Key result
AI-enhanced VCG detects reduced LVEF with ~88% AUC and identifies LVH.
Why the study?
Early diagnosis of reduced LVEF and LVH remains challenging, particularly in settings with limited access to advanced imaging, prompting evaluation of an AI-enhanced vectorcardiography analysis as a screening method.
Does vectorcardiography-based analysis enhanced by artificial intelligence accurately screen for reduced left ventricular ejection fraction and left ventricular hypertrophy compared to cardiac magnetic resonance imaging?
Case-Control (n=280)
No
Does vectorcardiography-based analysis enhanced by artificial intelligence accurately screen for reduced left ventricular ejection fraction and left ventricular hypertrophy compared to cardiac magnetic resonance imaging?
Answers are grounded in this paper and cite their sources.
Vectorcardiography combined with AI-based analysis shows diagnostic potential as a non-invasive screening tool to identify patients with reduced LVEF and LVH who may require advanced cardiac imaging.
Background Early diagnosis of reduced left ventricular ejection fraction (LVEF) and left ventricular hypertrophy (LVH) remains challenging, particularly in settings with limited access to advanced imaging. Both phenotypes are associated with substantial morbidity and mortality. Our study evaluated the diagnostic performance of a vectorcardiography (VCG)-based analysis enhanced by artificial intelligence (AI) as a screening method for identifying patients with reduced LVEF and LVH. Methods In this single-centre, prospective case-control study, 280 individuals were included: patients with reduced LVEF (<40%), patients with LVH (left ventricular mass > 55 g/m2), and healthy controls. All participants underwent cardiac magnetic resonance imaging and VCG recording. The AI-based algorithm analysed a total of 583 parameters. Relevant features were identified using the Mann–Whitney–Wilcoxon test (low p-values). Bayes’ theorem was applied to update prior probabilities and calculate posterior probabilities for reduced LVEF and LVH in comparison to the control group. Results Among 280 participants, 40 presented with reduced LVEF (78% male, mean age 56 +/- 16 years, median LVEF 31.5% IQR 23.4-36.3%) and 209 with LVH (87% male, mean age 60 +/- 16 years, left ventricular mass 67g/m2 IQR 61-77 g/m2). 31 served as healthy controls (61% male, mean age 50 +/- 16 years, LVEF 62 +/- 5.6% and left ventricular mass 40g/m2 IQR 35-47g/m2). For detection of reduced LVEF, one parameter (QT/RR ratio) was identified with a p-value < 0.001%, AUC of 87.5%, sensitivity of 80.0%, specificity of 85.7%, and accuracy of 82.9%. For detection of LVH, the combination of three parameters (T-wave azimuth, T-wave magnitude, and azimuth variability) showed a sensitivity of 74.5%, a specificity of 68.6% and an accuracy of 73.6%. Conclusion Our study demonstrated the potential of using VCG combined with AI-based analysis as a promising screening tool for reduced LVEF and LVH. This approach may facilitate early identification of individuals who require further advanced cardiac diagnostic evaluations.central figureparameters
No takes yet. Share an insight, caveat, or question.
Hein et al. (2026) conducted a case-control in Reduced left ventricular ejection fraction and left ventricular hypertrophy (n=280). Vectorcardiography (VCG)-based analysis enhanced by artificial intelligence (AI) vs. Healthy controls (verified by cardiac magnetic resonance imaging) was evaluated on Diagnostic performance for detecting reduced LVEF (<40%) and LVH (LV mass > 55 g/m2) (AUC 87.5% for reduced LVEF (accuracy 82.9%); accuracy 73.6% for LVH, p=<0.001%). AI-enhanced VCG analysis detected reduced LVEF with 87.5% AUC and LVH with 73.6% accuracy.
Synapse has enriched 10 related papers. Consider them for comparative context: