A Computationally Efficient Approach to Segmentation of the Aorta and Coronary Arteries Using Deep Learning.

Wing Keung Cheung,Kacy Chou,Joseph Jacob,Riyaz Patel,Daniel C Alexander,Simon Wan,Robert Bell,James C Moon,Ryo Torii,Arjun Nair,Leon J Menezes,Jiahang Chen,Rhodri H Davies

doi:10.1109/access.2021.3099030

Abstract

Early detection and diagnosis of coronary artery disease could reduce the risk of developing a heart attack. The coronary arteries are optimally visualised using computed tomography coronary angiography (CTCA) imaging. These images are reviewed by specialist radiologists who evaluate the coronary arteries for potential narrowing. A lack of radiologists in the UK is a constraint to timely diagnosis of coronary artery disease, particularly in the acute accident and emergency department setting. The development of automated methods by which coronary artery narrowing can be identified rapidly and accurately are therefore timely. Such complex computer based tools also need to be sufficiently computationally efficient that they can run on servers typically found in hospital settings, where graphical processing units for example are unavailable. We propose a fully automatic two-dimensional Unet model to segment the aorta and coronary arteries on CTCA images. Two models are trained to segment two regions of interest, (1) the aorta and the coronary arteries or (2) the coronary arteries alone. Our method achieves 91.20% and 88.80% dice similarity coefficient accuracy on regions of interest 1 and 2 respectively. Compared with a semi-automatic segmentation method, our model performs better when segmenting the coronary arteries alone. The performance of the proposed method is comparable to existing published two-dimensional or three-dimensional deep learning models. Importantly, the algorithmic and graphical processing unit memory efficiencies are maintained such that the model can be deployed without requiring graphical processing units, and therefore can be used in a hospital setting.

Highlights

Coronary artery disease (CAD) is one of the leading causes of death in the UK [1] and worldwide
The focus of the current study relates to deep learning methods for coronary artery segmentation and we briefly summarise the work published in the following paragraph
All the patients in the study had presented with chest pain and associated symptoms that indicated an intermediate risk of coronary artery disease

Summary

INTRODUCTION

Coronary artery disease (CAD) is one of the leading causes of death in the UK [1] and worldwide. An alternative approach for assessing CAD severity involves performing computational fluid dynamics (CFD) on the target arteries [11] It first requires identification of accurate geometries of the aorta and coronary arteries. There remains an unmet need to develop fast, objective and accurate automated computer-derived coronary artery segmentation algorithms that can be deployed in a hospital setting to assist clinicians diagnose CAD This is especially relevant in Accident and Emergency (A+E) departments where CTCA reviews are often delayed due to a lack of available specialists to read the CTCAs [21]. We propose a 2D Unet to perform aorta and/or coronary artery segmentation and demonstrate that this 2D Unet is practically feasible to be implemented given that the computational resources are limited to those available in a hospital network whilst maintaining a good segmentation accuracy. Our technique works well when the coronary arteries alone are segmented (accuracy ~89%)

METHOD

EXPERIMENTS AND RESULTS

EVALUATION

DISCUSSIONS

CONCLUSION