‘Selfies’ that save lives: AI detects heart disease using photographs

We currently live in the era of the “selfie,” a time when children and adults alike seem to take pictures of themselves whenever the opportunity arrives. But selfies are good for more than impromptu photo sessions and posting on social media—they can also be used to identify signs of heart disease.

That’s what a group of researchers out of China has determined, sharing its findings in the European Heart Journal.

“To our knowledge, this is the first work demonstrating that AI can be used to analyze faces to detect heart disease,” corresponding author Zhe Zheng, department of cardiovascular surgery at Fuwai Hospital, said in a statement. “It is a step towards the development of a deep learning-based tool that could be used to assess the risk of heart disease, either in outpatient clinics or by means of patients taking ‘selfies’ to perform their own screening. This could guide further diagnostic testing or a clinical visit.”

Zheng et al. used photos of more than 5,000 patients to train their AI model. All patients were treated at one of eight hospitals in China from July 2017 to March 2019. Photos of another 580 patients—taken from the same hospitals during the same time period—were set aside to validate the model.

The team’s algorithm was then tested on another 1,000 patients who were enrolled from April to July 2019—and it out-performed various other methods typically used to predict a person’s heart disease risk.

The study did have certain limitations—including a low specificity and the fact that it wasn’t tested on a larger, more diverse patient population—but the team’s findings could still be a significant step toward improved screening for a variety of potentially fatal conditions. The authors did note, however, that there are certain “ethical issues” related to this technology that must be taken seriously.

“We believe that future research on clinical tools should pay attention to the privacy, insurance and other social implications to ensure that the tool is used only for medical purposes,” Zheng said.

On Thursday, August 20, Cardiovascular Business highlighted a similar study that used smartphone cameras to detect diabetes. These handheld devices, it seems, are good for more than just texting silly emojis to everyone you know.