Developing a New ‘EyePhone’ mHealth App for Stroke Detection

Strokes are one of the deadliest killers in America, ranking fifth among all causes of death nationwide. The widespread prevalence of strokes is driving an urgent need for technology that enables rapid and accessible detection of the disease. As the popularity of mHealth smartphone applications continues to grow, Johns Hopkins investigators have developed a new one for stroke detection: the EyePhone.  

A stroke occurs when the blood supply to part of the brain is blocked or when a blood vessel in the brain bursts, according to the Centers for Disease Control and Prevention (CDC). Some of the symptoms of stroke include sudden numbness or weakness in the face, arm, or leg, sudden confusion, and sudden trouble seeing in one or both eyes. The time between stroke detection and treatment is critical, as the best options for treatment are only available within three hours of the first symptoms.

Though stroke care is widely available in the US, with 91 percent of the population living within one hour of a stroke care facility, various factors can hinder people from receiving this care, such as a lack of public awareness of stroke symptoms, racial disparities, and misdiagnoses. 

In fact, research also shows that stroke was the top cause of serious harm from misdiagnosis. Misdiagnoses become even more common when patients present with vague symptoms.

“The misdiagnosis of a stroke is something around 9 percent, but for these strokes that present with non-specific symptoms like dizziness and imbalance, they are a misdiagnosed at a rate of like 40 percent — and this is very high,” said Ali S. Saber Tehrani, MD, assistant professor of neurology at the Johns Hopkins University School of Medicine, in an interview with mHealthIntelligence. “You can compare it with the rate of misdiagnosis for heart attacks, which is around 2 percent.”

One approach that can help accurately diagnose strokes is the head impulse, nystagmus, and test of skew (HINTS) eye examination. However, research shows that since these eye movements are very subtle, proper training is required to identify and assess them, resulting in a significant barrier to diagnosing strokes through eye movement evaluations, Tehrani noted.

In a paper published in the Journal of the American Heart Association last month, a research team led by Johns Hopkins Medicine detailed the development and testing of a smartphone eye-tracking app that aims to mitigate misdiagnosis of posterior circulation strokes, that is, strokes occurring in the back of the brain.

According to Tehrani, about 250,000 posterior circulation strokes occur every year, and thus, the potential of using an mHealth app to help diagnose this type of stroke is wide-reaching. The apps can be used to detect strokes in the home or emergency departments (EDs), enabling timely triage and treatment. It can also be used in follow-up care, allowing clinicians to check whether any imbalance or dizziness the patient is experiencing may actually be a stroke.

“[mHealth] applications have tremendous potential for affecting all levels of stroke care from the patient’s home all the way to their follow-up,” Tehrani said.

The idea for the app was born out of a stroke case close to home. Two of the researchers and study authors, NYU Langone Health’s Max Parker, MD, and Johns Hopkins University School of Medicine’s Nathan Farrell, BSc, had a friend who suffered a posterior circulation stroke at 18 because of a vertebral artery dissection.

“They were smart young individuals — young students who wanted to do something about it, and they did,” Tehrani said.

Along with Tehrani and other researchers, they developed the EyePhone app, which uses the embedded facial recognition capabilities of the iPhone’s front camera to record eye movements. The eye-tracking data are then used to determine and quantify nystagmus, a condition where the eyes move rapidly and uncontrollably, or skew deviation, a vertical misalignment of the eyes. The recorded eye movements can also be used to conduct the head impulse test, which is a clinical assessment technique used to assess the angular vestibulo-ocular reflex. This involuntary reflex stabilizes the visual field.

“These are important signs for a stroke,” Tehrani said. “So, you use eye tracking, you try to figure out what those eye-tracking data mean, and then you get to a result of stroke, no stroke, or something else.”

Depending on the type of test, patients are asked to record different eye movements via the app.

“If we want to test nystagmus, we just have the patient look. We don't want them to look at a target; they can just stare, and then we just record their eye movements at various locations,” said Tehrani. “But in other tests, such as skew deviation, we will ask the patient to pick a target on the wall and keep looking at it.”

In the recently published study, the researchers examined whether the EyePhone app can detect and quantify induced nystagmus similar to the video oculography (VOG) goggles typically used for nystagmus quantification.

The researchers enrolled ten volunteers, recorded the velocity of induced nystagmus using the EyePhone app, and then compared the results with the VOG goggles. The EyePhone-recorded velocities highly correlated with the VOG recordings, with a correlation of 0.98 for horizontal and 0.94 for vertical nystagmus velocity recordings.

Though there have been challenges in developing and testing the app so far — including redoing the design and setup for inducing nystagmus and one of the app developers leaving for a job at Meta — the development process is ongoing.

Next, the research team will focus on further testing using the back camera, making the app more user-friendly, and doing some more feasibility testing, with eventual plans to pursue Food and Drug Administration approval.

With the advancements in digital health technology and broader clinical adoption of these tools, mHealth apps are playing a prime role in changing traditional diagnosis and treatment processes.

This is what Tehrani ultimately hopes the EyePhone app will be able to provide: a faster and more accessible route to stroke detection, diagnosis, and treatment.

“It's a great low-hanging fruit first step,” he said. “A first step in the hands of ED physicians or ED neurology consultants that would be very helpful. [Then] in the hands of primary care physicians and patients at home.”

Editor's note: A previous version of this article mistakenly attributed the development of the Eyephone app to Cedars-Sinai investigators in the first paragraph. The article was corrected and updated on March 1.