3D multimodal cardiac data reconstruction using computerized tomographic angiography and x-ray angiography registration

Computerized tomographic angiography (3D data representing coronary arteries) and X-ray angiography (2D X-ray image sequences providing information about coronary arteries and their stenosis) are standard and popular assessment tools utilized for medical diagnosis of coronary artery diseases. At...

Full description

Saved in:
Bibliographic Details
Main Author: Moosavitayebi, Seyed Rohollah
Format: Thesis
Language:English
Published: 2016
Subjects:
Online Access:http://psasir.upm.edu.my/id/eprint/69309/1/FSKTM%202016%201%20IR.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Computerized tomographic angiography (3D data representing coronary arteries) and X-ray angiography (2D X-ray image sequences providing information about coronary arteries and their stenosis) are standard and popular assessment tools utilized for medical diagnosis of coronary artery diseases. At present, the results of both modalities are analyzed individually by medical specialists, and it is difficult for them to mentally connect the details of these two techniques. This research aims to propose a new framework for coronary artery registration in both modalities to provide the 3D position of the stenosis diagnosed in X-ray angiography images. In this study, coronary arteries from two modalities are registered in order to create a 3D reconstruction and visualization of stenosis position. Over the last decade, some algorithms have been developed to register coronary arteries from the above modalities. However, most of them failed to register coronary arteries in order to estimate stenosis points from X-ray angiography on computerized tomographic angiography modalities. In this research work, we report on new contributions to the major parts of a 3D multimodal cardiac data reconstruction system. The first contribution is to propose a fast, accurate and fully automatic method for coronary artery segmentation and labeling from X-ray angiography, based on the proposed modified Starlet Wavelet Transform for segmentation, so that the method can be utilized for registering with other modalities. The average accuracy, sensitivity, specificity, and precision values of the proposed method in LCA angiograms from the data sets are 0.96934, 0.86014, 0.98439, and 0.87797, respectively, and in RCA angiograms, the values are 0.97425, 0.89962, 0.99587, and 0.93021, respectively. Obviously, the proposed method is robust in all performance metrics. Also, by comparing the results, it shows that the proposed method has minimum artifacts and it also segmented the whole parts of the coronary arteries. Furthermore, the running time for the proposed method is much better than other methods, as the whole process is done in less than 1 s. The second contribution is related to propose a fast and accurate method for 3D main coronary artery segmentation and labeling from CT Angiography based on the proposed Intersection Tracking method and Improved vesselness filter. By conducting experiments on the clinical data sets, it is proven that the proposed method improved the ability of 3D coronary artery segmentation and labeling from computerized tomographic angiography by increasing the overall overlap evaluation to 95.76%. In addition, since no registration is needed prior to applying the proposed intersection tracking method, and also segmentation and labeling are applied at the same time in this work, hence, it is faster compared to previous methods. The feature-based coronary artery registration in computerized tomographic angiography and X-ray angiography by using the provided features in both modalities is the topic of the third contribution in this work. Tests using the clinical data sets demonstrated that the proposed method aided the specialists to find the location of stenosis lesion and also to determine the visual relationship between the corresponding coronary arteries with the mean value 3D distance of 3.375mm and standard deviation of ±1.4137mm in a maximum processing time of 0.2 s for each coronary artery registration. The proposed research work is applicable and portable for common personal computers, as well as with respect to the standard medical acquisition methods. Moreover, the medical acquisition standards remain unchanged in this work, which means that no calibration in the acquisition devices is required, and it can be applied to most computerized tomographic angiography and X-ray angiography devices. Another benefit of this work would be each corresponding coronary arteries from these two modalities can be registered individually and finally, the results can be combined and displayed as a whole coronary arterial tree including stenosis.