A new AI tool could predict the risk of heart disease and death through retinal images
A new study has found that an artificial intelligence (AI) tool that images the retina's network of veins and arteries can accurately predict a person's risk of cardiovascular disease and death in less than a minute.
Moreover, it makes it possible to employ a non-invasive screening method that doesn't have to be done in a clinic for those with a medium to high risk of developing circulatory diseases.
"This AI tool could let someone know in 60 seconds or less their level of risk," Professor Alicja Rudnicka, the lead author of the study, told the Guardian. "If someone learned their risk was higher than expected, they could be prescribed statins or offered another intervention," she said.
Scanning the width of veins and arteries in the retina
Circulatory diseases such as cardiovascular disease, coronary heart disease, heart failure, and stroke are the leading causes of illness and mortality globally. Researchers suggest that although numerous risk frameworks exist, these aren't always able to accurately identify those who will go on to develop or die of circulatory diseases.
The new study, on the other hand, demonstrates that the width of veins and arteries in the retina could indicate circulatory disease early and accurately. In this direction, researchers developed a fully automated AI-enabled tool called Quartz to evaluate the potential of retinal vasculature imaging plus known risk factors to predict vascular health and death.
They applied Quartz firstly to scan the retinal images of 88,052 people who are UK Biobank participants between the ages of 40 and 69, and later to investigate the retinal images of 7,411 participants of the European Prospective Investigation into Cancer (EPIC)-Norfolk study, who were aged between 48 and 92.
The health of the subjects has been tracked for seven to nine years. The results show that the width, curviness, and width variation of veins and arteries in the retinas are important predictors of death from circulatory disease in men, while in women, artery area and width and vein curviness and width variation contributed to risk prediction.
Additionally, researchers discovered a strong correlation between the retinal data calculated by Quartz and cardiovascular disease, deaths, and strokes.
"AI-enabled vasculometry risk prediction is fully automated, low cost, non-invasive, and has the potential for reaching a higher proportion of the population in the community because of 'high street' availability and because blood sampling or [blood pressure measurement] are not needed," added the researchers.
The study was published in the British Journal of Ophthalmology.
Aims: We examine whether inclusion of artificial intelligence (AI)-enabled retinal vasculometry (RV) improves existing risk algorithms for incident stroke, myocardial infarction (MI) and circulatory mortality.
Methods: AI-enabled retinal vessel image analysis processed images from 88 052 UK Biobank (UKB) participants (aged 40–69 years at image capture) and 7411 European Prospective Investigation into Cancer (EPIC)-Norfolk participants (aged 48–92). Retinal arteriolar and venular width, tortuosity and area were extracted. Prediction models were developed in UKB using multivariable Cox proportional hazards regression for circulatory mortality, incident stroke and MI, and externally validated in EPIC-Norfolk. Model performance was assessed using optimism adjusted calibration, C-statistics and R2 statistics. Performance of Framingham risk scores (FRS) for incident stroke and incident MI, with addition of RV to FRS, were compared with a simpler model based on RV, age, smoking status and medical history (antihypertensive/cholesterol lowering medication, diabetes, prevalent stroke/MI).
Results: UKB prognostic models were developed on 65 144 participants (mean age 56.8; median follow-up 7.7 years) and validated in 5862 EPIC-Norfolk participants (67.6, 9.1 years, respectively). Prediction models for circulatory mortality in men and women had optimism adjusted C-statistics and R2 statistics between 0.75–0.77 and 0.33–0.44, respectively. For incident stroke and MI, addition of RV to FRS did not improve model performance in either cohort. However, the simpler RV model performed equally or better than FRS.
Conclusion: RV offers an alternative predictive biomarker to traditional risk-scores for vascular health, without the need for blood sampling or blood pressure measurement. Further work is needed to examine RV in population screening to triage individuals at high-risk.