Using the eye to predict risk for cardiovascular disease

: the living lab

Abstract

Introduction
Cardiovascular disease is associated with both structural and functional alterations and remodeling of the microcirculation. Chapter 1 (along with papers 1, 2 and 3) outline the investigation of conjunctival microvascular function as a strategy that may support cardiovascular risk prediction, as the microvasculature of the conjunctiva is ideal for non-invasive imaging. The main aim of this thesis is to examine measurements of the conjunctival microcirculation for application in cardiovascular disease risk prediction.

Methods
The first study of this thesis is described in Chapter 2. It investigates prior knowledge obtained from retinal research to determine the methods used, and to investigate their potential application within the conjunctival microvascular imaging application to measure various additional ocular parameters. Chapter 3 detail methods of recruitment to include the controls and high-risk cardiovascular disease cohorts recruited such as coronary microvascular disease patients and severe aortic stenosis patients referred for transcatheter aortic valve implantation (TAVI). The slit-lamp at a magnification of x40 and iPhone 11 Pro at x2 zoom was used in this optimised 4K imaging system following calibration. Image acquisition involved 10 second videos of the left temporal, left nasal, right temporal and right nasal conjunctival vessel view being obtained and a minimum of 60 stable and high quality frames were required for image processing. Chapter 4 comprises of the processes such as Manhattan Scoring and Convolutional Neural Networks involved in the development of a fully automated application (app) for processing the conjunctival vessel videos.

Results
As discussed within chapter 5, alterations of conjunctival arteriolar axial and cross-sectional velocities (0.57±0.12 mm/s to 0.62±0.12 mm/s, p=0.006 and 0.40±0.08 mm/s to 0.44±0.08 mm/s, p=0.001) wall shear rate (161±71 s-1 to 194±94 s-1, p=0.019), as well as venular wall shear stress (7.98±2.50 dynes/cm² to 6.57±1.63 dynes/cm², p

Conclusion
The findings presented in this thesis and related publications suggest that this objective ocular imaging tool may be able to detect changes in haemodynamics due to valvular disease, macro and micro-vascular disease, as impacted by cardiac structure and rhythm, providing an indication of cardiovascular health. Further longitudinal and serial imaging is required to better assess and stratify risk within the general population. The COVID-19 research had an international impact, but future research could also consider the impact of COVID-19 on the cardiovascular system.

Date of Award	Dec 2022
Original language	English
Sponsors	Northern Ireland Chest Heart and Stroke & The Heart Trust Fund
Supervisor	Andrew Nesbit (Supervisor), Tara Moore (Supervisor), Jim McLaughlin (Supervisor) & Mark Spence (Supervisor)