Google researchers announce technology that can predict heart disease simply by scanning the eyes

byMustafa ezz

It is a related subsidiary of GoogleVerilyResearchers developed a technique different from the conventional inspection method, "Scanning the eyes to know the risk of heart disease". In this method, from the scanned data, the software presumes the patient's age, blood pressure, smoker or smoker, etc., and predicts the risk of heart disease such as heart attack from here.

Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning | Nature Biomedical Engineering
https://www.nature.com/articles/s41551-018-0195-0

Google's new AI algorithm predicts heart disease by looking at your eyes - The Verge
https://www.theverge.com/2018/2/19/17027902/google-verily-ai-algorithm-eye-scan-heart-disease-cardiovascular-risk

In our eyes, Google's software sees heart attack risk - The Washington Post
https://www.washingtonpost.com/news/the-switch/wp/2018/02/19/google-used-artificial-intelligence-to-predict-heart-attacks-with-the-human-eye/

According to researchers, the method of "scanning the eyes" can predict heart disease with the same accuracy as existing methods. It is thought that doctors can analyze patients more easily than ever before because blood tests are unnecessary, but it seems that it is necessary to conduct further tests before being actually used.

To make software algorithms, Google and Verily scientists conducted machine learning using data from nearly 300,000 patients. The data included not only general medical data but also eye scan data. And the neural network searched for patterns of "hidden information" necessary for diagnosing cardiac disorder understood from the eye from the information.


Although the method of "predicting heart disease by scanning the eyes" seems strange, the back side inside the eyeball "Fundus"Blood vessels are densely packed and it has been done as" fundus examination "so far as to show the state of the whole body. Google's technology is a development of this technique.

Until now, a method called "risk score" is used for predicting heart disease, and it is said that the accuracy of prediction is 72%. The prediction of the new method developed by Google is 70% accurate, so it can be diagnosed with the same accuracy as the existing method, although slightly inferior. Professor Alun Hughes of University College London, who studies Cardiovascular Physiology & Pharmacology "The research to predict the risk of heart disease from the retina has a long history and that our approach is reliable" . Also, although artificial intelligence has the potential to accelerate existing medical analysis methods, we warned that the algorithm needs to repeat the test until it reaches a reliable level.