Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

Google-owned Verily Can Predict Heart Attacks By Looking in your Eyes


Eyes can reveal what’s going on in your heart….

Alphabet’s life sciences subsidiary Verily has developed a system that harnesses machine learning to predict a person’s risk of heart disease— simply by looking at their eyes.

Just by processing scans of the back  of a patient’s eye, the software can accurately glean information about that person’s age, blood pressure and smoking habits, which can then be further assessed to predict how likely they are to have a heart attack.

The algorithm was trained on a vast dataset of 300,000 patients— which included images of patient’s eyes and their medical data. By sifting through the information, Verily’s system spotted patterns linking feature visible  in the eye scans with characteristics (such as age and blood pressure) that make a person more prone to suffer from heart disease.

The back of the eye— scientific name: fundus— is particularly information-rich: as it is shot through by a huge quantity of blood vessels providing a good way to keep tabs on blood pressure and other key indicators.

Verily’s method— presented in a paper on Nature’Biomedical Engineeringis only slightly less accurate than current state-of-the-art diagnostic technique SCORE, which ranks likelihood of cardiovascular diseases by means of a blood test. SCORE can correctly predict risk about 72 percent of the time, whereas Verily’s technology has an accuracy of 70 percent.

Verily’s test, though, would have the advantage of being quicker and easier to administer than current methods. Many doctors rushed to underline that the technique will have to undergo several more tests before it can be considered a trustworthy  diagnostic procedure.

Even in that case, flesh-and-blood doctors should not be worried about losing their jobs to diagnosis-spewing algorithms. According to Verily, the AI’s aim is simply to help doctors by enhancing existing technology

“[This algorithm is] taking data that’s been captured for one clinical reason and getting more out of it than we currently do,” machine learning expert Luke Oakden-Rayner said in an interview with The Verge. “Rather than replacing doctors, it’s trying to extend what we can actually do.”

Image via


Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter