On the evaluation of vorticity using cardiovascular magnetic resonance velocity measurements

Vorticity and vortical structures play a fundamental role affecting the evaluation of energetic aspects (mainly left ventricle work) of cardiovascular function. Vorticity can be derived from cardiovascular magnetic resonance (CMR) imaging velocity measurements. However, several numerical schemes can be used to evaluate the vorticity field. The main objective of this work is to assess different numerical schemes used to evaluate the vorticity field derived from CMR velocity measurements. We compared the vorticity field obtained using direct differentiation schemes (eight-point circulation and Chapra) and derivate differentiation schemes (Richardson 4* and compact Richardson 4*) from a theoretical velocity field and in vivo CMR velocity measurements. In all cases, the effect of artificial spatial resolution up-sampling and signal-to-noise ratio (SNR) on vorticity computation was evaluated. Theoretical and in vivo results showed that the eight-point circulation method underestimated vorticity. Up-sampling evaluation showed that the artificial improvement of spatial resolution had no effect on mean absolute vorticity estimation but it affected SNR for all methods. The Richardson 4* method and its compact version were the most accurate and stable methods for vorticity magnitude evaluation. Vorticity field determination using the eight-point circulation method, the most common method used in CMR, has reduced accuracy compared to other vorticity schemes. Richardson 4* and its compact version showed stable SNR using both theoretical and in vivo data.