Correlation of breathing task derived cerebrovascular reactivity with baseline CBF, OEF and CMRO2

Ke Zhang^1*Simon M.F. Triphan¹Mark O. Wielpütz¹Johan Jende²Emilie Sleight³Christian Herbert Ziener²Mark E. Ladd²Heinz-Peter Schlemmer²Hans-Ulrich Kauczor¹Oliver Sedlaczek¹Felix T. Kurz⁴

• ¹Heidelberg University Hospital, Heidelberg, Germany
• ²German Cancer Research Center (DKFZ), Heidelberg, Baden-Württemberg, Germany
• ³Center for Biomedical Imaging (CIBM), Lausanne, Vaud, Switzerland
• ⁴University Hospitals of Geneva, Geneva, Geneva, Switzerland

Abstract Rationale and Objectives: Only a few studies examined the correlation between cerebrovascular reactivity (CVR) and other physiological parameters such as cerebral blood flow (CBF), oxygen extraction fraction (OEF) and cerebral metabolic rate of oxygen (CMRO2). In this study, these baseline parameters were measured using 3D MRI with whole brain coverage for the investigation of global and regional correlation between each other to enhance understanding of brain function and improve tumor diagnosis. Materials and Methods: All measurement were performed at 3T. CVR was derived from a breath-holding task. Baseline CBF was measured by pseudo-continuous arterial spin labeling. Baseline OEF was measured with a gradient-echo sampling of spin-echo pulse sequence. T1 weighted anatomical image (T1W) was measured using MPRAGE sequence. CVR was calculated using customized written programs. CBF was quantified by using ASLtbx. For OEF analysis, a feedforward artificial neural network was used. CMRO2 was calculated based on smoothed and normalized CBF and OEF. General linear regression analysis was used to examine the relations between CVR and other parameters in five lobes of gray matter including frontal, parietal, temporal, occipital and insula lobes in individual healthy subjects. Spearman correlation was performed to check the regional correlations in an Automated Anatomical Labeling (AAL) atlas. Results: 15 healthy volunteers and five patients with brain tumors were included. In the healthy subjects, five lobes had a positive correlation between CBF and CVR (p<0.05). Similarly, in five lobes positive correlations between CMRO2 and CVR were found (p<0.05), as well as significant inter-and intra-subject correlations (p<0.001). However, there were no significant correlations between OEF and other parameters. Conclusion: Our findings demonstrate that CVR is strongly associated with CBF and CMRO₂ at both global and regional levels in healthy brains, but not with OEF. These results provide new insight into the complex interplay between vascular reactivity, perfusion, and metabolism and underscore the potential of combined CVR-CBF-CMRO₂ imaging for assessing brain health and pathology.