A research program to use AI and machine learning to predict and detect heart failure, which can be difficult to diagnose and is often unrecognized, is being launched by a joint UK-US scientific team. Specifically, it will be a collaborative work between UK health-tech company Ultromics and Mayo Clinic in the US.
The team will use AI analysis of ultrasound heart scans to identify the markers of heart failure and develop an image analysis risk prediction model, that can alert doctors to potential heart failure.
The aim is to develop a diagnostic and predictive tool that can rapidly identify heart failure, reduce misdiagnosis, and enable its earlier prevention. By supporting medical professionals, it will free-up their time to provide greater patient care and ease the pressure on care-teams. Ultimately, the ambition is to help save and improve the quality of patient’s lives.
Why does it matter?
Heart failure is a chronic, progressive condition in which the heart muscle is unable to pump enough blood to meet the body’s needs for blood and oxygen. Globally it affects at least 26 million people and is increasing in prevalence. Worldwide, it is the leading cause of hospitalization in people over the age of 65. In the US it affects over 6.5 million adults, with 550,000 new cases diagnosed each year.
The research team will use the AI engine from Ultromics’ first product EchoGo Core, to analyze 10,000 echocardiograms (echos). It will analyze 2D-echocardiograms, including the assessment of systolic and diastolic information throughout the entire cardiac cycle.
This project will be led by Gary Woodward, CTO of Ultromics and Patricia A. Pellikka, M.D., cardiologist, and clinical researcher at Mayo Clinic. It is the third collaboration between Ultromics and Mayo Clinic.
On the record
“This project is focused on a critical aspect of cardiac disease as it affects so many people every day,” said CEO and co-founder of Ultromics, Dr. Ross Upton. “Using our pioneering AI technology stack, our objective is to map and scan databases of ultrasound images and develop detailed models to diagnose and hopefully even predict heart failure. Early intervention can make a huge difference to a patient’s treatment and quality of life – so the sooner we can identify the condition, the better.”
Dr. Upton added: “The study has two key objectives: the first is to identify novel biomarkers that can help identify early signs of heart failure. And the second is to develop a machine learning model using the novel biomarkers to provide an automated risk prediction of heart failure at the point of care.”
The context
According to one report, North American AI in healthcare diagnosis market is projected to grow from $1,7 billion in 2019 to $32 billion by 2027. This represents a CAGR of 44.3% from 2020 to 2027.
This massive growth is primarily attributed to the rising adoption of AI in disease identification and diagnosis, increasing investment in AI healthcare start-ups.