Skip to main content

Six challenges in heart disease that artificial intelligence is helping to solve

Vast amounts of data and artificial intelligence are helping to answer some of the biggest questions in cardiovascular disease. Here are six ways that BHF-funded data science research could solve pressing challenges in cardiovascular care.

Hero curve mask

29 April 2019, by Siobhan Chan

Artificial intelligence (AI) could “augment the skills of the NHS workforce” and “empower patients to manage their own health or seek appropriate health support”.

Those are the findings of the recent Topol Review on the digital future of healthcare.

But how will AI shape healthcare in the years to come?

Last year, the BHF – the single biggest independent funder of cardiovascular research in the UK – and the Alan Turing Institute awarded six research grant applications totalling over £550,000 through a joint funding scheme.

Machine learning is one of the AI techniques that data scientists involved in these schemes are using. This involves designing a set of instructions (an algorithm) that computers can use to detect patterns in vast datasets – for example, in heart scans, or across clinical databases. This allows us to spot trends that we wouldn’t otherwise have seen.

Techniques like this could help us solve challenges in cardiovascular disease (CVD) care such as prescribing medicines with potentially serious side-effects, predicting risk of heart disease, or what type of physical activity to recommend.

Read on to find out about the pioneering work the BHF is doing that will allow patients to benefit from AI research.

1. Can high-resolution photographs of blood cells help to predict heart attacks?

White blood cells and platelets can provide clues to a person’s risk of heart attack and stroke, according to researchers. But it isn’t known exactly what characteristics of these cells influence a person’s risk, and how.

High-resolution photographs of white blood cells may hold the answer. Researchers at the University of Cambridge are using algorithms to analyse blood sample images from 30,000 healthy people and measure certain properties of platelets and white blood cells, such as their shape and structure.

The team aims to identify the genes that affect these traits and see whether there is any overlap with parts of the genome known to influence CVD risk. This could help to identify drug targets for the prevention or treatment of heart attack and stroke.

2. How can we improve the accuracy of the NHS Health Check?

How can the NHS Health Check offer patients an even more personalised prediction of their CVD risk?

The national case-finding programme provides a snapshot of high-risk conditions such as hypertension and cholesterol, and gives a prediction of CVD risk overall, but does not assess risk levels for different conditions such as stroke.

Personalised risk prediction is becoming more likely thanks to the work of researchers in Cambridge, who are designing an algorithm to analyse the health records of two million people.

The algorithm will map trends in a person’s health and work out what information is needed to give a better risk score for coronary heart disease, stroke and other types of CVD. The researchers hope their work will help GPs to better treat and advise patients.

3. What type of activity is best for the heart?

Healthcare professionals often recommend that patients get more active to reduce their CVD risk. But what type of physical activity should you recommend, and how much?

Researchers are hoping to improve the lifestyle advice patients are given by looking at the types of activity that have the biggest impact on heart health. 

The team at the University of Oxford are using data from UK Biobank, a database containing health information from around 100,000 people, that includes readings from wearable devices.

From this data, they will develop machine learning methods that will work out which indicators of physical activity are best for improving cardiovascular health.

Easier access to data

The BHF is planning to fund a data science initiative that will give researchers faster and easier access to routinely collected healthcare data.

The initiative will aim to address barriers facing the analysis of healthcare data, paving the way to improve the cardiovascular health of the nation using the power of large-scale data. Proposals are currently in development with partners Health Data Research UK.

4. Who is at risk of side-effects from MI medication?

Antiplatelet medication is recommended for life following a myocardial infarction (MI) to reduce the risk of secondary cardiovascular events, but these can pose a higher risk of uncontrolled bleeding and it’s difficult to know who is most at risk.

A risk predictor tool is being developed by researchers at the University of Edinburgh and the Alan Turing Institute that uses data from over 50,000 heart attack patients in Scotland.

Once the tool has been fully tested, researchers hope it can equip doctors with the information they need to prescribe antiplatelets at the right dosage and for a safe length of time for those who are already at an increased risk of major bleeding.

5. How do certain parts of our genome affect CVD risk?

Around 130 sections of the human genome impact a person’s risk of CVD, but it’s unclear exactly how these genetic regions affect disease risk.

Previous research has investigated this by looking at how these genetic locations affect risk factors such as blood pressure or levels of specific proteins. But existing research methods struggle to distinguish whether it is a person’s genetic traits that are affecting these risk factors to cause CVD, or whether the genes, risk factors and the disease are all changing separately.

A new statistical tool being developed by a team at the University of Cambridge will help work out how the genome regions are linked to these factors.

The team hope its results will identify new drug targets, shed light on the aetiology of CVD, and pave the way for personalised preventative medicine.

6. How is calcium signalling in the heart affected by medicines?

If clinicians could more accurately predict whether an individual would have side-effects from certain medications, it would help them to prescribe more safely.

Some medications can cause arrhythmias, because they disrupt calcium signalling pathways within heart cells. Calcium signalling is controlled by proteins within cardiomyocytes, and differs between individuals.

Data scientists at King’s College London are trying to better understand the variation in calcium signalling between cells in the heart through data science.

They will then be able to generate virtual cell groups that represent the variability of cells in real hearts. They can then apply ‘virtual drugs’ to the virtual cell groups to find out whether certain cell types are more at risk of side-effects.

What do patients think about artificial intelligence?

The BHF has supported an inquiry by the All Party Parliamentary Group (APPG) on Heart and Circulatory Diseases to better understand patient perspectives on AI.

The APPG has produced a report that includes a survey of patient attitudes to AI, showing that 85% of patients support or strongly support doctors using AI technologies to assist them in diagnosing and treating CVD.

The report highlights the need to engage patients with the use of AI at an early stage to build trust with the technology.

Read the full report and its recommendations now.

We use cookies to ensure that we give you the best experience on our website. By continuing to use our site, you are agreeing to our use of cookies. Read our privacy and cookies policy for further information.