Artificial intelligence (AI) continues to amaze us with its capabilities and now it is taking on a new role in the healthcare sector. Scientists have introduced RETFound, which is said to be a groundbreaking AI tool that can predict the risk of various health conditions, including heart failure, by analyzing eye scans.
RETFound stands out as a potent AI instrument and capable of diagnosing as well as predicting the probability of various health conditions, spanning from ocular disorders to heart ailments and Parkinson’s disease. Its special ability is to predict things by looking at pictures of the back of the eye called retinal images.
In a recent Nature journal publication, researchers detailed how RETFound was trained using an innovative self-supervised learning technique. Unlike traditional methods that require manual labeling of vast amounts of images, RETFound learned from a massive dataset of 1.6 million retinal images without the need for extensive human labeling. This not only made the AI tool more efficient but also cost-effective in the realm of medical diagnosis.
RETFound works somewhat like ChatGPT and other large language models. Just as ChatGPT predicts the next word in a sentence based on the context, RETFound scans retinal photographs and predicts missing parts of images with incredible accuracy.
Pearse Keane, an ophthalmologist at Moorfields Eye Hospital NHS Foundation Trust in London and a co-author of the paper, explained that over millions of images, the model learns what a retina looks like and comprehends its features. This foundational knowledge makes RETFound a versatile tool that can be tailored for various medical applications.
The paper also highlights how RETFound leverages retinal scans to gain insights into the cardiovascular health of a person. Retinas offer a unique view of the body’s tiniest blood vessels, forming a complex network. This means that conditions affecting the circulatory system like hypertension can be directly observed through retinal images.