Two-Dimensional Tracking of Heart Wall for Detailed Analysis of Heart Function at High Temporal and Spatial Resolutions

For noninvasive and quantitative measurements of global two-dimensional (2D) heart wall motion, speckle tracking methods have been developed and applied. These methods overcome the limitation of tissue Doppler imaging (TDI), which is susceptible to aliasing, by directly tracking backscattered echoes by pattern matching techniques, i.e., the cross-correlation or the sum of absolute differences, in real time. In these conventional methods, the frame rate (FR) is limited to about 200 Hz, corresponding to the sampling period of 5 ms. However, myocardial function during the isovolumic contraction period obtained by these conventional speckle tracking methods remains unclear owing to low temporal and spatial resolutions of these methods. Moreover, the accuracy of the speckle tracking method depends on an important parameter, i.e., the size of the correlation kernel. To track backscattered echoes accurately, it is necessary to determine the optimal kernel size. However, the optimal kernel size has not been thoroughly investigated. In this study, correlation kernel size, which determines the tracking accurately, was optimized by evaluating root mean squared (RMS) errors in the lateral and axial displacements of a phantom estimated by speckle tracking methods at high spatial and temporal resolutions. For this purpose, the RF data from the longitudinal-axis cross-sectional view for the interventricular septum (IVS) were acquired on the basis of parallel beam forming (PBF) to improve temporal and spatial resolutions. A wide transmit beam scanned in 7 different directions sparsely and 16 receiving beams were generated for each transmission. The RF data of the phantom and the heart wall were obtained at high spatial (angle intervals of scan lines: 0.375 degrees) and temporal [frame rate (FR): 1020 Hz] resolutions. The determined optimal size of the correlation kernel was 7.9° ×4.8 mm. Estimated displacements of the phantom were in good agreement with the actual displacement at an RMS error of 0.34 mm. Furthermore, the IVS motion during the isovolumic contraction (IC) was analyzed in detail. The speckle tracking method using the optimal kernel size 7.9° ×4.8 mm was applied to multiple points in IVS to estimate 2D displacements during the IC period. In this period, a rapid displacement of IVS at a small amplitude of 1.5 mm, which suggests the expansion of the left ventricle and has not been measured by conventional tracking methods at a low temporal resolution, was detected by 2D tracking. Furthermore, the displacement on the apical side was found to be delayed by 10 ms compared with that on the basal side. These results indicate the potential of this method in the high-accuracy estimation of 2D displacements and detailed analyses of physiological function of the myocardium.