Computational methods for atherosclerotic plaque characterization (Doctoral thesis)
Αθανασίου, Λάμπρος Σ.
This thesis focuses on the atherosclerotic plaque characterization using invasive and non-invasive imaging modalities. The outermost aim is the development and implementation of atherosclerotic plaque characterization methodologies using Intravascular Ultrasound (IVUS), Optical Coherence Tomography (OCT) and Computed Tomography (CT) images of coronary vessels. The research was initiated after studying the limitations of existing automated plaque characterization methodologies, applied to widely used imaging modalities. In order to overcome these limitations, several methodologies have been developed. In the first chapter, the anatomy and physiology of the cardiovascular system is presented. It is described in detail and the relevant literature is provided for the following topics: heart functionality along with the vascular system, blood circulation and atherosclerosis. The imaging methods, invasive and non-invasive, for the investigation of plaque composition are presented also in detail. In the second chapter the up to date literature and their limitations are highlighted. The methodologies developed for segmenting and characterizing atherosclerotic plaque using IVUS, OCT and CT images are presented. In the third chapter two novel methodologies for the automated identification of different plaque components in grayscale IVUS images, are presented. The first methodology is based on a hybrid approach that incorporates both image processing techniques and classification algorithms. The methodology allows classification of the plaque into three different categories, providing reliable results when compared to experts' annotations. The second IVUS plaque characterization methodology classifies the plaque into four classes. In order to train and validate the methodology, we used IVUS frames acquired from a commercially available software. In the fourth chapter a novel methodology that processes OCT images in a fully automated manner is presented. The proposed methodology is able to detect the lumen borders in OCT frames, identify the plaque region and detect four plaque types. The efficiency of the developed methodology was evaluated using experts' annotations. In the fifth chapter a new methodology for automated three-dimensional (3D) reconstruction of coronary arteries and characterization of plaque morphology in CT images, is presented. The methodology detects the inner - outer wall vessel borders, detects the calcium plaque and reconstructs the vessel and plaque into 3D surfaces. The methodology was validated using the estimations of the recently presented IVUS plaque characterization methodology. In the sixth chapter a study for propagating the imaging errors and image segmentation errors in plaque characterization methodologies applied to 2D vascular images, is presented. The maximum error that can be propagated to the plaque characterization results is calculated, assuming worst case scenarios. The proposed error propagation methodology is validated using methods applied to real datasets, obtained from IVUS and OCT systems. Finally, in the seventh chapter, a framework for the inflation of histology/micro-CT images based on IVUS is presented. The proposed methodology inflates the deformed histological/micro-CT images, which are the current gold standard in plaque characterization, based on the IVUS and micro-CT/histological lumen contour difference. In order to validate the proposed image inflation methodology plaque areas in the inflated micro-CT and histological images are compared with the ones in the IVUS images. The contribution of this thesis is related to the following: (i) the implementation of new plaque characterization methodologies which overcome the literature limitations, (ii) the emergence of producing plaque characterization methodologies that can be applied to recently developed imaging systems, (iii) the presentation of methodologies able to produce 3D arterial and plaque models for investigating the evolution of the plaque, (iv) the study and analysis of the error propagation of imaging systems to plaque detection, and (v) the automated image registration of intravascular images with histology.
|Institution and School/Department of submitter:||Πανεπιστήμιο Ιωαννίνων. Σχολή Θετικών Επιστημών. Τμήμα Μηχανικών Επιστήμης Υλικών|
|Subject classification:||Atherosclerotic plaque|
|Keywords:||Atherosclerotic plaque,Cardiovascular system|
|Appears in Collections:||Διδακτορικές Διατριβές|
Files in This Item:
|Δ.Δ. ΑΘΑΝΑΣΙΟΥ ΛΑΜΠΡΟΣ Σ. 2015.pdf||9.1 MB||Adobe PDF||View/Open|
Please use this identifier to cite or link to this item:This item is a favorite for 0 people.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.