This App Helps Diagnose Rare Genetic Disorders from a Picture

FDNA's Face2Gene technology analyzes patient biometric data using artificial intelligence, identifying correlations with disease-causing genetic variations.

(Credit: FDNA)


Medical geneticist Omar Abdul-Rahman had a hunch. He thought that the three-year-old boy with deep-set eyes, a rounded nose, and uplifted earlobes might have Mowat-Wilson syndrome, but he'd never seen a patient with the rare disorder before.

"If it weren't for the app I'm not sure I would have had the confidence to say 'yes you should spend $1000 on this test."

Rahman had already ordered genetic tests for three different conditions without any luck, and he didn't want to cost the family any more money—or hope—if he wasn't sure of the diagnosis. So he took a picture of the boy and uploaded the photo to Face2Gene, a diagnostic aid for rare genetic disorders. Sure enough, Mowat-Wilson came up as a potential match. The family agreed to one final genetic test, which was positive for the syndrome.

"If it weren't for the app I'm not sure I would have had the confidence to say 'yes you should spend $1000 on this test,'" says Rahman, who is now the director of Genetic Medicine at the University of Nebraska Medical Center, but saw the boy when he was in the Department of Pediatrics at the University of Mississippi Medical Center in 2012.

"Families who are dealing with undiagnosed diseases never know what's going to come around the corner, what other organ system might be a problem next week," Rahman says. With a diagnosis, "You don't have to wait for the other shoe to drop because now you know the extent of the condition."

A diagnosis is the first and most important step for patients to attain medical care. Disease prognosis, treatment plans, and emotional coping all stem from this critical phase. But diagnosis can also be the trickiest part of the process, particularly for rare disorders. According to one European survey, 40 percent of rare diseases are initially misdiagnosed.

Healthcare professionals and medical technology companies hope that facial recognition software will help prevent families from facing difficult disruptions due to misdiagnoses.

"Patients with rare diseases or genetic disorders go through a long period of diagnostic odyssey, and just putting a name to a syndrome or finding a diagnosis can be very helpful and relieve a lot of tension for the family," says Dekel Gelbman, CEO of FDNA.

Consequently, a misdiagnosis can be devastating for families. Money and time may have been wasted on fruitless treatments, while opportunities for potentially helpful therapies or clinical trials were missed. Parents led down the wrong path must change their expectations of their child's long-term prognosis and care. In addition, they may be misinformed regarding future decisions about family planning.

Healthcare professionals and medical technology companies hope that facial recognition software will help prevent families from facing these difficult disruptions by improving the accuracy and ease of diagnosing genetic disorders. Traditionally, doctors diagnose these types of conditions by identifying unique patterns of facial features, a practice called dysmorphology. Trained physicians can read a child's face like a map and detect any abnormal ridges or plateaus—wide-set eyes, broad forehead, flat nose, rotated ears—that, combined with other symptoms such as intellectual disability or abnormal height and weight, signify a specific genetic disorder.

These morphological changes can be subtle, though, and often only specialized medical geneticists are able to detect and interpret these facial clues. What's more, some genetic disorders are so rare that even a specialist may not have encountered it before, much less a general practitioner. Diagnosing rare conditions has improved thanks to genomic testing that can confirm (or refute) a doctor's suspicion. Yet with thousands of variants in each person's genome, identifying the culprit mutation or deletion can be extremely difficult if you don't know what you're looking for.

Facial recognition technology is trying to take some of the guesswork out of this process. Software such as the Face2Gene app use machine learning to compare a picture of a patient against images of thousands of disorders and come back with suggestions of possible diagnoses.

"This is a classic field for artificial intelligence because no human being can really have enough knowledge and enough experience to be able to do this for thousands of different disorders."

"When we met a geneticist for the first time we were pretty blown away with the fact that they actually use their own human pattern recognition" to diagnose patients, says Gelbman. "This is a classic field for AI [artificial intelligence], for machine learning because no human being can really have enough knowledge and enough experience to be able to do this for thousands of different disorders."

When a physician uploads a photo to the app, they are given a list of different diagnostic suggestions, each with a heat map to indicate how similar the facial features are to a classic representation of the syndrome. The physician can hone the suggestions by adding in other symptoms or family history. Gelbman emphasized that the app is a "search and reference tool" and should not "be used to diagnose or treat medical conditions." It is not approved by the FDA as a diagnostic.

"As a tool, we've all been waiting for this, something that can help everyone," says Julian Martinez-Agosto, an associate professor in human genetics and pediatrics at UCLA. He sees the greatest benefit of facial recognition technology in its ability to empower non-specialists to make a diagnosis. Many areas, including rural communities or resource-poor countries, do not have access to either medical geneticists trained in these types of diagnostics or genomic screens. Apps like Face2Gene can help guide a general practitioner or flag diseases they might not be familiar with.

One concern is that most textbook images of genetic disorders come from the West, so the "classic" face of a condition is often a child of European descent.

Maximilian Muenke, a senior investigator at the National Human Genome Research Institute (NHGRI), agrees that in many countries, facial recognition programs could be the only way for a doctor to make a diagnosis.

"There are only geneticists in countries like the U.S., Canada, Europe, Japan. In most countries, geneticists don't exist at all," Muenke says. "In Nigeria, the most populous country in all of Africa with 160 million people, there's not a single clinical geneticist. So in a country like that, facial recognition programs will be sought after and will be extremely useful to help make a diagnosis to the non-geneticists."

One concern about providing this type of technology to a global population is that most textbook images of genetic disorders come from the West, so the "classic" face of a condition is often a child of European descent. However, the defining facial features of some of these disorders manifest differently across ethnicities, leaving clinicians from other geographic regions at a disadvantage.

"Every syndrome is either more easy or more difficult to detect in people from different geographic backgrounds," explains Muenke. For example, "in some countries of Southeast Asia, the eyes are slanted upward, and that happens to be one of the findings that occurs mostly with children with Down Syndrome. So then it might be more difficult for some individuals to recognize Down Syndrome in children from Southeast Asia."

There is a risk that providing this type of diagnostic information online will lead to parents trying to classify their own children.

To combat this issue, Muenke helped develop the Atlas of Human Malformation Syndromes, a database that incorporates descriptions and pictures of patients from every continent. By providing examples of rare genetic disorders in children from outside of the United States and Europe, Muenke hopes to provide clinicians with a better understanding of what to look for in each condition, regardless of where they practice.

There is a risk that providing this type of diagnostic information online will lead to parents trying to classify their own children. Face2Gene is free to download in the app store, although users must be authenticated by the company as a healthcare professional before they can access the database. The NHGRI Atlas can be accessed by anyone through their website. However, Martinez and Muenke say parents already use Google and WebMD to look up their child's symptoms; facial recognition programs and databases are just an extension of that trend. In fact, Martinez says, "Empowering families is another way to facilitate access to care. Some families live in rural areas and have no access to geneticists. If they can use software to get a diagnosis and then contact someone at a large hospital, it can help facilitate the process."

Martinez also says the app could go further by providing greater transparency about how the program makes its assessments. Giving clinicians feedback about why a diagnosis fits certain facial features would offer a valuable teaching opportunity in addition to a diagnostic aid.

Both Martinez and Muenke think the technology is an innovation that could vastly benefit patients. "In the beginning, I was quite skeptical and I could not believe that a machine could replace a human," says Muenke. "However, I am a convert that it actually can help tremendously in making a diagnosis. I think there is a place for facial recognition programs, and I am a firm believer that this will spread over the next five years."

Dana Smith
Dana Smith is a freelance science writer specializing in brains and bodies. She has written for the Atlantic, the Guardian, NPR, Scientific American, Discover, and Fast Company, among others. In a previous life, she earned a PhD in Experimental Psychology from the University of Cambridge. You can find more of her writing at danagsmith.com.
Get our top stories twice a month
Follow us on

David Kurtz making DNA sequencing libraries in his lab.

Photo credit: Florian Scherer

When David M. Kurtz was doing his clinical fellowship at Stanford University Medical Center in 2009, specializing in lymphoma treatments, he found himself grappling with a question no one could answer. A typical regimen for these blood cancers prescribed six cycles of chemotherapy, but no one knew why. "The number seemed to be drawn out of a hat," Kurtz says. Some patients felt much better after just two doses, but had to endure the toxic effects of the entire course. For some elderly patients, the side effects of chemo are so harsh, they alone can kill. Others appeared to be cancer-free on the CT scans after the requisite six but then succumbed to it months later.

"Anecdotally, one patient decided to stop therapy after one dose because he felt it was so toxic that he opted for hospice instead," says Kurtz, now an oncologist at the center. "Five years down the road, he was alive and well. For him, just one dose was enough." Others would return for their one-year check up and find that their tumors grew back. Kurtz felt that while CT scans and MRIs were powerful tools, they weren't perfect ones. They couldn't tell him if there were any cancer cells left, stealthily waiting to germinate again. The scans only showed the tumor once it was back.

Blood cancers claim about 68,000 people a year, with a new diagnosis made about every three minutes, according to the Leukemia Research Foundation. For patients with B-cell lymphoma, which Kurtz focuses on, the survival chances are better than for some others. About 60 percent are cured, but the remaining 40 percent will relapse—possibly because they will have a negative CT scan, but still harbor malignant cells. "You can't see this on imaging," says Michael Green, who also treats blood cancers at University of Texas MD Anderson Medical Center.

Keep Reading Keep Reading
Lina Zeldovich
Lina Zeldovich has written about science, medicine and technology for Scientific American, Reader’s Digest, Mosaic Science and other publications. She’s an alumna of Columbia University School of Journalism and the author of the upcoming book, The Other Dark Matter: The Science and Business of Turning Waste into Wealth, from Chicago University Press. You can find her on http://linazeldovich.com/ and @linazeldovich.


Reporter Michaela Haas takes Aptera's Sol car out for a test drive in San Diego, Calif.

Courtesy Haas

The white two-seater car that rolls down the street in the Sorrento Valley of San Diego looks like a futuristic batmobile, with its long aerodynamic tail and curved underbelly. Called 'Sol' (Spanish for "sun"), it runs solely on solar and could be the future of green cars. Its maker, the California startup Aptera, has announced the production of Sol, the world's first mass-produced solar vehicle, by the end of this year. Aptera co-founder Chris Anthony points to the sky as he says, "On this sunny California day, there is ample fuel. You never need to charge the car."

If you live in a sunny state like California or Florida, you might never need to plug in the streamlined Sol because the solar panels recharge while driving and parked. Its 60-mile range is more than the average commuter needs. For cloudy weather, battery packs can be recharged electronically for a range of up to 1,000 miles. The ultra-aerodynamic shape made of lightweight materials such as carbon, Kevlar, and hemp makes the Sol four times more energy-efficient than a Tesla, according to Aptera. "The material is seven times stronger than steel and even survives hail or an angry ex-girlfriend," Anthony promises.

Keep Reading Keep Reading
Michaela Haas
Michaela Haas, PhD, is an award-winning reporter and author, most recently of Bouncing Forward: The Art and Science of Cultivating Resilience (Atria). Her work has been published in the New York Times, Mother Jones, the Huffington Post, and numerous other media. Find her at www.MichaelaHaas.com and Twitter @MichaelaHaas!