Evaluation of Accuracy of Automatic Out-of-plane Respiratory Gating for DCEUS-based Quantification Using Principal Component Analysis

Diya Wang,Zhe Su,Qiang Su,Xinyu Zhang,Zhen Qu,Na Wang,Yujin Zong,Yaning Yang,Mingxi Wan

DOI: https://doi.org/10.1016/j.compmedimag.2018.10.004

IF: 7.422

2018-01-01

Computerized Medical Imaging and Graphics

Abstract:The accuracy of abdominal multi-parametric quantification based on dynamic contrast-enhanced ultrasound (DCEUS) is limited by out-of-plane severe distortion induced by respiratory motion. This study developed a fully automatic respiratory gating scheme by using principal component analysis to remove distortions and disturbances in free-breathing DCEUS-based quantification. Taking the known in-vitro perfusions as ground truths, we further evaluated the respiratory gating accuracy from multiple perspectives in a controllable rotary distortion flow model with out-of-plane severe distortion induced by respiratory motion. Compared with those without respiratory gating, the signal-to-clutter ratio and correlation coefficients of perfusion parameters with respiratory gating improved by 3.99 ± 1.71 dB (p < 0.01) and 0.39 ± 0.17 (p < 0.01), respectively. The corresponding mean square error decreased by 1893.9 ± 763.16 (p < 0.001). Results with respiratory gating were significantly consistent with the ground truths without respiratory motion disturbances. Quantitative results demonstrated the effective removal of out-of-plane serious distortion and other disturbances induced by the respiratory kinetics on DCEUS-based quantification. The accuracy and robustness of DCEUS-based quantification were significantly improved by the proposed respiratory gating strategy, which not only benefits physicians in identifying transient hemodynamic characteristics in tumor angiogenesis, but also provides accurate diagnoses for abdominal tumors.