Rotation Elastogram Estimation Using Mechanical Assisted Spatial Compounding: An Experimental Validation Study

Abstract

In quasi-static ultrasound elastography, breast tumor classification can be performed by using the rotation fill-in signature present in the rotation elastogram. This rotation fill-in signature is a benign tumors marker obtained from lesion rotation as a product of the lateral asymmetric stress field. However, the well-known low lateral resolution limitation of ultrasound imaging devices reduces the image quality of the rotation elastograms. Studies using beam-steering, synthetic transmit aperture (STA), diverging beam with STA, and sub-pitch translation of the ultrasonic beam methods have shown that the rotation elastogram’s quality can be significantly improved. In this context, we aim to study the feasibility of improving the rotation elastogram quality by mechanical-assisted spatial compounding of displacement images from different ultrasound probe angles. Recently, through numerical simulations, we have shown the theoretical feasibility of this technique. Here, we present the corresponding experimental validation by using tissue-mimicking gelatin phantoms. Our experimental results show that the contrast-to-noise ratio of the rotation elastogram can be improved by approximately 5 dB by increasing the number of displacement images used in the spatial compounding for small scanning angles of about 1°. This result confirms the prediction of previous numerical simulations. In addition, we show that the estimation of rotation can be used, in conjunction with the shear strain, to compute a new parameter to quantify the rotation possibility of the inclusion with respect to its shearing nature. In conclusion, our method is technically feasible but requires an exhaustive ultrasound probe position control in synchronism with the ultrasound image acquisition process.