Cardiology Research, ISSN 1923-2829 print, 1923-2837 online, Open Access
Article copyright, the authors; Journal compilation copyright, Cardiol Res and Elmer Press Inc
Journal website https://www.cardiologyres.org

Review

Volume 12, Number 3, June 2021, pages 132-139

Heart, Eye, and Artificial Intelligence: A Review

Table

Table 1. Studies Using Retinal Imaging to Predict Cardiovascular Disease Characteristics
 
StudyCountryNumberPopulationDeep learning involvementRetinal metric(s)CV outcomeResults
MI: myocardial infarction; ICA: internal carotid artery; SBP: systolic blood pressure; DBP: diastolic blood pressure; BMI: body mass index; AUC: area under the ROC curve; CV: cardiovascular; CVD: cardiovascular disease; DL-FAS: deep learning-funduscopic atherosclerosis score; FRS: Framingham risk score; OR: odds ratio; CI: confidence interval; CAC: coronary artery calcification; AVR: arterial-venous ratio.
Cheung et al, 2007 [24]USA4,593Men and women aged 45 - 84 years without history of clinical CV diseaseNoneRetinal vascular caliber, retinopathyLeft ventricular concentric remodelingNarrower retinal arteriolar caliber (OR: 2.06) and retinopathy (OR: 1.31) associated with increased odds of concentric remodeling
Wong et al, 2008 [23]USA6,814Men and women aged 45 - 84 years without history of clinical CV diseaseNoneRetinopathy, retinal arteriovenous nicking, retinal arteriolar caliber, retinal venular caliberCACPresence of retinopathy was associated with presence of any CAC (OR: 1.22, 95% CI: 1.04 - 1.43); retinal arteriovenous nicking weakly associated with higher CAC scores; no significant associations of retinal arteriolar or venular caliber with CAC scores.
Kromer et al, 2018 [29]Germany106Males who experienced MI before age 50 and age-matched healthy malesNoneCentral retinal vessel caliber, AVRMINo significant differences in central retinal arterial/venous equivalent or AVR between MI group and control group.
Poplin et al, 2018 [28]UK, USA, and other countries in EyePACS database297,360Patients from UK Biobank and EyePACS databaseUsed model to predict CV risk factors from retinal fundus imagesRetinal fundus imagesCV risk factors, major adverse cardiovascular events (MACEs) within 5 yearsModel significantly predicted certain CV risk factors within given margin (P < 0.0001): age within 5 years at 78% accuracy; SBP within 15 mm Hg at 72% accuracy; DBP within 10 mm Hg at 79% accuracy; BMI within 5 at 80% accuracy. Model achieved AUC 0.70 (95% CI: 0.648 - 0.740) of predicting 5-year MACE from retinal fundus images alone compared to AUC 0.72 for European SCORE risk calculator.
Li et al, 2019 [27]China162Patients with ICA stenosisNoneSubfoveal choroidal thickness, choroidal vascular indexICA stenosisSubfoveal choroidal thickness (P < 0.05) and choroidal vascular index (P = 0.001) lower in severe ICA stenosis group.
Chang et al, 2020 [31]Korea38,824Participants who completed exams at HPC-SNUH between 2005 - 2016 and received retinal fundus examUsed model to predict atherosclerosis and risk of CV death relative to FRSRetinal fundus imagesPresence of atherosclerosis (DL-FAS), risk of CVD mortalityModel predicted atherosclerosis with accuracy of 58.3% and demonstrated significant association between DL-FAS and CVD mortality.
Dabrowska et al, 2020 [25]Poland12088 patients with essential hypertension and 32 healthy participants matched in age and genderNoneRetinal capillary flowArterial stiffness (measured by pulse wave velocity)Lower retinal capillary flow in patients with pulse wave velocity (PWV) > 10m/s compared to those with PWV ≤ 10 m/s (P = 0.02).
Druckenbrod et al, 2020 [26]USA143US veterans aged 29 - 91NoneSubfoveal choroidal thicknessCV disease risk factorsDiabetes associated with thinner choroid (P = 0.001). Hypertension (P = 0.006) and hyperlipidemia (P = 0.05) associated with thicker choroid.
Zhang et al, 2020 [30]China625Participants aged 24 - 83 yearsUsed model to predict CV risk factors from retinal fundus imagesRetinal fundus imagesHypertension, hyperglycemia, dyslipidemia, and other CV risk factorsModel predicted hypertension, hyperglycemia, and dyslipidemia with accuracies of 68.8%, 78.7%, and 66.7%, respectively.