Software diagnoses DiGeorge syndrome to assist Atlas of Human Malformations in Diverse Populations

Orange County, CA - March 27th 2017 - DiGeorge syndrome, also known as 22q11.2 deletion syndrome, is a disorder produced by the deletion of a portion of chromosome 22. The disorder affects between 1 in 3,000 to 1 in 6,000 children worldwide and can cause a variety of body defects depending on the affected body systems.  Commonly, those diagnosed suffer from heart defects, poor immune system function, cleft palate, and delayed behavioral and emotional development. The number, extent, and time of symptom development, vary making diagnosis difficult for trained physicians.

While most of those affected will grow up to live a long a normal life, some symptoms of the disease can be so severe that early detection and intervention is crucial for an increased quality of life. A difficult task, for in cases of malformation syndromes various complications manifest differently around the world, negating any uniform method of diagnosis. To assist in the early detection of DiGeorge syndrome, researchers from the National Human Genome Research Institute (NHGRI) developed facial recognition software able to diagnose the disease in Africans, Asians and Latin Americans.

Beginning the study, researchers analyzed the clinical information of 106 participants and the photographs of 101 participants all hailing from 11 countries in Africa, Asia and Latin America. The facial recognition software was then trained and used to compare a group of 156 Caucasians, Africans, Asians and Latin Americans with DiGeorge syndrome to 156 matched controls without the disease. From this information, the researchers were able to correctly diagnose with 96.6% accuracy. The team hopes to expand the software’s capabilities to eventually allow physicians to take photos of their patients with a cell phone for analysis and quick retrieval of results