Abstract Body (Do not enter title and authors here): Background: Left atrial cardiomyopathy (LACM) is a recognized contributor to atrial fibrillation (AF), thromboembolic events, and heart failure (HF). In acute settings, a prompt and comprehensive evaluation is crucial but often limited by restricted time availability. Artificial intelligence (AI)-implemented echocardiographic tools may facilitate efficient acquisition and improve functional assessment.
Aims: To evaluate the utility of AI-assisted echocardiography for rapid assessment of LA function in patients with acute decompensated HF, and to investigate the association between LA strain dynamics and adverse events during hospitalization.
Methods: We prospectively enrolled 43 patients hospitalized for acute HF with a mean age of 66±13 years; 23 (50%) were male, with median ejection fraction (EF) of 46 (21) %. All undergone standard and AI-assisted echocardiography within 24 hours of admission. LA strain parameters—reservoir (LASr), conduit (LAScd), and contractile (LASct)—and their variations (ΔLASr, ΔLAScd, ΔLASct) during hospitalization were recorded. The incidence of new-onset AF and other complications was documented.
Results: In hospital stay was 6±3 days. AI-assisted echocardiography significantly reduced acquisition time (12.4±2.4 vs. 3.5±1.2 minutes; p<0.001). Two patients (4.6%) experienced systemic thromboembolism, and 9 (20.9%) developed new-onset AF. LA strain parameters improved significantly during hospitalization: after repeated assessments LASct increased from –10.8% to –13%, LASr from 23.6% to 26%, and LAScd from –12% to –13% (all p<0.001). Patients without AF showed greater improvements in ΔLASr (7.2±1.9 vs. 2.9±1.1), ΔLASct (4.6±1.3 vs. 2.3±0.9), and ΔLAScd (4.1±1.3 vs. 2.6±0.9) compared to those with AF (all p<0.001) (Figure 1). Logistic regression showed that increases in ΔLASr (OR: 0.121; 95% CI: 0.016–0.903; p=0.039), ΔLASct (OR: 0.022; 95% CI: 0.001–0.49; p=0.016), and ΔLAScd (OR: 0.163; 95% CI: 0.034–0.794; p=0.022) were independently associated with a lower risk of AF onset.
Conclusions: AI-implemented echocardiography enabled time-efficient and reproducible assessment of LA function in acute HF. Dynamic changes in LA strain parameters were associated with in-hospital onset of AF. These findings suggest a potential role for LA strain monitoring in risk stratification and management of acute HF patients.