Non-invasive small vessels imaging of human thyroid using motion-corrected power Doppler imaging

Imaging of microvascular pathways plays a significant role in detection of metastasis of cancerous masses in human thyroid. Contrast-free ultrasound based Doppler blood flow imaging is potentially useful to reduce the large number of benign findings of suspicious thyroid nodules. However, Doppler imaging of thyroid nodules is particularly demanding because the large motion due to its proximity with the carotid artery can impact both tissue clutter suppression and power Doppler (PD) integration. To address these issues, thyroid motion was estimated as components of axial and lateral displacements using a two-dimensional (2D) normalized cross correlation (NCC) - based speckle tracking of the Doppler ensemble, prior to singular value decomposition (SVD) clutter filtering. Moreover, we evaluated the feasibility of using non-redundant Doppler ensembles to track small-inter frame displacements, whose quantitative estimation is quite challenging because of the poor lateral resolution of ultrasonic images. This hypothesis was corroborated by applying the motion correction technique on phantom and in vivo data. Correlation matrices were used throughout the analysis as performance indices. Increased values of mean correlation coefficient after motion correction in phantom instances guaranteed the efficiency of the proposed optimized technique in detecting small inter-frame displacements and encouraged its validation on the in vivo dataset. The results from the in vivo study proved that the motion-corrected PD images importantly improved the visualization of the small vessels. In fact, motion induced signal distortion, blurring and appearance of shadow vessel were considerably reduced upon motion correction and noise suppression, without affecting vessel morphology.