Stroke, multiple sclerosis and Alzheimer’s disease each affect the brain in different ways — but it’s difficult for clinicians and researchers to truly understand just how, and when, while patients are still living.

The eyes offer valuable insights about these disease processes: clues that can be detected noninvasively and used to fine-tune diagnosis and improve the effectiveness of treatments. At Johns Hopkins, researchers are harnessing ophthalmic technology and pioneering new techniques to capitalize on those clues, paving the way for improved management of these diseases.

An EyePhone App for Better Stroke Diagnosis

Ali Tehrani

When patients arrive at an emergency department (ED) showing classic stroke symptoms like slurred speech and weakness in one side of the face or body, the misdiagnosis rate is about 4%. But when patients display symptoms like dizziness or lack of balance, which are also common in other conditions, the misdiagnosis rate jumps to 40%, notes Johns Hopkins neurologist Ali Tehrani.

Whenever a stroke goes undiagnosed, doctors miss the opportunity to intervene and minimize harm to a patient’s brain. Misdiagnosis also increases the chances that the patient will suffer another stroke in the ensuing days, doing further damage to the brain.

David Newman-Toker

David Newman-Toker, the director of Johns Hopkins Medicine’s Armstrong Institute Center for Diagnostic Excellence, helped develop a three-part eye exam called HINTS (short for Head-Impulse, Nystagmus and Test-of-Skew) in 2009. The HINTS exam measures subtle eye movements to distinguish benign vestibular disorders from more severe conditions like strokes. Newman-Toker and his colleagues later piloted the use of U.S. Food and Drug Administration-approved video goggles in the HINTS testing process, which produced far more accurate results than a doctor making a diagnosis with the naked eye.

“Successful implementation [of HINTS and related eye movement diagnostics] would likely result in improved quality of care for hundreds of thousands of peripheral vestibular patients and tens of thousands of stroke patients, as well as an estimated national health care savings of roughly $1 billion per year,” Newman-Toker and colleagues noted in a 2022 article in Seminars in Neurology.

But there’s one problem: Many EDs can’t afford a pair of the video goggles, which can cost between $35,000–$40,000, nor do they have a technician on staff who’s trained to operate them.

That’s when Newman-Toker and Tehrani started talking.

“We thought, everyone has a smartphone in their pocket. Could we use that to record eye movements?” Tehrani says.

From those musings, the duo developed the EyePhone app. First, a clinician points a smartphone camera at a patient’s eyes. Then, the app analyzes eye movements during the HINTS exam. Finally, the app produces measurements that can help clinicians determine whether the patient is having a stroke.

A study supported by a $2.5 million grant from the American Heart Association tested EyePhone’s HINTS exam capabilities against those of the video goggles. In a paper published in the Journal of the American Heart Association in January 2024, the team demonstrated that the app’s recordings correlated strongly with the goggles’.

“This technology, because it relies solely on the mobile phone and not attachments, tripods or other equipment, is potentially transformational,” Newman-Toker says. “It’s the difference between everyone being able to use it and only a tiny subset of people being able to use it.”

The duo are planning to submit the EyePhone app for FDA approval, meaning the tool is not cleared for use in clinical settings yet. Meantime, Tehrani and Newman-Toker plan to spend the next year improving the app’s interface through usability tests and questionnaires with ED physicians around the country. They are also working to connect the app to automated diagnostic systems typically found in EDs.

“Ideally, a physician will be able to look at a patient who is dizzy or having neurologic symptoms, point the phone, and not only will the app objectively measure eye movements but also determine whether those results are consistent with a stroke without the physician having to do anything further,” Tehrani says.

Steven Zeiler

Beyond its applications for stroke, Tehrani says, EyePhone is being studied as a tool to measure eye movements in patients with ALS. Eye-tracking technology has long been used to help late-stage ALS patients communicate with others, but it hasn’t been explored as a diagnostic or prognostic tool for the disease. In a 2024 paper published in Digital Biomarkers, Tehrani, Newman-Toker, Johns Hopkins neurologist Steven Zeiler and their collaborators argue that continuing this line of research could identify “valuable biomarkers that could shape the future of ALS assessment and resource allocation.”

“This app, if FDA-approved, can be used as a biomarker for disease progression in many cases,” Tehrani says. “I think that the future is bright. There’s lots of room for collaborations, whether for initial stroke diagnosis or potentially, response to ALS treatment.”

A New, Noninvasive Way to Diagnose MS

Multiple sclerosis is notoriously difficult to diagnose. There isn’t a silver bullet test to determine whether a person has the disease, and its early symptoms can mimic those of many other conditions.

Since 2001, the McDonald criteria have served as a guide for physicians when diagnosing MS. But it wasn’t until recently that those criteria included symptoms that are visible in eye imaging.

In late 2024, lesions in the optic nerve were added as a “fifth separate topology or affected area” in the McDonald criteria. This change, noted an article by the Multiple Sclerosis Academy, “should decrease the risk of misdiagnosis and decrease time to diagnosis, allowing for more rapid management to be put in place.”

Peter Calabresi

The change occurred in large part because of research by Johns Hopkins’ Peter Calabresi and his long-standing collaborator, Shiv Saidha, professor of neurology. Calabresi, co-director of the Johns Hopkins Multiple Sclerosis Precision Medicine Center of Excellence, was among the first researchers in the world to establish that an ophthalmic technology, optical coherence tomography (OCT), could help researchers and clinicians identify optic nerve neuropathy.

Shiv Saidha

The retina and optic nerve in the back of the eye are part of the central nervous system. Calabresi, Saidha and their team asserted in 2008 that damage to those areas can occur asymptomatically and serve as analogs for brain atrophy and progression of MS. Optic neuropathy — a frequent clinical feature of MS — can be present with as little as a 4-micron loss of retinal thickness in one eye.

Without technology like OCT, Calabresi says, it’s difficult for even experienced doctors to detect that kind of change. OCT, which is like a light-based ultrasound of the back of the eye, can catch such microscopic changes in tissue. It’s also noninvasive, which means patients tolerate it well.

In the years since establishing OCT’s utility for detecting brain atrophy in MS, Calabresi and his team have studied the ways the technology can be used to monitor MS disease progression and aid in prognosis. OCT results can also help doctors and patients make important decisions about courses of treatment. It’s even become a routine clinical procedure for patients with MS who are treated at The Johns Hopkins Hospital.

Calabresi says that OCT equipment is getting faster, easier to use, and less expensive, which means more clinicians have access to this powerful tool. The quality of the imaging technology is improving, too, allowing researchers to see changes in deeper layers of the retina. He and his team are continuing to learn new things about the back of the eye, and that’s leading to new areas of research in the relationship between the organ and MS.

For example: Calabresi is studying whether the eyes can help researchers better understand MS’ heterogeneous nature. Some patients live with MS for 50 years and show no outward symptoms or impairments. But in others, the disease progresses so quickly that a person can go from fully functional to a wheelchair in five years.

Calabresi’s team is currently using OCT to investigate whether genetics plays a role in how MS severity can vary so widely between people. Such a connection would be transformative for MS treatment, because identifying a gene that is associated with more severe MS would help identify patients at risk for that type of disease faster. That could lead to more aggressive or targeted treatment and, ideally, a slower progression of symptoms. Calabresi says the team has already found a gene that looks promising.

“This is just one example of how this technology can help the people with MS right in front of you, clinically, but can also help many others through research,” he says. “OCT is allowing us to ask new research questions that will lead to the therapies of the future.”

An Early Window Into Alzheimer’s

Researchers have studied the connection between Alzheimer’s disease and the eyes for decades. This research, however, is complicated by the fact that late-onset Alzheimer’s disease (LOAD) patients are elderly and have comorbidities. It’s difficult to pinpoint whether eye damage, particularly to the retina, is from Alzheimer’s or hypertension, diabetes or glaucoma.

Patients with autosomal-dominant Alzheimer’s disease (ADAD), however, offer a vital new pathway in this research.

ADAD is inherited and extremely rare, accounting for only 1% of Alzheimer’s cases, but it is caused by mutations in proteins that are implicated in LOAD. ADAD patients develop symptoms at a much younger age than typical Alzheimer’s patients, sometimes as early as 25. Amir H. Kashani, the Boone Pickens Professor of Ophthalmology at Johns Hopkins’ Wilmer Eye Institute, is one of a handful of researchers in the world who focus on the ADAD population.

“These patients provide a unique opportunity to see what’s purely Alzheimer’s and what is age-related or other disease-related changes,” he says.

In two separate studies, Kashani and his team have identified potential biomarkers in the eye that correlate with what’s happening in the brain during Alzheimer’s disease.

In one, Kashani used optical coherence tomography angiography (OCTA) to measure blood flow and vessel density in the retina. His team’s findings, published in a 2021 Alzheimer’s & Dementia article, showed that ADAD patients had increased capillary blood flow in the retina well before they showed any outward Alzheimer’s symptoms.

In another study, Kashani used OCT and diffusion tensor imaging (DTI) to measure retinal thickness in the eyes of ADAD patients. The results — published in a 2023 Alzheimer’s & Dementia article — showed that ADAD patients had thinning in retinal cell layers that significantly correlated with optic nerve changes.

Taken together, these studies have exciting implications for Alzheimer’s diagnosis and treatment. Alzheimer’s is typically diagnosed with invasive techniques and only after symptoms develop. By then, the brain has already deteriorated. But by using noninvasive retinal capillary measurements and imaging to screen patients, clinicians may better predict who’s at a greater risk for developing Alzheimer’s and suggest earlier therapeutic interventions to slow the disease’s progression.

Kashani cautions that there are hurdles to clear before these findings can be applied clinically, though.

First, these studies are limited by their small population size. There are other groups around the world who are studying ADAD patients, Kashani explains, but studies of the eyes in these patients are relatively uncommon.

“One idea is to establish a consortium of investigators to carefully study these retinal findings in larger cohorts,” he says. “The papers we’ve published have hopefully raised awareness that this is something worth looking into very early in the disease, and that we need to do a multicenter study to do this work with more people.”

Second, detecting these biomarkers requires sophisticated equipment that many clinics do not have. Until recently, Kashani says, technology powerful enough to measure capillary blood flow and retinal thickness has not been widely available in a clinical setting.

“With retinal thickness, we’re potentially looking at just a few micron changes in thickness. This level of precision — while achievable in a state-of-the-art research center — is pretty tough to do,” he says. “Developing clinical devices that have higher resolution will make it easier for more people to identify those differences.”

OCT and other technologies are growing more powerful and more accessible every day. Kashani thinks a continued focus on the eyes will transform the way researchers and clinicians approach Alzheimer’s disease. And he’s confident that the discoveries have only just begun.

What makes him so optimistic?

Years ago, when Kashani was working with a patient with ADAD, he noticed something in the patient’s eye imaging that he believed were amyloid lesions — a hallmark pathological feature of Alzheimer’s disease. 

“Those have been described in people who’ve passed away, but no one had looked at the retina (in vivo) and said, ‘These lesions are related to Alzheimer’s,’” Kashani says. Two years later, when the patient unfortunately died of natural causes, histopathology of the eyes confirmed that what Kashani saw in those images were, in fact, amyloid lesions around the blood vessels and neurosensory retina.

“This is only one example,” he says, “but we think that this is something that, if people begin to look for it, they’ll see it in many more patients even earlier in disease.”