AI-based ultrasound for better diagnosis of heart disease

Friday, June 28, 2024

Philips has presented a new AI-based platform for cardiovascular ultrasound that enables faster cardiac ultrasound analysis and reduced workload. The new AI applications have received 510(k) approval from the US FDA and are integrated into the EPIQ CVx and Affiniti CVx ultrasound systems. This significantly improves cardiovascular imaging and diagnostic solutions by automating measurements and speeding up workflows. As a result, productivity increases.

This new form of ultrasound with AI can help detect heart failure, among other things. Heart failure is a common condition that affects some 64 million people worldwide. The Netherlands is estimated to have about 241,000 people with heart failure. Heart failure leads to a high mortality rate and poor quality of life and places a significant burden on health care systems around the world.

Cardiovascular ultrasound

To diagnose heart failure and other heart diseases earlier, cardiovascular ultrasound is often used. This is the least invasive way to image the structure and function of the heart. By integrating AI into this imaging technology, multiple analysis steps can be automated. This allows doctors and nurses to detect, diagnose and monitor various heart conditions more quickly, reliably and efficiently.

"By harnessing the power of AI in our echocardiography solutions, we provide cardiologists and ultrasound technicians with better diagnostic capabilities, ultimately improving patient care and outcomes in the treatment of heart disease and valve disease," said Anine van den Hurk, Head of Ultrasound at Philips Benelux. "For patients, this means that images will be interpreted more consistently, potentially reducing the need for a second scan. In addition, the procedures are shorter and more effective, so patients may recover faster."

Detecting RWMAs accurately

Roberto Lang, MD, director of the Noninvasive Cardiac Imaging Lab at the University of Chicago Medicine in the US, along with other physicians, presented new insights at the American Society of Echocardiography's annual meeting (ASE2024) held June 14-16 in the US city of Portland. This included an explanation of how AI algorithms developed in collaboration with Philips can very accurately detect regional wall motion abnormalities (RWMA).

RWMAs can be an indicator of cardiovascular events and death in patients with cardiovascular diseases, such as myocardial infarction and congenital heart disease. Automating the assessment of RWMAs using machine learning will take work out of the hands of cardiologists and ultrasound technicians. The AI functions in the cardiovascular ultrasound systems are trained using anonymized patient data sets from clinical practice. They improve the quality and reproducibility of cardiac imaging and ease the work of the users of these systems. The combination of AI features supports the interpretation of ultrasound images, allowing users to analyze images more quickly, efficiently and accurately, regardless of their ultrasound experience.

AI and heart failure

Developments in the field of AI and heart failure are rapidly following one another. Just recently, researchers at UVA Health developed a new risk assessment tool to predict adverse outcomes in patients with heart failure. The new tool improves on current heart failure risk assessment methods by using machine learning and AI to determine, on a patient-by-patient basis, how much risk a person is at to develop adverse outcomes in heart failure. The study was published in the international journal American Heart Journal, and the researchers have made the tool available to medical providers at no cost.

And earlier this year, Eko Health and the Mayo Clinic reported that they have developed an AI stethoscope that can diagnose heart failure within 15 seconds. The smart device can be used during a physical exam and can detect low ejection fraction, a key indicator of heart failure. The AI stethoscope has been trained with a dataset of more than 100,000 ECGs and echocardiograms from unique patients.

By innovation partner