Atrial fibrillation causes uneven blood flow, which results in subtle changes in facial skin color. Researchers were able to use that to diagnose AF patients in a small, pilot study by employing video of a person's face in conjunction with a software analysis of skin color.
Eventually, this could translate into a diagnostic tool that complements a standard electrocardiogram (ECG) or even enables remote diagnosis. It could also allow for continuous monitoring of patients at risk for or diagnosed with AF while they use a computer, smartphone or tablet.
The study used a web camera to record a person's face and the subtle heartbeat-to-heartbeat variations in the skin color in 11 subjects who had been referred for electrical cardioversion, a series of low-voltage electrical shocks used to restore a normal heart rhythm, before and after the procedure. While simultaneously recording the video, researchers also conducted an ECG before and after electrical cardioversion.
The researchers quantified the pulse variability detected in the facial videos and used it to diagnose AF. Use of this video measurement of pulse variability had a 20% error detection rate, while ECG measurements had a 17% to 29% error rate. The University of Rochester School of Medicine and Dentistry and Xerox developed the technology in partnership; the study was published in the journal Heart Rhythm.
"This technology holds the potential to identify and diagnose cardiac disease using contactless video monitoring," Jean-Philippe Couderc with the University of Rochester Medical Center's Heart Research Follow-up Program said in a statement. "This is a very simple concept, but one that could enable more people with atrial fibrillation to get the care the care they need."
Globally, 33.5 million people have AF, but it's estimated that 30% of those with the condition don't know they have it. The fatigue and weakness associated with it often come and go. If left untreated, AF is associated with potentially fatal blood clots or stroke.
The video scan requires the subject to remain still for 15 seconds. Sensors in digital cameras record three colors: red, green and blue. Hemoglobin, a component of blood, absorbs more of the green spectrum light. This change can be detected by the camera's sensor and the face is the ideal place to look, since the skin is thinner and blood vessels are closer to the surface.
The researchers are conducting a larger trial that will include people without atrial fibrillation. The expectation is that these new studies, along with the application of image stabilizing technology and the improvement of camera resolution, will reduce the error rate.
- here is the study and the press release