Royal Papworth Hospital AI study could detect early heart valve disease

The sounds of thousands of heartbeats are being collected in the hope that artificial intelligence will be able to diagnose heart valve disease earlier.

Valvular heart disease (VHD) affects almost two million people in the UK and this is expected to double by 2040.

A study led by Royal Papworth Hospital and the University of Cambridge is collating sounds of healthy and unhealthy hearts.

It is hoped an app could diagnose disease and be used by non-specialists.

About half of those affected by VHD are unaware of their condition, because symptoms often do not develop until the disease has become severe, the hospital said.

Since its inception, more than 1,200 patients with suspected heart valve disease or congenital heart disease have signed up to the study across five NHS hospital sites.

Heart recordings are collected by a Bluetooth stethoscope and an echocardiogram test.

Thousands of sound recordings will then be uploaded to a machine learning programme, from which a team at the University of Cambridge will be able to build a database of the noises of heart valve diseases.

This will then create an artificially intelligent stethoscope capable of analysing heart murmurs and providing an instant diagnosis, or recommending whether or not someone needs further investigation.

Debbie Bygrave, 53, from Hitchin, had an operation to repair a hole in her heart and is now taking part in the study.

She said: "Anything that I can do... if it can get someone else treated earlier than I was, and they can lead a full life afterwards, then that's good."

She began getting palpitations at 45 but was not diagnosed with a hole in her heart until she was 47 - after several hospital appointments.

"Maybe if this new stethoscope can pick it up earlier - or even when [the patient is] a child, surely that's better for their quality of life and for the NHS in general."

Andrew McDonald, a research associate at the University of Cambridge's engineering department, which is developing the app, said the software "compares the sound of a heartbeat against those of a healthy heart and produces a diagnosis".

"Using a stethoscope requires a lot of skill and experience so it's traditionally only used by GPs.

"We're hoping that our device can be a really accessible screening tool that can be used by practice nurses, pharmacists or any healthcare professional without the need for specialist training.

Fatima Hajee, a senior research nurse in cardiology, leading the study at Royal Papworth Hospital, said: "It's enormously important for patients to be diagnosed sooner rather than later.

"By using a stethoscope that will connect to a database - together - you'll get a bit of information that comes up and says this person needs further investigation and may have possible valve disease.

"When you use a traditional stethoscope you hear a murmur, whereas the AI stethoscope will connect to a database which will highlight that there is a possible valve issue."

"Patients who are diagnosed later will experience things such as breathlessness, dizzy spells and will generally struggle, walking around.

"Those patients have severe valve disease and by bringing them in sooner, we'll be able to monitor that and treat them so that they have a better quality of life later on."

The results of the study will need to be analysed before the device can be rolled out to test pharmacies to see if the disease is being detected early, and if patients' lives are being improved.

Find BBC News: East of England on Facebook, Instagram and Twitter. If you have a story suggestion email eastofenglandnews@bbc.co.uk