The FaceMatch Initiative

Background 

As 50% of people with significant developmental delay or moderate intellectual disability (ID) have facial features which can provide a clue to genetic diagnosis, the overall aim of the FaceMatch project is to develop a sensitive and specific computer-vision phenotyping tool to aid and enhance diagnosis. It achieves this by comparing the facial features of a child without a diagnosis, to a pool of images from children with genetically confirmed forms of ID and those children who remain undiagnosed.  There are currently 600 known unique forms of ID that typically show characteristic facial features.

The FaceMatch computer-vision technology will have clinical utility to 1) triage early and appropriate genetic testing by primary care doctors and paediatricians 2) guide laboratory interpretation of DNA sequence data and 3) facilitate the discovery of novel ID genes or genetic mechanisms.  Understanding the complex and interacting pathways involved in normal cognition is the first step towards developing targeted treatments in the future.

A pilot study published in 2017 confirmed that the FaceMatch technology correctly identified a top match more than expected by chance (P<.00001) in patients with known syndromic forms of ID (1).  The results of this study support the need for further research into the clinical utility of deep phenotyping using FaceMatch combined with human phenome ontology (HPO) terms for individuals with an undiagnosed ID. Achieving an early diagnosis for conditions like Mowat-Wilson Syndrome is essential to ensure early educational intervention, guide management and triage eligibility for new therapeutic trials.

Objective of the FaceMatch Mowat-Wilson Syndrome subproject.  

The overall objective of this project is to test the clinical utility of  FaceMatch at prompting a diagnosis of Mowat-Wilson Syndrome by the age of 12 months. This project will expand the current FaceMatch database and provide proof-of-concept data for the clinical utility of the  FaceMatch technology at prompting earlier consideration of Mowat-Wilson Syndrome.

FaceMatch recruitment platform 

The FaceMatch team recently launched [April 2019] an international secure FM platform (facematch.org.au) to allow parents and doctors around the world to upload images of both diagnosed and undiagnosed people with intellectual disability. Parents are encouraged to upload a number of images across the age range, and data is securely collected and stored through encryption and a coded patient identifier.  Every new case image entered into the image database is automatically and efficiently matched against every existing image generating a ranking list of the closest matches.

FaceMatch aims to work with parents to develop a large secure facial image database which can suggest a diagnosis in people where genetic testing has not provided an answer or where genetic testing is unavailable. People with Mowat-Wilson Syndrome can often share similar facial features. Using advanced computer vision technology to match the faces of children from around the world, the project aims to provide a diagnosis for more families and to assist the discovery of new developmental genes. The main benefit of this project is the potential to help families find a diagnosis.

What does participation in the project involve? 

Parents are invited to visit the FaceMatch site at FaceMatch.org.au to complete registration and online consent process. Consenting parents will provide information for their son or daughter’s profile page in the secure FaceMatch platform. This will include uploading photographs of them at different ages and providing relevant medical information.

Parents of children or adults who have a confirmed genetic diagnosis of Mowat-Wilson Syndrome will be asked to upload a scanned or photographed copy of the letter or genetic result confirming the diagnosis.

Parents of children or adults with a clinical diagnosis of Mowat-Wilson Syndrome who have tested negative on genetic analysis will be asked to nominate an associated doctor to work with the Facematch team.

Privacy and confidentiality  

Any information collected for this research project will be treated as confidential. Information and photographs are stored securely away from the publicly viewed FaceMatch website. All uploaded photographs and data will be encrypted and stored behind a firewall in a password protected computer at the Hunter Medical Research Institute. We will remove names and give participants a special code. Only the research team can match the names to their code. De-identified data generated by the computer vision software from the photograph is used for matching purposes.  Images can only be viewed within the secure FaceMatch site by the FaceMatch team. When the FaceMatch team writes or talks about the results of this project, information will be provided in such a way that people cannot be identified. Images will not be published without additional consent.

For more information see the FaceMatch Website