ABSTRACT
This paper presents the AutoQual elastography method: a novel algorithm that improves the quality of 2D displacement field calculation from ultrasound radio frequency (RF) sequences of acutely ruptured Achilles tendons to determine image-lateral strain fields and has potential use for ligaments and muscles. This method uses 2D bicubic spline interpolation of the RF signal, Quality Determined Search, Automatic Search Range and Adaptive Block Size components as a novel combination that is designed to improve continuity and decrease displacement field noise, especially in areas of low signal strength. We present a simple experiment for quantitatively comparing the AutoQual method to a multiscale (MS) elastography method from ultrasound RF sequences of a 5% agar phantom for rigid body motion and known lateral strain loads with speeds up to 5mm/s. We finally present examples of four in vivo Achilles tendons in various damage states and with manual or artificially controlled passive flexion of the foot. Results show that the AutoQual method offers a substantial improvement on the MS method, achieving similar performance for rigid body tracking at all speeds, a lower normalized square error at all strains induced and a more continuous strain field at higher compression rates. AutoQual also showed a greater average normalized cross correlation for image blocks in the area of interest, a lower standard deviation of the strain field and a visually more acceptable point tracking for in vivo examples. This work demonstrates lateral ultrasound elastography which is robust to the complex passive motion of the Achilles and to various imaging artifacts associated with imaging tendon rupture. This method potentially has a wide clinical application for assessing in vivo strains in and hence mechanical function of any near skin surface tissues that are longitudinally loaded.
