Novel Coronary Tortuosity Extraction Using Intravascular Ultrasound Sequence Alone

摘要

Considering Intravascular Ultrasound (IVUS) frames, the lack of three-dimensional (3-D) information does not allow to recognize the real shape of the coronary vessel, and thus to analysis atherosclerosis treatment completely. In this research, a novel framework is proposed to extract vessel tortuosity by mapping the IVUS sequence frames based on catheter center in 3-D space without any external information. In the other word, tilt angle of catheter is distinguished by finding an angle that two mapped sequence frames have a maximum similarity coefficient. The obtained tortuosity is validated with its real 3-D shape reconstructed using a developed algorithm that fuses IVUS and angiography images. With the obtained parametric curvature difference of 0.178 ± 0.035 mm?1 between the real curvature and the obtained curvature, we showed that the presented framework is suitable for full 3-D registration of IVUS sequences. This simplifies a wide range of coronary assessments like full analysis of atherosclerosis treatment and predicting of plaque progression.