Quantitative Mapping of Cerebral Blood Flow Change using Phase Information of SWI

Introduction Phase image of MRI contains information about local difference of magnetic susceptibility, originated from blood product, iron, etc. [1]. One of the recent utilities of phase image is susceptibility weighted imaging (SWI), which emphasizes susceptibility differences by using the phase shift of the spin derived from the paramagnetic substances mentioned above. By using the bulk susceptibility differences included in SWI, manual calculation of the blood flow changes and oxygen saturation in veins has been reported [2,3]. For an application of this theory, automated calculation of cerebral blood flow (CBF) changes can be achieved as a whole brain CBF mapping. The purposes of this study were (1) to develop the automated method to obtain a map of CBF change, (2) to assess the reliability and stability of the automated method in measuring rest CBF, compared to the manual method reported previously [3], and (3) to evaluate the accuracy of this method in measuring CBF changes induced by the respiratory stress, compared to arterial spin labeling (ASL). Materials and Methods A 1.5T MR imager (Avanto, Siemens, Erlargen) was used to obtain phase images of SWI, with a fully velocity-compensated, three-dimensional fast low-angle shot (3D-FLASH, TR = 49 ms, TE = 40 ms, flip angle = 20°, and bandwidth = 80 Hz/pixel). The phase images were reconstructed and imported to our original program, Perfusion Mismatch Analyzer (PMA), to make CBF change (∆ƒ) map. First, a high pass filter was applied to the phase image (Fig. 1A). Then, the veins were extracted by using sliding window and local threshold (Fig. 1B). Finally, the venous phase shifts (∆φ) were calculated locally (Fig. 1C). ∆ƒ map (Fig. 1D) was created according to the following formulas with oxygen saturation changes (∆Y) [2,3].