Near Future: Send a selfie to your doctor for Coronary Artery Disease (CAD)

By Lisa O’Keefe

If you are like me, you prefer any diagnostic test that is non-invasive over one that requires something be introduced into your body. So, when I read in the European Heart Journal that in the near future it will be possible to send a selfie to a doctor and have them detect heart disease I had to know more.  And the quick answer is that deep learning computer algorithms may be the key to making this a reality.

So, this is how it would work: 

(1)    First, you would need to answer some basic questions on lifestyle choices, medical history and social economic status. Then you would  

(2)    have 4 photos taken of your face – one frontal, two profiles and one view of the top of the head. These photos would be analyzed  by AI to determine features that  are known to be associated with increased coronary artery disease (CAD) such as grey or thinning hair, wrinkles, ear lobe creases, small yellow deposits of cholesterol underneath the skin (usually around the eyelids) and fat and cholesterol deposits that appear as a hazy white, grey or blue opaque ring in the outer edges of the cornea.    

Now I know what you are thinking – all of us over 60 have at least one of those facial features so wouldn’t we all then be classified with heart disease?  Believe it or not – no. 

In a recent study, 6800 people took part in this research and according to Zhe Zheng  the Vice Director of the National Center for Cardiovascular Diseases, the AI analysis was correct 80% of the time in detecting CAD.  This is significant as 80% accuracy is better than the two most predictive diagnostic tools we have to date – those being the  Diamond-Forrester (DF) model which is predictive of obstructive coronary artery disease (CAD) and is often used to risk stratify acute chest pain patients and the CAD consortium clinical (CADCC) score which has superior diagnostic accuracy for obstructive CAD in patients referred for coronary CTA or invasive angiography.

Zhe Zheng believes this selfie screening tool is a positive step in the right direction as it is an inexpensive, non-invasive way to identify patients who may need follow-up tests and procedures. 

However as good as this all sounds, at present, there are three concerns you should be aware of with this tool:

(1)    The false-positive rate is high which in turn could result in increased cost to the patient who will be advised to go for further tests or procedures as well as increased stress to the patient who might have elevated blood pressure as a result.

(2)     The testing of this diagnostic tool needs to be applied to a larger population for validity.

(3)    In today’s world where many for-profit companies are paying for facial photos there is the possibility that the photos may be misused and therefore this gap will need to be closed to ensure privacy.

Lisa O'Keefe

Lisa O'Keefe

Lead columnist MedTech Visionaries
Editor AgeDisruptor magazine
Leftbrain Marketing L4S