TL;DR:
- AI technology may help forecast MACE using Optical Coherence Tomography.
- A recent study presented at the 2023 ARVO Annual Meeting showed the potential of deep-learning models in predicting MACE using retinal OCT scans.
- The study was led by Dr. Mark Chia of the University College London Institute of Ophthalmology and used an in-house pre-training strategy that outperformed the baseline supervised strategy.
- The retina holds key insights into a patient’s overall health, including the prediction of CVD, the leading cause of death globally.
- The study analyzed retinal imaging and systemic disease data from over 5,000 patients aged 40 and above, defining MACE as ischemic stroke, myocardial infarction, heart failure, and atrial fibrillation.
- The deep-learning model performed “reasonably well” in predicting MACE within three years.
- The team suggests expanding this work to other populations holds significant potential, with the promise of identifying high-risk patients during routine eye exams.
Main AI News:
A recent study presented at the 2023 Association for Research in Vision and Ophthalmology (ARVO) Annual Meeting in New Orleans has revealed the potential of deep-learning models in predicting major adverse cardiovascular events (MACE) using optical coherence tomography (OCT). The research team, led by Dr. Mark Chia of the University College London Institute of Ophthalmology, utilized an in-house pre-training strategy that outperformed the baseline supervised strategy, demonstrating the model’s effectiveness in predicting MACE within three years.
The study drew upon the growing body of evidence that the retina holds key insights into a patient’s overall health, including the prediction of cardiovascular disease (CVD), the leading cause of death globally. Despite the potential, few studies have explored the use of the 3-dimensional imaging capabilities of OCT in predicting MACE. The research team addressed this gap by developing a deep-learning model using retinal OCT scans.
The study, conducted using data from the AlzEye study, analyzed retinal imaging and systemic disease data from over 5,000 patients aged 40 and above. MACE was defined as ischemic stroke, myocardial infarction, heart failure, and atrial fibrillation according to the International Classification of Disease, 10th revision. The dataset was split into a train, validation, and test sets in a 55:15:30 ratio, and two models were developed using pre-training strategies for comparison.
The results showed that the deep-learning model performed “reasonably well” in predicting MACE within three years. The team suggests that expanding this work to other populations holds significant potential, with the promise of identifying high-risk patients during routine eye exams.
Conlcusion:
The advancements in AI and deep-learning technology for predicting MACE using Optical Coherence Tomography (OCT) hold great promise for the medical industry. The recent study presented at the 2023 ARVO Annual Meeting has shown the potential of retinal OCT scans in predicting MACE within three years, with the deep-learning model performing “reasonably well” in predictions. The retina’s role in providing insights into a patient’s overall health, including the prediction of cardiovascular disease, underscores the significance of this research.
The potential for identifying high-risk patients during routine eye exams could greatly benefit the medical industry, and the team suggests expanding this work to other populations. This research opens up new opportunities for the development of innovative solutions in the medical industry and holds great potential for the future of health care.
