Improving Cardiac Phase Extraction in IVUS Studies by Integration of Gating Methods

Abstract

Goal: Coronary intravascular ultrasound (IVUS) is a fundamental imaging technique for atherosclerotic plaque assessment. However, volume-based data retrieved from IVUS studies can be misleading due to the artifacts generated by the cardiac motion, hindering diagnostic, and visualization of the vessel condition. Then, we propose an image-based gating method that improves the performance of the preexisting methods, delivering a gating in an appropriate time for clinical practice. Methods: We propose a fully automatic method to synergically integrate motion signals from different gating methods to improve the cardiac phase estimation. Additionally, we present a local extrema identification method that provides a more accurate extraction of a cardiac phase and, also, a scheme for multiple phase extraction mandatory for elastography-type studies. Results: A comparison with three state-of-the-art methods is performed over 61 in-vivo IVUS studies including a wide range of physiological situations. The results show that the proposed strategy offers: 1) a more accurate cardiac phase extraction; 2) a lower frame oversampling and/or omission in the extracted phase data (error of 1.492 ± 0.977 heartbeats per study, mean ± SD); 3) a more accurate and robust heartbeat period detection with a Bland-Altman coefficient of reproducibility (RPC) of 0.23 s, while the second closest method presents an RPC of 0.36 s. Significance: The integration of motion signals performed by our method shown an improvement of the gating accuracy and reliability.