O₂ is exhausted) , t₃≤15:00min. Results of the glioma WHO II -patient revealed significantly decreased CMRO₂ in tumor tissue (0.59–0.66±0.16μmol/g/min) compared to contralateral normal appearing brain tissue (1.10–1.22±0.08 μmol/g/min, control compartment, Tab.1). Quantitative data are shown in Fig.2 and a map of relative ¹⁷O signal increase visualizes the decrease of oxygen metabolization in the tumor region (Fig.3). Further, CMRO₂ was significantly increased in gray matter (2.34–2.58±0.20μmol/g/min) compared to white matter tissue (0.56–0.62±0.06μmol/g/min) after application of PVC. CMRO₂ determination in the glioblastoma patient, revealed an even more pronounced metabolization drop in the tumor tissue, with lowest values in the necrotic (NE) as well as contrast enhancing (CE) tumor region (Tab.1).

DISCUSSION

Application of a simulation tool allowed protocol optimization and estimation of expected fitting errors due to information reduction in individual breathing phases. The protocol and utilized inhalation setup allowed measurements in two patients and show the feasibility of transferring the suggested method to a clinical environment. The drop in quantified functional parameters in the malignant tissue (CE, NE, PE) corresponds to the Warburg effect, which describes decreased CMRO₂ in tumors due to the shift in glucose metabolism from oxidative phosphorylation to lactate production for energy generation as reported for ¹⁷O MRI data for the first time by Hofmann et al. [10]. Overall functional parameters are in good agreement with previous reports (Tab.1) [10, 11]. Due to large voxel volumes and rapid transverse relaxation, the quantified H₂¹⁷O concentrations might be underestimated despite the application of a dedicated PVC algorithm. Uncertainties in the prior ¹⁷O enrichment factor as well as limitations of the PVC as discussed in [5] are the main sources of error in the suggested method. However, CMRO₂ values of healthy tissue (GM, WM) and cerebrospinal fluid (CSF) are in correspondence with previous work [4,5].

CONCLUSION

This work presents the first results of an ongoing study investigating brain tumor metabolism in glioma-patients with the help of dynamic ¹⁷O MRI. Further application may provide new insights into tumor pathophysiology through the visualization of the cerebral metabolic rate of oxygen consumption.

Acknowledgements

The authors want to thank NUKEM Isotopes Imaging GmbH for their generous supply of ¹⁷O₂-gas and support of this project.

References

[1] Miles KA, Williams RE., Cancer Imaging 2008 (8):81-86; [2] Beal MF., Ann Neurol; 1992 (2):119-130; [3] Maurer I, Neurobiology of Aging 2000;21(3):455-462; [4] Atkinson IC, Neuroimage 2010 (51):723-733; [5] Niesporek SC, Magn Reson Med 2017, doi: 10.1002/mrm.26952; [6] Nagel et al., Magn Reson Med 2009 (62):1565-73; [7] Chan, R.W. et al., Magn Reson Med 2009(61): p. 354–363; [8] Rousset et al., J Nucl Med 1998(5):904-911, 1998, [9] Niesporek et al., NeuroImage, 2015(112): 353–363; [10] Hoffmann et al., MAGMA 2014(27):579-87; [11] Ito, M., Neuroradiology 1982 (23): 63–74