AI Powered Face Scanning App For Detecting Rare Genetic Disorders
The number of Genetic Disorder detection per year is accelerating at a high pace. Physicians often land themselves in the soup, wherein its very difficult to diagnose a patient suffering from a rare genetic disorder simply by noting the symptoms. An AI app based on deep learning algorithm is here for the rescue. It will help the doctors identify a rare genetic disorder by analyzing peoples face. The app can be downloaded on the phone to take a photo of the patients face for analysis.
Face2Gene, a smartphone app is currently used by 70% of the world’s geneticists across 2,000 clinical sites in 130 countries. Developed by the Boston-based company FDNA, it uses machine-learning algorithms and brain-like neural networks to differentiate facial phenotypes in photo os patients with congenital and neurodevelopmental disorders. It is developed on a novel facial image analysis framework, DeepGestaltTM. Based on the patterns obtained after analysis it suggests possible diagnosis available with other options.
The researchers behind the app have described the technicality behind the app in a paper published in nature medicine on 7th January. Their study describes how the technology used in the app helps in analyzing complex human physiological data by
transforming phenotyping – capture, and structuring. The app was tested with a dataset of over 150,000 patients. For further analysis and to test the accuracy 17,000 patient images representing more than 200 syndromes were analyzed. As per reports in the paper – DeepGestalt achieves 91% top-10-accuracy in identifying the correct syndrome on 502 images and outperformed expert clinicians in three experiments.
Co-Author of the paper – Karen Gripp, MD, chief of the division of medical genetics, Nemours/Alfred I. DuPont Hospital for Children, chief medical officer at FDNA said that the paper they have published states how the algorithm was trained and its functionality. Out of many such systems in the market, currently, there is none with as many cases and conditions being analyzed. This paper can be used as a standard to compare other systems. The concept of AI in facial phenotyping can be applied in other imaging systems as well. She also emphasized the role of AI in precision medicine.
Gripp further explained that the app does not have an age limit to adhere. It can be readily used on young babies as well as old.
Eventually, FDNA wishes to develop this technology to assist other businesses filter, prioritize and translate genetic variants of unknown importance during DNA analysis. However, to train its own versions, FDNA wants information.
Face2Gene program is now available free of charge to health care professionals, a lot of whom utilize the machine for a sort of instant opinion for diagnosing infrequently seen hereditary ailments, says Karen Grimm. Additionally, it may provide a beginning point in situations in which a physician does not know what to make of a patient’s symptoms. “It is just like a Google search,” Grimm says.
However, it does raise a range of legal and ethical issues, say investigators. These include cultural bias in training data collections as well as the industrial fragmentation of databases, each of which may limit the range of the diagnostic instrument.
Published Paper – “Identifying Facial Phenotypes of Genetic Disorders Using Deep Learning“
