Some diseases are so rare that a primary care physician may see only one or two cases in an entire career. The problem with the math in daily practice, however, comes from the sheer number of rare diseases--more than 8,000 known so far. Even low prevalence disorders add up to the likelihood that one in ten patients who walk through a primary physician's door are likely to have or develop a debilitating or life-threatening rare disease.
The difficulty in detecting one rare disease among so many possibilities leads patients on an average diagnostic journey of more than seven years and seeing an average of seven different specialists.
Although high throughput molecular testing is becoming more available, the odds of finding a diagnosis is still only 25 percent without enough patient data to sort out which genetic variants are clinically relevant.
However, a new tool that is as near as your smart phone is improving the odds and providing clues on where to look next. The Face2Gene app developed by FDNA, Inc. (Facial Dysmorphic Novel Analysis) is free to physicians and available in Iphone, Android, PC and Mac formats.
"You simply upload a photo of the patient to a HIPAA secure and verified server," UAB medical geneticist and pediatrician Anna Hurst, MD said. "The image is translated into mathematical calculations of more than 140 data points. Measurements and ratios are then compared to a vast database of facial variants linked to many different disorders. The patient's features are analyzed in seconds and the app generates a list of syndromes that match or show similarities.
"The app doesn't take the place of the physician's judgment, but it can provide added confirmation when you are considering additional testing. It helps to narrow the field when there are too many possibilities. And when you've eliminated everything likely, it provides clues on where to look next."
When Hurst was in training, her mentors were beta testers of the app, and today she is on the advisory board as a clinician using Face2Gene in real world practice.
"In some disorders with a genetic basis, there are characteristic differences in facial features, such as in Down syndrome. Other variations may be less defined, but are common in specific genetic diseases. The analysis looks at things like spacing and angle of the eyes; the position and shape of the ears; the structures of the nose; corners of the mouth and size of the chin; placement of the neck bones and overall facial proportions," Hurst said.
The app can search the London Medical Database for correlations between patient photos and such disorders as Noonan's syndrome, Smith-Magenis syndrome and Russell-Silver syndrome, as well as diseases with facial differences that may be less apparent.
"UAB is active in Neurofibromatosis research, and we often see patients who have been referred to us for diagnosis and treatment. The Face2Gene app can be helpful in confirming the diagnosis and in identifying other factors that might be involved," Hurst said.
The entire database has been crowd-sourced from real-world patient cases through broad networks of clinical, lab and research users, and additional data is being contributed on an ongoing basis. Using facial analysis, artificial intelligence, and deep learning algorithms, the app transforms data into actionable genomic intelligence to help clinicians recognize syndrome-related phenotypes and disease-causing genes.
In research, as new clinical features and genes are discovered, it can assist in developing disease-predicting biomarkers. The data can also be accessed by drug companies to develop, test and market new therapeutics for rare diseases around the world.
"As of now, the app is being used clinically primarily by specialists involved in evaluations of genetic disorders and rare diseases, but physicians who are distant from a referral facility or who have a case they suspect may have a genetic basis might want to download the app," Hurst said. "It is free and available from Google Playstore and the App Store, but you will need to register, since it can only be used by health care professionals.
"The patient's privacy is also well protected. Only the person who uploads the photo can see it. The photo is immediately translated into mathematical data that is encrypted so even the people who manage the database can see only numbers and no faces.
"However, the data Face2Gene returns can help you look at a case more clearly. Any matches are prioritized in order of correlation so you can see the most likely places to start looking to find answers for your patients."
Anna Hurst, MD
Face2Gene, FDNA, Facial Dysmorphic Novel Analysis, UAB, genetic disease, Anna Hurst MD, rare disease, London Medical Database, Noonan's syndrome, Smith-Magenis syndrome, Russell-Silver syndrome, neurofibromatosis