Facial recognition diagnoses rare disease with 96.6% accuracy

Researchers with the National Human Genome Research Institute (NHGRI) have used facial recognition software to diagnose rare genetic diseases in African, Asian and Latin American populations with 96.6 percent accuracy.

The study, published in the American Journal of Medical Genetics, tests the accuracy of facial recognition technology in diagnosing 22q11.2 deletion syndrome, more commonly known as DiGeorge syndrome. The genetic disease affects children and causes a variety of defects including cleft palate, heart defects, a certain facial appearance and learning difficulties. The variety of symptoms of the disease make it difficult for physicians to diagnose, especially in a diverse population.

"Human malformation syndromes appear different in different parts of the world," said Paul Kruszka, MD, MPH, a medical geneticist in NHGRI's Medical Genetics Branch. "Even experienced clinicians have difficulty diagnosing genetic syndromes in non-European populations."

Developed by Marius George Linguraru, DPhil, MA, MB, an investigator at the Sheikh Zayed Institute for Pediatric Surgical Innovation at Children's National Health System in Washington, D.C., the facial recognition technology was trained using 106 participants and 101 photographs of participants with the disease from 11 different countries. A total of 156 individuals with the disease were compared to people without DiGeorge syndrome. The technology was able to diagnose patients with 96.6 percent accuracy.

Researchers believe the technology can diagnose patients earlier.

"Healthcare providers here in the United States as well as those in other countries with fewer resources will be able to use the atlas and the facial recognition software for early diagnoses," said Maximilian Muenke, MD, chief of NHGRI's Medical Genetics Branch. "Early diagnoses means early treatment along with the potential for reducing pain and suffering experienced by these children and their families."