Precision of AI-Based Cardiac Ultrasound for LVEF in the Elderly

Heart failure (HF) is a major chronic illness, particularly common in older adults. With advances in healthcare and an aging population, HF is increasingly affecting people over 65 years old. In fact, 80% of HF patients are over 65. HF is associated with high mortality rates and is the leading cause of hospitalization after age 80, and even after age 65 in some countries like France.

In older adults, HF symptoms are often atypical due to multiple other health conditions, increased frailty, and associated geriatric syndromes, making diagnosis difficult. In this context, echocardiography (an ultrasound of the heart) is essential for accurately diagnosing HF.

Evaluating the left ventricular ejection fraction (LVEF) through echocardiography is a fundamental step in diagnosing HF and deciding on treatment strategies. This evaluation helps refine the HF diagnosis, propose appropriate treatments, and monitor changes in heart function over time.

One major challenge in geriatric units and nursing homes (EHPADs) is the difficulty in accessing echocardiography due to growing demand and a limited number of specialized doctors.

Recent studies have shown that automated LVEF measurements assisted by artificial intelligence (AI) using portable ultrasound devices are reliable and produce results comparable to traditional methods. This AI-assisted automatic LVEF calculation (AutoEF-AI) could be a major advantage in geriatric departments, providing a credible alternative to conventional echocardiography for evaluating LVEF in HF patients.

The main goal of this study was to evaluate the relevance and accuracy of AutoEF-AI echocardiography in elderly patients. The secondary goal was to assess whether geriatricians without specialized training in echocardiography could capture images of sufficient quality to be analyzed by automatic LVEF algorithms with acceptable accuracy.

In simple terms, this study aims to determine if using portable ultrasound devices, assisted by artificial intelligence, can provide diagnostics as precise as traditional methods. This could make evaluating heart function more accessible and effective for older adults with heart failure, even when specialists are not available.

Inclusion criteria

• At least 75 years and a clinical presentation of acute heart failure consistent with the criteria of the European Society of Cardiology guidelines

Exclusion criteria

• unstable patient