Using machine learning to predict intestinal inflammation in relatives of people with Crohn’s

What this research is looking at

Close relatives (such as a parent, brother or sister) of people with Crohn’s Disease have an increased risk of developing the condition themselves.

The aim of this research is to be able to identify those close relatives who are at greatest risk of developing Crohn’s Disease and diagnose them early. Early diagnosis is important because it can allow early treatment. It has been shown that early treatment with biologic medicines has better outcomes for people with Crohn’s. This includes better control of symptoms, more time in remission (periods of good health, free from active disease), fewer complications and less need for surgery.

The researchers will gather information from close relatives of people with Crohn’s Disease. They will look at their DNA (genetic material) to see if they have any known genetic risk factors for Crohn’s. They will then link this to the person’s medical assessment and gut images from a capsule endoscopy (pill camera test). This should help the researchers see if there are any early signs of gut inflammation that might indicate Crohn’s Disease.

A method known as machine learning will then be used to learn from patterns in this information. Machine learning works by giving a computer a large amount of information. Using this information, the computer will then start to work out patterns by itself.

It is hoped that machine learning will help predict outcomes such as whether an individual will develop Crohn’s Disease or not.

What this could mean for people with Crohn’s Disease

Close relatives of people with Crohn’s Disease are more likely to develop Crohn’s Disease themselves compared with the general population. Using this method, it is hoped that doctors will be able to predict which of these people are most likely to develop Crohn’s Disease. These people can then be monitored more closely and, if necessary, be treated earlier to reduce their risk of complications.

Who is leading this research: Dr Natalie Prescott, King's College London
Our funding: £99,754
Duration: 24 months
Official title of application: Using machine learning to predict intestinal inflammation in asymptomatic relatives of patients with Crohn’s Disease