Kidney Week

Abstract: TH-PO620

A Deep Learning System for Retinal Vessel Calibre Improves Cardiovascular Risk Prediction in Asians With CKD

Session Information

• Hypertension and CVD: Epidemiology, Risk Factors, Prevention
  November 03, 2022 | Location: Exhibit Hall, Orange County Convention Center‚ West Building
  Abstract Time: 10:00 AM - 12:00 PM

Category: Hypertension and CVD

• 1501 Hypertension and CVD: Epidemiology‚ Risk Factors‚ and Prevention

Authors

• Lim, Cynthia Ciwei, Singapore General Hospital, Singapore, Singapore

Cynthia Ciwei Lim,

• Chong, Crystal, Singapore Eye Research Institute, Singapore, Singapore

Crystal Chong,

• Tan, Gavin, Singapore Eye Research Institute, Singapore, Singapore

Gavin Tan,

• Cheng, Ching-Yu, Singapore Eye Research Institute, Singapore, Singapore

Ching-Yu Cheng,

• Sabanayagam, Charumathi, Singapore Eye Research Institute, Singapore, Singapore

Charumathi Sabanayagam,

Background

Cardiovascular disease (CVD) and mortality is elevated in chronic kidney disease (CKD). Retinal vessel calibre in retinal photographs is associated with CVD risk and automated measurements from retinal photographs by a deep learning system (DLS) may aid CVD risk prediction.

Methods

Retrospective cohort study of 838 Chinese, Malay and Indian participants aged 40-80 years with CKD (estimated glomerular filtration rate (eGFR) <60 ml/min/1.73 m²) at the baseline visit (2004-2011) of the population-based Singapore Epidemiology of Eye Diseases Study. Retinal vessel caliber measurements were generated by DLS: central retinal arteriolar equivalent and central retinal venular equivalent at Zone_B and Zone_C (CRAE_B, CRAE_C and CRVE_B, CRVE_C, respectively). Incidence of CVD (defined as non-fatal acute myocardial infarction (AMI) and stroke, and death due to AMI, stroke and heart failure) in those who were free of CVD at baseline was ascertained until 31st December 2018. Risk factors (traditional and retinal-DLS measures) were examined using Cox proportional hazards regression models. Improvement in risk prediction by addition of retinal vascular parameters was assessed by net reclassification improvement (NRI).

Results

Baseline mean age was 67.8 (SD 8.8) years and eGFR 47.3 (11.8) ml/min/1.73 m². Incidence of CVD was 30.3%. Multivariable regression models adjusting for traditional risk factors showed that eGFR and retinal parameters independently predicted 10-year CVD risk (Table 1). Addition of eGFR, retinal arteriolar calibre and retinopathy improved overall reclassification of incident CVD beyond traditional risk factors with NRI of 38.2% for Zone_B and 44.9% for Zone_C (p<0.001for both).

Conclusion

Addition of kidney function and retinal vessel caliber measurements by DLS improved CVD risk prediction among Asians with CKD.