The EarliPoint Evaluation has demonstrated positive phase 3 trial results, aiding in autism diagnosis and assessment in children aged 16 to 30 months. Warren Jones, PhD, chief scientific officer and cofounder of EarliTec Diagnostics, Inc., discusses the eye-tracking tool, how it is currently being used, and what could be in store for the future.
Transcript (edited for clarity):
Hi, and thank you so much for visiting Contemporary Pediatrics®. I'm editor Joshua Fitch.
Warren Jones, PhD:
I'm Warren Jones, I'm the director of research at the Marcus Autism Center, Children's Healthcare of Atlanta, and then the Norman Nien Distinguished Chair in Autism at Emory University School of Medicine. I'm also the scientific co-founder of EarliTec Diagnostics. That's a company that develops technology to aid in early diagnosis and treatment of children with autism and gives revenue to support treatment of children with autism. I'm primarily affiliated with Emory but work as a consultant to EarliTec and all my work with EarliTec is reviewed and approved by Emory University's conflict of interest management office to manage any potential for conflict. Thank you for having me here today.
Of course, Dr. Jones, thank you for joining us. Can you explain what this EarliPoint Evaluation test is and how it is used to evaluate autism spectrum disorder in children? When was it FDA cleared, and can you kind of explain its indications please?
Absolutely. The EarliPoint Evaluation for autism was cleared this past June 29, 2023, for use as really the first objective tool to aid in the diagnosis and assessment of autism, for children seen in specialty centers. So it was cleared just this past June, I can tell you, though, that the research that's being published on it, the clinical trials that are coming out, in JAMA, and in a simultaneous publication in JAMA Network Open, are really the culmination of 20 years of past research, developing our understanding of measures of how children with autism, look at and learn from the surrounding visual world. So, we really measure the way that children look at social information as presented to them, we measure their eye movements 120 times per second, to quantify what each child is paying attention to, and the cues that that child may or may not be missing. This really builds on the gold standard methods of diagnosing and assessing autism right now, where a highly trained, experienced clinician would put the child through a series of social presses and see how the child is responding. So, we spent really what was the past 20 years trying to quantify the clinical intuitions of best experienced clinicians and turn that into actionable measures that other clinicians could use. So, we're really measuring how children look at social information, what they see what they miss and those are measures that we can use as an objective biomarker to identify autism, and to identify early emerging vulnerabilities for social disability and for delays in verbal communications, language skills, and other nonverbal cognitive delays. So, what we've really tried to do is measure the way that a child is behaving to return actionable clinical results.
Thank you, Dr. Jones. Can you further explain in what settings the evaluation is given? How does that relate to the general pediatrician with this knowledge now, and with the evaluation at hand?
That's a great question. In terms of how this new tool can be used, the papers that are being published summarize results for more than 1500 children who are tested. Two papers, 3 different studies are presented, 1 initial discovery, 2 replications for that, and what we tested was the extent to which these automated measures of child behavior, objective measures of what the child is paying attention to, can actually predict diagnoses that are given by experts using best practice, current clinical gold standard assessments that take anywhere from 3 to 4 to 6 hours of assessment time per child, as well as measures of the individual strengths and vulnerabilities of each child. We've made all these tests within the papers and the clinical data that was submitted to the FDA within the context of specialty centers for autism diagnosis and treatment. So, the multi-site trial that's being published in JAMA was conducted at 6 of the best specialty centers for autism diagnosis and treatment across the country. We wanted to test in double-blinded fashion, the extent to which these eye-tracking-based measures of what the child is looking at when seeing scenes of social interaction could predict the diagnostic labels given by expert clinicians. With respect to general pediatric providers, what our hope is that we can reduce the time from parent concern from pediatrician concern to actually getting an actionable diagnosis for that child. Parents are often concerned with before the age of 2 years, pediatricians often share those parents’ concerns. But unfortunately, in terms of getting children access to experienced expert clinicians, there are lengthy waitlist all around the country. A recent paper found that it was a 2-year average wait time between the ages of first concerns and actually getting an actionable diagnosis of autism. A 2-year wait in the life of a 2-year-old child is obviously far, far too long. We know that pediatricians in the community know this already very well. The hope is that this tool can make the expert diagnostic process much more streamlined because we're providing objective measures that facilitate that process to help a child go from concerns to diagnosis and then on to the supports that may be most beneficial.
Thank you, Dr. Jones, for the explanation. I want to ask about the studies recently published in the various JAMA publications, you know, can you kind of name them? I know you broke them down a little bit, can you go through the results and what you found and any potential limitations, please if possible.
The first is "Development and replication of objective measurements of social visual engagement to aid in early diagnosis and assessment of autism." That's the paper that's published now and will be published on this Tuesday in JAMA Network open. That describes the initial development of the technology and algorithms to measure what we termed social visual engagement. That's just a way of saying what children look at and learn from in the surrounding social visual environment. So, we use eye tracking technology to measure 120 times per second how the child's eyes move, while presenting scenes of social information. In a way, this is a lot like a treadmill test in cardiology, where you present a certain context in which you can observe how the patient is responding, we present scenes of social interaction. And each one of those scenes of social interaction has many hundreds of social cues happening on a continuous basis and we compare the patient's response to normative data about the expectations of age-matched peers. So, the paper that's being published now in JAMA Network open studied 1089 children. A first cohort was used to develop the algorithms and then we perform the first replication on a first prototype device out in the community to test the performance. That set the stage for the second publication, which was a multi-site clinical trial register on clinicaltrials.gov is now being published in JAMA. That one is "Eye-tracking-based measurement of social visual engagement compared with expert clinical diagnosis of autism." In that study, it was important to take the next step, and actually have standalone devices cited in each one of these clinics across the country. They were all operated by site staff who had just under an hour's worth of training on how to operate the device. They were able to collect the data, all the data were blinded, all of the centers also provided the current best practice gold standard behavioral evaluation for autism, and at the end of the study, we were able to unblind and look at the performance of the results. Those are the data that led to the clearance of this tool by the FDA as a tool to aid in the diagnosis and assessment of autism. What's critical to us is that this is really the first objective performance-based biomarker for diagnosing and assessing autism. We're able to collect these measures in automated fashion and provide results in an automated fashion and our hope is that this can move the current age of diagnosis, which has been sort of stuck at the median age of 4 to 5 years for many years now, downwards under the age of 3, so this is the reason we call the test EarliPoint because of the emphasis on early diagnosis and early intervention. We can help children with autism succeed at any age, but if we can identify kids prior to the age of 3, or a lot of interventions that can be more supportive to the child and family at those early ages when the brain is still most plastic to the beneficial effects of early intervention. In essence to test whether these automated measures these objective measures of child behavior can predict the diagnoses given by expert clinicians. We collect both of those data in parallel in blinded fashion. So best practice current evaluations are performed and all of the eye tracking based measures on the early point evaluation are collected, and then we test the extent to which they match. So the way that we measure the performance of the diagnostic tool is the accuracy of the match the sensitivity and specificity of the match, and then we also measured in 3 critical areas for dimensional levels of how impacted a child is with autism. We measure the extent of social disability In the verbal ability of the child, their verbal communication skills and the nonverbal cognitive skills of the child. Those last 2 would be kind of like verbal and nonverbal IQ, but in the in the toddler age range. Each one of those dimensions is very important because for thinking about intervention planning. Which children need really intensive 1 on 1 support to the greatest extent possible, which children might be ready for integration in group settings with peers where they're ready to learn from some peers, depend upon some individual variability in those measures. With the gold standard behavioral assessment by expert clinicians, those measures can take 3 to 4 more hours of direct assessment by highly trained experts. What EarliPoint does is actually present scenes of social interaction, it's about 10 minutes of viewing where the child is watching those scenes, but through that, we can effectively proxy the same measures that come out of those lengthy behavioral assessments. I should say that, to be very clear, this tool is not meant to replace expert clinicians by any stretch of the imagination. We need more expert clinicians, but right now the community reality is we don't have enough, and that's one of the reasons why the wait lists are so long. So the goal for this is not a replacement but can this tool actually give clinicians the same or more information than they're getting from current assessments in a way that helps them see more children more rapidly and accelerate that pace from parent and pediatrician first concern, to actionable diagnosis and on to support to that child's development and intervention as needed. Right now, this tool is indicated with the FDA for use with kids who have concerns, use in specialty centers. We have not yet had an indication that would be general pediatrician, population-wide screening. We are working on studies that test the performance in those settings, but knowing for a pediatrician audience who are already trying to conduct some screenings at 18 and 24 months, this tool is really for kids for whom we have concerns. Our next step and we're doing it right now in a few large ongoing studies is to test if we're out in the community for at pediatricians’ practices, can we get the same level of performance in those general population settings. That's next on the horizon, this really shows us that right now, we can match expert evaluations at some of the best centers across the country.
Warren Jones, PhD, chief scientific officer and founder of EarliTec, director of research, Marcus Autism Center, Children’s Healthcare of Atlanta, associate professor, Emory University School of Medicine, discloses that he is a board member at EarliTec Diagnostics, Inc., consultancies or paid advisory boards at EarliTec Diagnostics, Inc., is employed at Emory University School of Medicine, has received grants/research funding from the National Institute of Mental Health (NIMH), has received patents for EarliTec Diagnostics, Inc., and owns stock of EarliTec Diagnostics, Inc.