Singular Value Decomposition Filtering for High Frame Rate Speckle Tracking Echocardiography

Recently, high frame rate (HFR) imaging approaches have been proposed together with dedicated speckle tracking (STE) algorithms capable of handling small interframe displacements. However, the compromised quality of HFR images affects the feasibility of STE. Therefore, the aim of this study was to assess whether Singular Value Decomposition (SVD) filtering could reduce HFR imaging artifacts and, possibly, improve STE feasibility. Specifically, 2-step SVD filter approach was proposed and tested on in vivo recordings, which were characterized by different heart rates, artefacts and noise levels. For both steps, the filter parameters were optimized to reject artifacts/noise. Then, filtered and original B-mode clips were shown side-by-side to an expert cardiologist, to select the preferred image based on visual reading. Although the clutter was suppressed as expected, the expert cardiologist preferred filtered images in 25% of the cases only. Moreover, the extracted systolic peak values of the Global Longitudinal Strain were not significantly different (p = 0.07), even if there was a tendency for slightly lower values after filtering. In general, these preliminary findings suggest that SVD filtering might not significantly affect HFR STE in a clinical setting.