Kidney Week

Abstract: FR-PO0017

Exploring the Utility of a Cardiovascular Risk-Predicting Retinal Artificial Intelligence (AI) Model in Identifying Severe Coronary Calcification in Patients with and Without CKD

Session Information

• Artificial Intelligence and Digital Health at the Bedside
  November 07, 2025 | Location: Exhibit Hall, Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Artificial Intelligence, Digital Health, and Data Science

• 300 Artificial Intelligence, Digital Health, and Data Science

Authors

• Thakur, Sahil, Mediwhale Inc., Gangnam-gu, Korea (the Republic of)

Sahil Thakur, MBBS, MS, PhD

• Nam, Dongjin, Severance Hospital, Seodaemun-gu, Seoul, Korea (the Republic of)

Dongjin Nam, MD

• Rukmini, Annadata V., Mediwhale Inc., Gangnam-gu, Seoul, Korea (the Republic of)

Annadata V. Rukmini, MD, MBBS, PhD

• Park, Junseok, Seoul National University College of Medicine, Jongno-gu, Seoul, Korea (the Republic of)

Junseok Park, MD

• Cho, Jungkyung, Seoul National University College of Medicine, Jongno-gu, Seoul, Korea (the Republic of)

Jungkyung Cho, MD

• Park, Tae Hyun, Mediwhale Inc., Gangnam-gu, Seoul, Korea (the Republic of)

Tae Hyun Park, MD

• Seo, Jaewon, Mediwhale Inc., Gangnam-gu, Seoul, Korea (the Republic of)

Jaewon Seo, MSc

• Nusinovici, Simon, Mediwhale Inc., Gangnam-gu, Korea (the Republic of)

Simon Nusinovici, PhD

• Rim, Tyler Hyungtaek, Mediwhale Inc., Gangnam-gu, Korea (the Republic of)

Tyler Hyungtaek Rim, MD, PhD, MBA

• Chun, Kyeong-hyeon, Gangnam Severance Hospital, Seoul, Korea (the Republic of)

Kyeong-hyeon Chun, MD

• Lee, Chan Joo, Severance Cardiovascular Hospital Department of Cardiology, Seodaemun-gu, Seoul, Korea (the Republic of)

Chan Joo Lee, MD, PhD

• Park, Sungha, Severance Cardiovascular Hospital Department of Cardiology, Seodaemun-gu, Seoul, Korea (the Republic of)

Sungha Park, MD, PhD

Background

Dr.Noon CVD is a retinal AI model for cardiovascular (CV) risk prediction, validated for estimating coronary calcification and 5-year CV events. Given the clinical relevance of a coronary artery calcium score (CACS)≥400 in guiding intensive prevention, this study assessed the model’s ability to identify such high-risk individuals, comparing performance between patients with and without chronic kidney disease (CKD).

Methods

We analyzed 1,130 CMERC-HI participants, classified into six groups by Dr.Noon CVD risk level (low/moderate [<41], high [41–51], very high [>51]) and CKD status based on eGFR (non-CKD: ≥60 mL/min/1.73m²; CKD: <60). Discriminatory performance for CACS≥400 was assessed using ROC analysis, with AUCs compared by CKD status. Adjusted odds ratios (ORs) were estimated via multivariable logistic regression, adjusting for traditional CV risk factors.

Results

Dr.Noon CVD showed comparable discrimination in the non-CKD (AUC=0.784) and CKD (AUC=0.798) groups (p=0.687). The prevalence of CACS≥400 increased from 3.6% in the non-CKD low/moderate-risk group to 42.3% in the CKD very high-risk group. In both CKD subgroups, the very high-risk group had significantly greater odds of CACS ≥400 compared to the high-risk group (adjusted OR=2.05 and 3.16; p=0.019 and <0.001, respectively)

Conclusion

Dr.Noon CVD accurately identifies individuals at increased risk of severe coronary calcification (CACS≥400) regardless of CKD status, supporting its clinical utility for targeted preventive strategies in both CKD and non-CKD populations.

Figure 1. Prevalence (A) and adjusted ORs (B) of CACS≥400 by Dr.Noon CVD risk group and CKD status

Funding

• Commercial Support – Mediwhale Inc., Seoul, South Korea

Digital Object Identifier (DOI)

• doi: 10.1681/ASN.2025kr2tp1h6