Assessment of myocardial elastography performance in phantoms under combined physiologic motion configurations with preliminary in vivo feasibility

Abstract

Myocardial elastography (ME) is a non-invasive, ultrasound-based strain imaging technique, which can detect and localize abnormalities in myocardial function. By acquiring radio-frequency (RF) frames at high frame rates, the deformation of the myocardium can be estimated, and used to identify regions of abnormal deformation indicative of cardiovascular disease. In this study, the primary objective is to evaluate the effect of torsion on the performance of ME, while the secondary objective is to image inclusions during different motion schemes. Finally, the phantom findings are validated with an in vivo human case. Phantoms of homogeneous stiffness, or containing harder inclusions, were fixed to a pump and motors, and imaged. Incremental displacements were estimated from the RF signals, and accumulated over a motion cycle, and rotation angle, radial strain and circumferential strain were estimated. Phantoms were subjected to four motion schemes: rotation, torsion, deformation, and a combination of torsion and deformation. Sonomicrometry was used as a gold standard during deformation and combined motion schemes. In the rotation scheme, the input and estimated rotation angle agree in both the homogeneous and inclusion phantoms. In the torsion scheme, the estimated rotation angle was found to be highest, closest to the source of torsion and lowest farthest from the source of torsion. In the deformation scheme, if an inclusion was not present, the estimated strain patterns accurately depicted homogeneity, while if an inclusion was present, abnormalities were observed which enabled detection of the inclusion. In addition, no significant rotation was detected. In the combined scheme, if an inclusion was not present, the estimated strain patterns accurately depicted homogeneity, while, if an inclusion was present, abnormalities were observed which enabled detection of the inclusion. Also, torsion was separated from the combined scheme and was found to be similar to the pure torsion findings. This study shows ME to be capable of accurately depicting and distinguishing between different types of motion schemes, and to be sensitive to stiffness changes in localized regions of tissue-mimicking phantoms under physiologic cardiac motion configurations, while strains estimated in the combined motion scheme were noisier than in individual motion schemes. Finally, ME was shown to be capable of distinguishing between deformation and rotation in a normal human heart in vivo.