Citation
Beeran Kutty, Suhaili
(2020)
Three-dimensional coronary artery reconstruction based on single-plane angiography and intravascular ultrasound registration.
Doctoral thesis, Universiti Putra Malaysia.
Abstract
The mortality rate of heart failure or heart disease has been increasing all over the
world. Coronary artery disease (CAD) is one type of heart disease that is caused by the
atherosclerosis process. The process starts with the building up of plaque along the
inner walls of the coronary arteries. The obstruction of the artery will reduce the blood
flow that circulates through the heart. The reduced blood flow to the heart can cause
angina or chest pain and heart attack. X-ray coronary angiography is the gold standard
test to investigate the affected artery, however, this modality only visualizes the
silhouette of the artery. To help the diagnostic process, Intravascular Ultrasound
(IVUS) based on ultrasound technology, is used to show the internal structure of the
artery. Coronary angiography and IVUS complement each other. However, to use
coronary angiography and IVUS simultaneously, cardiologists view both modalities on
the same screen, but in a different window. They are merged mentally using the
biological landmarks on the modalities and also based on the cardiologist’s knowledge
and experiences. Mental registration can cause errors in the interpretation of the plaque
and is time-consuming. Because of that, research on the registration of Angiography
and IVUS have been highlighted until today. Most of the previous registration
approaches are based on the bi-plane Angiography that required greater radiation dose
and a common technique for the registration, based on the epipolar geometry. The
disadvantages of this technique are that it requires at least two views of coronary
angiography, and approximation is used in determining the correspondence points. In
this research, the registration proposes to use a single-plane Angiography, which is the
medical imaging technology that is still relevant and used in many catheterization labs
around the world to reconstruct the 3D artery based on the registration with IVUS
modality. Firstly, the IVUS modality; the catheter shape, the lumen and the mediaadventitia
are completely segmented using labeling and a parametric deformable
model. Next, the trajectory of the artery is reconstructed based on the catheter path of
IVUS and the biological landmarks found on the coronary angiography image. A new
method for reconstructing 3D coronary artery based on the circle trigonometry equation for the registration of single-plane coronary angiography and IVUS was
introduced in this research. To help the CAD diagnostic process, this research offers an
accurate measurement of the lumen cross-sectional area, minimum lumen diameter,
and maximum lumen diameter. In IVUS segmentation, the result shows that 100% of
the catheter shape on IVUS image is detected correctly using the labelling approach.
The accuracy of the lumen segmentation increases with a percentage of area difference
(PAD) 0.07, and the media-adventitia segmentation could be done without manual
initialization process. It is also proven that the trajectory of the artery could be
generated using only one view of Angiography that contains the catheter path of IVUS.
Furthermore, the accuracy of the equation used in the registration experiment is
acceptable, where it is shown by is 0.9999 more than the significant value. As a
result, the new approach of the registration of the single-plane Angiography and IVUS
to produce a 3D coronary artery is accepted by the medical practitioners. The 3D
model shows the precise location of the catheter in the artery and the coronary artery is
measured through the cross-sectional area of the lumen and the minimum and
maximum diameters. The result of this study may help us to understand the
longitudinal view of the coronary artery and the morphology of the arterial wall. It
could be used by cardiologists to make decisions in diagnosing coronary artery disease.
Download File
Additional Metadata
Actions (login required)
![View Item View Item](/style/images/action_view.png) |
View Item |