Automatic quantification of intravascular optical coherence tomography

Abstract

As mentioned above, IVOCT, as a novel imaging modality, has played an active role in a wide range of CAD applications, including research and clinical routine. Due to its unparalleled high resolution and the ability to delineate complex vascular structures, IVOCT technology makes many precise measurement and novel applications possible. However, currently, a lot of analyses in IVOCT images are still relying on the manual work which decreases their value. The goal of this thesis is to develop robust and accurate (semi)automated methods that can detect and segment the interesting components in IVOCT pullback runs, such as implanted stent struts and side branches in 3D for accurate measurement, so that the results could contribute to medical research as well as for clinical decision-making. My thesis presents four different automated algorithms to detect metallic stent struts, bioresorbable vascular scaffold struts, side branches and all the common components in IVOCT images. It also presented a semi-automated method to assess the stent support to vessel wall and the stent-jailed side branch access through stent cells in 3-dimentional space