The British Heart Foundation (BHF) Data Science Centre, led by Health Data Research UK, is hosting an open challenge which invites competitors to explore the potential of Artificial Intelligence to improve the use of electrocardiogram (ECG) for cardiovascular disease patient care. The challenge has been co-designed with members of public and patients affected by cardiovascular disease.
The BHF Data Science Centre is collaborating with experts from the University of Edinburgh to use a synthetic imaging dataset made up of approximately 20,000 simulated electrocardiogram (ECG) images. Competitors will be invited to develop algorithms which can make clinical diagnoses, based on the information held within the ECG images.
Professor Michelle Williams, Associate Director of the BHF Data Science Centre, said: “Open Challenges provide a mechanism for researchers to collaborate on some of the most pressing challenges in health data research. The analysis of imaging data is one of the most challenging topics we face at the BHF Data Science Centre and we are excited to launch this opportunity for the research community to make some important discoveries.
“This dataset mimics pictures of ECGs in realistic environments, along with cardiologist verified clinical diagnoses. Turing tests have been performed to show that these ECG images are indistinguishable from real world ECGs, and the dataset will be made publicly available with a hold-out set reserved for testing.”
Many existing algorithms are limited to analysing digitised signal data, rendering them ineffective for the paper-based ECGs which are still prevalent in many clinical settings. The BHF Data Science Centre hopes that the algorithms developed using this synthetic data set could eventually be applied to real world datasets, ultimately improving the diagnosis, treatment and prevention of cardiovascular disease.
“We’ve decided to host this Open Challenge to encourage out-of-the-box thinking,” said Professor Steffen Petersen, Director of the BHF Data Science Centre. “Solutions can come from people involved in any discipline. For example, researchers who are usually involved in environmental studies may have solutions that could help with a healthcare problem, and the open challenge is an opportunity for them to explore this. Individuals are encouraged to form multi-disciplinary teams, and collaborate with those out with their immediate area of expertise.”
A working group which includes three members of the public has been established to develop plans for the open challenge, and will continue to support throughout. The input from our public contributors has been immensely helpful – supporting us to develop a challenge that keeps patient’s needs and hopes at the heart of the process.
The open challenge will run until 16 December 2024. The winning group will have the opportunity to present their results at the British Cardiovascular Society Annual Conference 2025 and to work with the centre’s team to publish the results of the open challenge.
If you are interested in forming a team to participate in the challenge, please contact bhfdsc@hdruk.ac.uk or visit Open Challenge - British Heart Foundation - Data Science Centre (bhfdatasciencecentre.org) for more information.