AI-Enabled Retinal Vasculometry Provides Alternative Biomarker of Vascular Health

Professor Alicja Regina Rudnicka

New research indicated that artificial intelligence (AI)-enabled retinal vasculometry is accurate in predicting cardiovascular disease (CVD) and death, without the need for blood tests or blood pressure (BP) medicine.

Based on these findings, clinicians may have a greatly effective, non-invasive screening test for individuals at medium to high risk of circulatory disease that can be done outside of the clinic, suggest the investigators.

“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 [BP measurement] are not needed,” wrote study investigator Professor Alicja Regina Rudnicka, Population Health Research Institute, St George's University of London.

Diseases of the circulatory system including cardiovascular disease, coronary heart disease, heart failure, and stroke are major causes of illness and death worldwide. Several risk frameworks exist, but they are not always able to accurately identify patients who will develop or die of circulatory diseases.

Prior findings suggest the width of the arterioles and venules of the retina could help early, accurate prediction of circulatory diseases. However, it is unclear if this will apply both consistently and equally to men and women.

Researchers in the current study created a fully automated AI-enabled algorithm named Quantitative Analysis of Retinal Vessels Topology and Size (QUARTZ) to develop models to assess the potential of retinal vasculature imaging plus known risk factors to predict vascular health and death.

Then, QUARTZ was applied to retinal images from 88,052 UK Biobank participants aged 40 - 69 years. It looked at the width, vessel area, and tortuosity of the arterioles and venules in the retina to develop prediction models for stroke, heart attack, and death from circulatory disease.

Models were subsequently applied to retinal images of 7411 participants aged 48 - 92 years of the European Prospective Investigation into Cancer (EPIC)-Norfolk study. The performance QUARTZ was compared with the Framingham Risk Score framework, both separately and jointly.

The study tracked the health of all participants for an average of 7 to 9 years. In that time, a total of 327 circulatory disease deaths were observed among 64,144 UK Biobank participants (average age, 56 years) and 201 circulatory deaths among 5862 EPIC-Norfolk participants (average age, 67 years).

Male predictors of death from circulatory diseases included arteriolar and venular width, tortuosity, and width variation, while in women, arteriolar and venular area and width and venular tortuosity and width variation contributed to risk prediction.

Investigators added the predictive impact of retinal vasculature on circulatory disease death interacted with smoking, BP medications, and previous heart attack. Overall, the predictive models were reported to have captured between half and two-thirds of circulatory disease deaths in individuals most at risk.

Data show retinal vasculature models captured about 5% more cases of stroke in UK Biobank men, 8% more cases in UK Biobank women, and 3% more cases among EPIC-Norfolk men most at risk, but nearly 2% fewer cases among EPIC-Norfolk women. Further, Framingham Risk Scores captured more cases of heart attack among those most at risk.

The addition of retinal vasculature to Framingham Risk Scores led to marginal changes in the prediction of stroke or heart attack, according to investigators. However, they added a simpler non invasive risk score based on age, sex, smoking, medical history, and retinal vasculature performed as well as the Framingham Risk Score, without the need to blood tests or BP measurement.

“In the general population, it could be used as a non-contact form of systemic vascular health check, to triage those at medium-high risk of circulatory mortality for further clinical risk assessment and appropriate intervention,” Rudnicka wrote.

The study, “Artificial intelligence-enabled retinal vasculometry for prediction of circulatory mortality, myocardial infarction and stroke,” was published in the British Journal of Ophthalmology.