Children with rare conditions could be diagnosed quicker thanks to 3D facial analysis software.
Richard Palmer at Curtin University in Western Australia and his colleagues have developed a tool that can spot subtle, but important, differences in facial geometry.
Around one in three rare and genetic diseases show up in facial features. For example, Foetal Alcohol Spectrum Disorders (FASD), a group of conditions caused by alcohol exposure in the uterus, often leads to a thin upper lip, a smoother area between the top lip and bottom of the nose and a shorter distance between the corners of the eyes.
Advertisement
The team鈥檚 tool, called , scans a person鈥檚 face.聽It then creates a 3D image of the face and measures the distances between the person鈥檚 facial features. It compares this with the average measurement for their age, sex and ethnicity.
The way you would expect a face to change as someone ages depends on their ethnicity and sex. If the measurements taken by Cliniface are too far from the average for someone in that demographic, it will flag the deviations and label the symptom. A human clinician would still make the diagnosis.
At the moment, the tool uses databases with predominantly European Caucasian measurements, but Palmer says users can add other databases to the system.
鈥淢any of the syndromes we deal with are incredibly rare, so as an individual clinician you may not have seen the condition the patient in front of you has,鈥� says Natasha Brown at the Victorian Clinical Genetic Service in Australia, who wasn鈥檛 involved in the work.
Humans still required
Currently, diagnosing some conditions requires doctors to measure the face with callipers and look at a set of reference images, which is a difficult process. In contrast, any picture taken with a suitable camera can be run through the software, says Palmer.
At present, the tool takes 18 distinct measurements to identify a total of 25 different facial characteristics. The team doesn鈥檛 yet have information on every one of the facial differences for the 7000 or so rare diseases in existence, but it is hoping that users of the tool will augment the databases with their own findings.
The tool is still a long way from replacing human expertise, says Brown. 鈥淣o one tool is perfect, and so what we need to do is synthesise the information with our clinical experience and with the experience of our colleagues,鈥� she says.
While distance between facial features is a well-established tool to detect facial differences, the team is looking to incorporate the curves of the face and genomic data to shed more light on underlying conditions.
One of the most interesting applications will be to use these 3D visualisations to get more accurate data on how faces change over time with different conditions, says Brown.
