UCLA

Eddy currents induce velocity errors (veddy) that influence PC-MRI derived parameters; static tissue corrections aim to minimize these veddy influences. Local and global static tissue corrections have been proposed; however, these separate strategies have yet to be thoroughly characterized. The dependence of local and global static tissue corrections on SNR and the amount of static tissue used for correction is not well understood. To characterize these relationships, corrections were analyzed within a static phantom and flow phantom experiment. Corrections were repeated for a range of SNR and for increasing amounts of static tissue used for polynomial fitting. Local and global corrections were separately applied to reduce velocity offset within ROIs within a static tissue phantom. Global correction reduced velocity offset in 100% of ROIs and local correction reduced velocity offset in ~96% of ROIs. Local correction introduced error in ~4% of ROIs (always when SNR < 30). Correction differences between local and global strategies in the static phantom experiment were on the order of 0.9 cm/s for low SNR protocols and 0.2 cm/s for high SNR protocols. Velocity offset within static ROIs was reduced by up to an additional ~0.6 cm/s by increasing static tissue coverage for low SNR protocols (~0.05 cm/s for high SNR protocols). Local and global corrections were separately applied to correct total volumetric flow estimates within a flow phantom. Veddy errors induce differences between total flows measured at different locations along a flow phantom. Static tissue corrections reduced these differences between total flow estimates in 90% of corrections. Of these cases, global correction provided a larger reduction in flow difference ~67% of the time. Local fitting introduced error in ~27% of total corrections, and errors were observed for a range of SNR. Global fitting introduced one instance of error (for lowest SNR). Flow differences were reduced by up to an additional ~2 mL by increasing static tissue coverage. This study indicates that: 1) Local static tissue correction is suitable only for higher SNR PC-MRI applications and 2) Maximizing the amount of static tissue used during polynomial fitting improves performance of applied corrections for low SNR applications.