Eye Tracking Shows High Sensitivity As A Biomarker For Concussion, Study Finds
Robert Glatter, MD
Figuring out a way to rapidly diagnose a concussion has been a challenge that clinicians and researchers alike have been investigating for many decades. A multitude of definitions of concussion currently exist, but figuring out who truly has a concussion has been a vexing problem.
As no two patients are alike, the variability in signs and symptoms resulting from impacts to the body and head that produce concussive symptoms make it difficult to say who truly has a concussion after the initial impact.
As the eyes are literally the window to the brain, researchers have long recognized the potential that eye movements may provide early clues to neurologic dysfunction before other more serious signs of brain injury may become apparent.
And while the race to find an accurate and reliable blood marker to identify concussions has not been promising to date, the practice of using batteries of self-reported symptoms combined with balance testing and rapid number naming tests such as King-Devick have become accepted ways to identify concussions.
The increase in head injuries in youth sports leagues, possibly due to increased reporting and awareness, has further accelerated research to find a way to detect and monitor concussions.
In fact, the CDC reported that from 2001-2009, emergency department visits for youth sports-related traumatic brain injuries increased by 57%. Additional data from the CDC indicates that brain injury is the number one cause of death and disability in Americans under age 35. Of the 1.4 million annual traumatic brain injuries (TBI) in the United States, 50,000 persons die and another 235,000 will require hospital admission.
Many states have now passed laws mandating that youth athletes who are suspected of having a concussion must be examined by a licensed health care provider before being allowed to return to play. Such widespread state laws are the result of the Zachary Lystedt law enacted in Washington state in May of 2009, after Zachary suffered a life threatening injury after improperly returning to play before he was medically able.
But the real issue is that there is currently no single reliable test that serves as a gold standard to alert trainers, coaches, and medical professionals on the sidelines when a collision is serious enough to mandate removal from play.
We can describe the clinical effects of head trauma, and its resulting effects on orientation, memory and even thinking that may characterize a concussion. The only problem is that we don’t have a simple blood test or biomarker to objectively measure if a patient truly has a concussion, such as the way a patient might come to the emergency department with chest pain and receive an EKG and blood tests to evaluate if the pain is due to a heart attack.
Well, this may all be changing, thanks to new eye tracking technology that Dr. Uzma Samadani, Chairperson for Traumatic Brain Injury at Hennepin County Medical Center and an Associate Professor of Neurosurgery at the University of Minnesota, recently developed while working with brain injured patients.
In fact, promising results from Samadani’s team indicate that eye tracking can detect concussions with sensitivity that is comparable to that of simple blood tests used to diagnose heart attacks.
Unique eye tracking technology developed by Samadani, which follows patients’ eye movements and traces patterns, represents a more accurate measure of brain injury compared with current methods and clinical approaches to evaluate patients for concussions.
The research was published online August 7th in the Journal, Concussion.
Samadani, now with Hennepin County Medical Center and the University of Minnesota, in collaboration with researchers at the Steven and Alexandra Cohen Veterans Center at NYU Langone Medical Center, developed the technology that can serve as a biomarker for concussion by tracking patients’ eye movements as they watched music videos.
In her trial, subjects watched a music video for 220 seconds while eye movements were measured using a specialized tracking camera. Each eye’s movement was recorded as they watched a video and the eyes were then compared to their positions over time. Data obtained from these measurements was then used to differentiate between those persons with a concussion, and those who have a normal exam.
Her approach to defining a concussion was simple: develop a unique mathematical model that transforms data from a patient’s eye movements into a predictor for concussion status.
Her research group tested 255 subjects, including 8 persons who had concussions, and found that their eye tracking test had a sensitivity of 88% and specificity of 87%.
The encouraging news is that the eye tracker correctly identified 7 of the 8 persons with confirmed concussions, but unfortunately recorded false positives in up to 20% of healthy uninjured subjects. The impact of additional factors such as sleep deprivation, alcohol use, and other medications will need to be considered in future trials of the eye tracker.
A Concussion Biomarker
The ideal biomarker would have no false positives, but that would be quite difficult to achieve in reality. A biomarker that is not useful would have a 50% chance of being accurate, much like a coin toss.
It turns out that useful clinical tests typically have a sensitivity of greater than 80%. For example, cardiac biomarkers such as serum troponin have a sensitivity ranging in various studies from 76%-96%.
As a comparison, in this study, eye tracking achieved a sensitivity of 88%, quite comparable to tests used to screen for heart attacks.
However, as Samadani explains, validating any technology as a potential biomarker for concussion first requires an accurate definition for concussion itself. There are more than 42 published definitions that currently exist so the challenge is quite daunting to say the least.
“When doctors look for a biomarker for heart attack, it is relatively easy to check the accuracy of a potential candidate because they can perform a cardiac catheterization and confirm that the heart vessel is blocked and an attack has occurred,” said Samadani. “There is no analogous capability with brain injury – there is no gold standard diagnostic, no blood test, and no imaging study for definitively concluding that a patient has experienced a concussion.”
“We use symptom severity scales and standardized cognitive examination assessments but the imperfect nature of these may result in incorrect subject classification. Potentially, eye tracking may be more accurate than it appears, because of its objective appraisal of a complicated process of coordination that may be impaired,” added Samadani.
In Samadani’s research, concussion is defined as trauma to the head with a normal CT (computed tomography) scan of the brain, symptom severity score of 40 or greater on SCAT3 testing and standardized assessment of concussion (SAC) score less than 24.
The symptom severity score measures the self-reported severity of 22 concussion symptoms ranging from headache to dizziness and irritability. The SAC measures orientation, memory, and concentration – characteristics which can vary even among uninjured healthy control subjects. Self-reporting clearly makes the score less objective and reproducible.
In an accompanying editorial that also appears in the journal, Samadani predicts that eye tracking will help diagnose and classify brain injury and concussion, especially in patients with severe TBI with elevated intracranial pressure.
“The ultimate goal for brain injury” said Dr. Samadani, “is to achieve the same level of diagnostic capability and care as currently exists for other medical conditions.” “Right now when someone comes in to the emergency room with chest pain, doctors perform an EKG, blood test, imaging, and treatment.”
“With brain injury we need to be able to achieve the same level of care – to assess all aspects of the problem rigorously, classify, and treat accordingly. We already know that there is much more to brain injury than what is seen on a CT scan,” she explained.
It’s the unseen and unmeasured effects that have baffled researchers searching for the holy grail to identify a concussion at outset—that one could say “yes or no” after the initial impact occurs.
“Eye tracking tells us how well the brain is working regardless of how it looks,” said Samadani, “and represents the beginning of a solution to this problem. It is non-invasive, inexpensive and extremely quick. Testing does not require reading nor language skills which makes it useful for multiple patient populations.”
Dr. David Cifu, the Herman J. Flax, M.D. Professor and Chair of Rehabilitation at Virginia Commonwealth University and Senior TBI Specialist with the U.S. Department of Veterans Affairs and Principal Investigator of the VA/Department of Defense Chronic Effects of NeuroTrauma Consortium believed the technology holds promise for helping patients.
“This innovative research by Samadani and colleagues highlights a novel approach to objectively and rapidly support the diagnosis of acute concussion using a novel technique of assessing eye tracking,” he explained. “This publication may represent the first step in the development of a more exacting method of diagnosing and monitoring recovery from traumatic brain injury. Computerized assessment of eye tracking may represent the first truly useful biomarker of TBI.”
The Alternatives and Where We Stand Now
Current approaches to evaluating patients for concussion include tests that assess cognition and memory as well as motor function such as SCAT3 and Impact testing, but a baseline examination and score is required. As with any baseline evaluation, many athletes can purposely underperform in order to demonstrate improvement if they happen to suffer an injury.
While there are other current techniques in use–such as rapid number naming (King-Devick test) to evaluate early signs of a potential concussion on sidelines–eye tracking doesn’t require literacy as Samadani explains, and does not require a baseline for comparison that both King-Devick, SCAT3 and Impact testing require.
And while balance testing may reveal abnormalities early after a concussion (as part of the SCAT3 test), there is still no standardized measurement of reporting with this type of assessment.
One potential issue with Samadani’s study was the nearly 20% false positive rate in healthy subjects, which in some cases might be viewed as acceptable in relation to other types of tests, but still seen as a limitation which needs to be addressed.
And further validation studies that include more patients with concussions, compared to the numbers in the current study, will be necessary to demonstrate the feasibility of eye tracking as a method for rapid concussion evaluation.
Samadani also said that data from this study will be reviewed by the FDA in hopes to gain approval for clinical decision making and care in the hospital setting and physician offices.
One issue raised by FDA officials was whether eye tracking itself was influenced by a particular type of video. Samadani explains that the type of video is not important, but rather if your eyes move together or not. That’s the key point.
The reality is that when someone hits their head, it’s possible they will be fine and will not experience any further symptoms. If they do go on to experience nausea or dizziness, it’s important to be able to follow them with a reliable screening test that can determine whether they should be kept out of work, school or competition. And likewise, it’s also important that the test can also be a way to monitor them for return to work, school or competition. Such a modality, if validated in further trials, may help guide decisions such as these.
Samadani ultimately believes that her research has the potential to accomplish this goal.
“Our study shows that eye tracking has a higher sensitivity and specificity than any other methodology used to assess concussion,” she explained. “I think eye tracking will become an absolutely essential test for assessment of brain injury severity.”
But possibly the biggest advantage to eye tracking is that it can provide a window into an often neglected aspect of brain injury.
“I believe it provides information about function that is not readily apparent with conventional imaging and that it will ultimately change the way we currently diagnose and classify brain injury,” added Samadani.
Of note, Samadani has also spun her talents into her start-up, Oculogica, that is currently evaluating how noninvasive technology such as eye tracking can be developed into a more usable neurodiagnostic to benefit persons who suffer TBI or may have other neuropathology.