Abstract
PURPOSE: There is a clinical need for stiffness mapping of the heart; however, current cardiac magnetic resonance elastography (cMRE) has limited spatiotemporal resolution. Therefore, we developed 2D spiral multifrequency MRE of the human heart and conducted a study to analyze the consistency and reproducibility of motion-encoding and stiffness mapping with and without external vibration. METHODS: Eleven healthy volunteers were studied using single-slice gradient-echo spiral cMRE with cardiac triggering and encoding of harmonic shear wave fields at 70, 80, and 90 Hz frequency generated by either external drivers or endogenous heart motion. Tissue displacement was monitored synchronized to the cardiac cycle, and frequency-resolved shear wave speed (SWS) maps were reconstructed as a proxy for left ventricular (LV) stiffness variations. After several days, all subjects underwent repeat scanning for reproducibility analysis based on intraclass correlation coefficients (ICCs). RESULTS: cMRE with external vibration showed LV SWS to be highest in end-systole (ES) (2.17 ± 0.23 m/s), followed by diastole (DIA) (1.94 ± 0.15 m/s) and isovolumetric contraction (IVC) (1.78 ± 0.18 m/s). ICCs decreased with distance from the R-wave from excellent (0.93) in IVC to moderate (0.68) in ES. Without external vibration, sufficient LV harmonic displacement amplitudes permitted SWS reconstruction, resulting in similar SWS, but lower ICC values than with external vibration in IVC and ES. CONCLUSIONS: Multifrequency cMRE offers high spatiotemporal resolution and reproducibility with external vibration. In addition, the technique allows the encoding of endogenous shear waves during cardiac phases with pronounced wall motion.