Quantification of normal cerebral oxygen extraction and oxygen metabolism by phase‐based MRI susceptometry: evaluation of repeatability using two different imaging protocols

Summary Introduction Global oxygen extraction fraction ( OEF ) and cerebral metabolic rate of oxygen ( CMRO 2 ) were quantified in a test–retest study. Cerebral blood flow ( CBF ) data, required for CMRO 2 estimation, were obtained using dynamic susceptibility contrast MRI ( DSC ‐ MRI ). OEF and CMRO 2 were quantified using two separate data sets, that is, conventional high‐resolution ( HR ) gradient echo ( GRE ) phase maps as well as echo planar imaging ( EPI ) phase maps taken from the baseline (precontrast) part of the DSC ‐ MRI time series. The EPI phase data were included to elucidate whether an extra HR ‐ GRE scan is needed to obtain information about OEF and CMRO 2 , or if this information can be extracted from the DSC ‐ MRI experiment only. Methods Twenty healthy volunteers were scanned using 3 T MRI on two occasions. Oxygen saturation levels were obtained from phase data measured in the great cerebral vein of Galen, based on HR ‐ GRE as well as EPI phase maps. In combination with DSC ‐ MRI CBF , this allowed for calculation of OEF and CMRO 2 . Results High‐resolution‐gradient echo‐ and EPI ‐based phase images resulted in similar OEF spread and repeatability, with coefficients of variation/intraclass correlation coefficients of 0·26/0·95 and 0·23/0·81, respectively. Absolute OEF values ( HR ‐ GRE : 0·40 ± 0·11, EPI : 0·35 ± 0·08) were consistent with literature data. CMRO 2 showed similar repeatability, somewhat increased spread and reasonable absolute values ( HR ‐ GRE : 3·23 ± 1·26 ml O 2 /100 g min −1 , EPI : 2·79 ± 0·89 ml O 2 /100 g min −1 ). Discussion In general, the results obtained by HR ‐ GRE and EPI showed comparable characteristics. The EPI methodology could potentially be improved using a slightly modified DSC ‐ MRI protocol (e.g. with regard to spatial resolution and slice gap).