Using AI to create an app that can diagnose genetic disorders

- patent application from UI and Landspítali University Hospital in progress 
 
A discovery by scientists at the University of Iceland and Landspítali University Hospital has revealed the possibility of diagnosing genetic disorders using a smartphone. The technology that originally led to the creation of smartphones was developed through basic research in universities. This new solution uses AI that can diagnose at least one rare genetic disorder by analysing fingerprints. Automated fingerprint analysis is used to generate results that are believed to be very reliable. Scientists have developed a specialised app, and the idea has now been approved for patent protection. The team also recently published a paper in the journal Genetics in Medicine Open, explaining what they have discovered.

The scientists began by working on diagnosing a disease called Kabuki syndrome, which is known to cause atypical fingerprint patterns. People with Kabuki syndrome may have mild to moderate intellectual disabilities, growth delays, heart defects, visual and hearing impairments, as well as communication difficulties. 

“We developed an app for smartphones that collects fingerprints, as well as other important information such as age, sex, ethnicity and genetic mutations, if known, and compared the data from individuals with Kabuki syndrome to the fingerprint patterns of the control group,” explains Lotta María Ellingsen. Lotta is professor of medical image analysis at the UI Faculty of Electrical and Computer Engineering and led the team responsible for this discovery along with Hans Tómas Björnsson, professor at UI and senior consultant in clinical genetics at Landspítali University Hospital. 

Lotta María Ellingsen.

Lotta explains that the team’s research confirmed using two independent methods that fingerprint analysis can be used to diagnose Kabuki syndrome, which is the first step towards potentially using smartphones and AI to help diagnose other genetic disorders.
“We tested two methods of distinguishing between the two groups: logistic regression and convolutional neural networks. These methods use different data to produce a diagnosis, based solely on fingerprint images, and both methods successfully distinguished between individuals with Kabuki syndrome and the control group with a reasonable degree of accuracy.”

Innovation and a successful patent application

The Strategy for UI places particular emphasis on innovation in research. “This project has generated the world’s first diagnostic tool based on automated fingerprint analysis. This is technology based on an old field of study that has not managed to make much headway except for the purposes of personal identification. So this is a new application for a method that the public are very familiar with and use in their daily lives, which could revolutionise the diagnosis of genetic disorders and, in the future, enable diagnosis of genetic disorders using mobile apps,” says Lotta.

“Our initial findings show that we can use advances in automated analysis using AI to get clinical value from fingerprints. The end goal would be to develop a simple mobile app that can be paired with a fingerprint scanner and used to diagnose certain genetic disorders, such as Kabuki syndrome, improving access to genetic analysis in places where there is a lack of specialist knowledge and/or genetic testing, for example due to poverty, in rural areas, or because of other factors.”
 

One of the metrics that leading universities often use to measure the success of research and innovation is the number of patents resulting from discoveries made. This means it is particularly gratifying to know that the patent application submitted by Lotta, Hans Tómas and the medical student Viktor Ingi Ágústsson, who is first author of the paper, has been approved.

“The patent will protect our invention on behalf of the University of Iceland and Landspítali University Hospital, enabling us to develop a tool to support diagnosis by healthcare professionals who lack specialised training in genetics or are working in regions with limited access to knowledge of genetics. The patent process is still ongoing but we are optimistic that the patent will be granted and we will be able to create a diagnostic tool to benefit as many people as possible. It would then be possible to found a company based on the idea, or sell the patent to another party.” 

Came up with the idea round the kitchen table

Lotta says that they first came up with the idea chatting round the kitchen table at home. The research bridges the specialisms of Lotta and her husband Hans Tómas Björnsson, medical doctor and professor at the UI Faculty of Medicine.  Lotta is an expert in automated image analysis, while Hans Tómas is a genetics specialist and one of the world’s leading scientists in Kabuki syndrome research. 

Hans Tómas Björnsson.

The target market for the solution is doctors and other healthcare professionals involved with diagnosing congenital genetic disorders, both GPs and specialists, particularly in regions with limited resources.  “This would ensure that a much broader group of patients would receive the right diagnosis and appropriate treatment, which in turn would lead to improved health and wellbeing and increased social equality,” says Lotta. 

Better diagnostic technology would also make it easier to diagnose conditions early, which would be extremely helpful and could be key to the prognosis for people with disorders such as Kabuki syndrome. “Recent research by Hans Tómas and his team indicates that specialised early intervention could reduce the level of intellectual disability caused by the syndrome. More efficient diagnostic methods could also reduce the costs of an otherwise lengthy process and save clinical staff valuable time, giving them more time to treat more patients.”

The University of Iceland is committed to advancing the UN Sustainable Development Goals, which place emphasis on the importance of improving global health and equality. This research supports Goal 3, good health and wellbeing, Goal 9, industry, innovation and infrastructure, and Goal 10, reduced inequalities.