June 2023
Volume 64, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2023
Optical coherence tomography for the prediction of major adverse cardiovascular events in AlzEye
Author Affiliations & Notes
  • Mark Chia
    Institute of Ophthalmology, University College London, London, London, United Kingdom
    Moorfields Eye Hospital NHS Foundation Trust, London, London, United Kingdom
  • Yukun Zhou
    Centre for Medical Image Computing, University College London, London, London, United Kingdom
    Moorfields Eye Hospital NHS Foundation Trust, London, London, United Kingdom
  • Siegfried Wagner
    Institute of Ophthalmology, University College London, London, London, United Kingdom
    Moorfields Eye Hospital NHS Foundation Trust, London, London, United Kingdom
  • Dominic Williamson
    Institute of Ophthalmology, University College London, London, London, United Kingdom
    Moorfields Eye Hospital NHS Foundation Trust, London, London, United Kingdom
  • Robbert Struyven
    Institute of Ophthalmology, University College London, London, London, United Kingdom
    Moorfields Eye Hospital NHS Foundation Trust, London, London, United Kingdom
  • Axel Petzold
    Institute of Ophthalmology, University College London, London, London, United Kingdom
    Queen Square Institute of Neurology, University College London, London, London, United Kingdom
  • Alastair K Denniston
    University Hospitals Birmingham NHS Foundation Trust, Birmingham, Birmingham, United Kingdom
  • Jugnoo Rahi
    Great Ormond Street Institute of Child Health, University College London, London, London, United Kingdom
    Great Ormond Street Hospital for Children NHS Foundation Trust, London, London, United Kingdom
  • Mario Cortina Borja
    Great Ormond Street Institute of Child Health, University College London, London, London, United Kingdom
  • Pearse Keane
    Institute of Ophthalmology, University College London, London, London, United Kingdom
    Moorfields Eye Hospital NHS Foundation Trust, London, London, United Kingdom
  • Footnotes
    Commercial Relationships   Mark Chia None; Yukun Zhou None; Siegfried Wagner None; Dominic Williamson None; Robbert Struyven None; Axel Petzold Novartis,, Code C (Consultant/Contractor), Heidelberg Engineering, Roche, Code R (Recipient); Alastair Denniston None; Jugnoo Rahi None; Mario Cortina Borja None; Pearse Keane Apellis, Code C (Consultant/Contractor), Allergan, Topcon, Heidelberg Engineering, Novartis, Roche, Bayer, Code F (Financial Support), Big Picture Medical, Code I (Personal Financial Interest)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2023, Vol.64, 547. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Mark Chia, Yukun Zhou, Siegfried Wagner, Dominic Williamson, Robbert Struyven, Axel Petzold, Alastair K Denniston, Jugnoo Rahi, Mario Cortina Borja, Pearse Keane; Optical coherence tomography for the prediction of major adverse cardiovascular events in AlzEye. Invest. Ophthalmol. Vis. Sci. 2023;64(8):547.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : The human retina can deliver insights into a range of systemic health conditions. As the leading cause of death globally, cardiovascular disease represents a particularly promising area of research. Few studies have harnessed the three-dimensional imaging capabilities of optical coherence tomography (OCT) to explore this link. We therefore developed a deep learning model to predict the three-year incidence of major adverse cardiovascular events (MACE) using retinal OCT.

Methods : AlzEye is a retrospective cohort study linking retinal imaging of patients aged ≥40 years with systemic disease data from hospital admissions between January 2008 and March 2018. MACE was defined as ischaemic stroke (I63-I64), myocardial infarction (I21-I22), heart failure (I50), and atrial fibrillation (I48), according to the International Classification of Disease, 10th revision. We used the left retinal OCT from a single visit for each patient, and split the dataset into train, validation, and test sets in the ratio 55:15:30. Two models were developed using distinct pre-training strategies to allow comparison: (1) an in-house strategy trained on unlabelled natural and retinal images, and (2) a baseline strategy trained on labelled natural images from ImageNet-21K. In both cases, fine-tuning was performed on the AlzEye training set and evaluated on the internal test set.

Results : Within AlzEye, 5,382 patients had a MACE incident within three years of undergoing retinal OCT. The primary model using our in-house strategy achieved an AUROC of 0.796 [95% CI: 0.795, 0.797] and AUPR of 0.769 [95% CI: 0.768, 0.771]. The F1-score and sensitivity was 0.713 [95% CI: 0.711, 0.714] and 0.700 [95% CI: 0.699, 0.702], respectively. Our in-house pre-training strategy significantly outperformed the supervised baseline strategy, which achieved an AUROC of 0.743 [95% CI: 0.741, 0.744], AUPR 0.713 [95% CI: 0.710, 0.715], F1-score 0.660 [95% CI: 0.659, 0.662], and sensitivity 0.637 [95% CI: 0.635, 0.639].

Conclusions : Our deep-learning model showed promising performance for the task of three-year incident MACE prediction using OCT. Our in-house pre-training strategy significantly outperformed the baseline supervised strategy. Subsequent research will explore multimodal prediction models, external validation of our findings, and the potential for clinical translation.

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×