Citation: Huang, F.; Lian, J.; Ng,
K.-S.; Shih, K.; Vardhanabhuti, V.
Predicting CT-Based Coronary Artery
Disease Using Vascular Biomarkers
Derived from Fundus Photographs
with a Graph Convolutional Neural
Network. Diagnostics 2022, 12, 1390.
https://doi.org/10.3390/
diagnostics12061390
Academic Editor: Michael Henein
Received: 21 March 2022
Accepted: 2 June 2022
Published: 4 June 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Predicting CT-Based Coronary Artery Disease Using Vascular
Biomarkers Derived from Fundus Photographs with a Graph
Convolutional Neural Network
Fan Huang
1
, Jie Lian
1
, Kei-Shing Ng
1
, Kendrick Shih
2
and Varut Vardhanabhuti
1,
*
1
Department of Diagnostic Radiology, LKS Faculty of Medicine, The University of Hong Kong,
Hong Kong, China; fhuang@hku.hk (F.H.); jlian@connect.hku.hk (J.L.); dougng@hku.hk (K.-S.N.)
2
Department of Ophthalmology, LKS Faculty of Medicine, The University of Hong Kong, Hong Kong, China;
kcshih@hku.hk
* Correspondence: varv@hku.hk; Tel.: +852-2255-3307
Abstract:
The study population contains 145 patients who were prospectively recruited for coronary
CT angiography (CCTA) and fundoscopy. This study first examined the association between retinal
vascular changes and the Coronary Artery Disease Reporting and Data System (
CAD-RADS
) as
assessed on CCTA. Then, we developed a graph neural network (GNN) model for predicting the
CAD-RADS as a proxy for coronary artery disease. The CCTA scans were stratified by CAD-RADS
scores by expert readers, and the vascular biomarkers were extracted from their fundus images.
Association analyses of CAD-RADS scores were performed with patient characteristics, retinal
diseases, and quantitative vascular biomarkers. Finally, a GNN model was constructed for the task
of predicting the CAD-RADS score compared to traditional machine learning (ML) models. The
experimental results showed that a few retinal vascular biomarkers were significantly associated
with adverse CAD-RADS scores, which were mainly pertaining to arterial width, arterial angle,
venous angle, and fractal dimensions. Additionally, the GNN model achieved a sensitivity, specificity,
accuracy and area under the curve of 0.711, 0.697, 0.704 and 0.739, respectively. This performance
outperformed the same evaluation metrics obtained from the traditional ML models (p < 0.05). The
data suggested that retinal vasculature could be a potential biomarker for atherosclerosis in the
coronary artery and that the GNN model could be utilized for accurate prediction.
Keywords:
CAD-RADS; coronary artery disease; fundoscopy; fundus image analysis; graph convo-
lutional neural network
1. Introduction
Atherosclerosis is a chronic inflammatory disease of the arteries which is due to the
buildup of plaques adhering to the inner vessel wall. The early detection of atherosclerosis is
crucial for early treatment and prevention. However, current clinical diagnosis techniques,
such as coronary computed tomography angiography, tend to only identify the plaques at
their advanced stages rather than in the early stages. More importantly, medical imaging
is usually utilized when clear symptoms of atherosclerosis, such as acute chest pain, are
observed in high-risk patients. Early subclinical disease detection remains a challenge.
Hence, finding additional variables for the risk stratification or even the early detection of
atherosclerosis is needed.
Changes in the micro-vasculature, such as vessel nicking/narrowing, have been
recognized as early indicators for macro-vascular abnormalities. It was believed that the
common pathophysiologic processes of atherosclerosis may underlie both macro-vascular
and micro-vascular disease [1–3].
The retinal vasculature shares similar anatomical and physiological characteristics
with the coronary circulations [
4
–
6
]. It can be non-invasively acquired by a fundus camera
Diagnostics 2022, 12, 1390. https://doi.org/10.3390/diagnostics12061390 https://www.mdpi.com/journal/diagnostics
