Key result
Dual-domain SCG deep learning achieves ~65% accuracy in classifying primary valve lesions.
Why the study?
Accessible screening tools for heart valve disease remain limited. Although tri-axial seismocardiogram signals capture mechanical signatures of stenosis and regurgitation, their non-stationary nature and subtle inter-lesion morphological differences challenge robust classification using a single representation.
Does an ECG-free dual-domain deep learning framework fusing temporal and spectrotemporal seismocardiogram representations improve primary valve lesion classification accuracy compared to unimodal encoders and baseline models in patients with valvular heart disease?
Observational (n=94)
Does an ECG-free dual-domain deep learning framework fusing temporal and spectrotemporal seismocardiogram representations improve primary valve lesion classification accuracy compared to unimodal encoders and baseline models in patients with valvular heart disease?
Combining temporal and spectrotemporal seismocardiogram representations via deep learning enables superior ECG-free classification of valvular heart disease compared to unimodal models under patient-independent evaluation.
Supports SCG fusion research in valvular disease; leaves open clinical adoption pending prospective validation.
Heart valve disease (HVD) is a growing cause of cardiovascular morbidity, yet accessible screening tools remain limited. Tri-axial seismocardiogram (SCG) signals capture mechanical signatures of both stenosis and regurgitation, but their non-stationary nature and subtle inter-lesion morphological differences challenge robust classification from a single representation. This study proposes an electrocardiogram (ECG)-free dual-domain deep learning framework that combines a 1D convolutional-recurrent encoder of raw SCG waveforms with a 2D EfficientNet encoder of continuous wavelet transform scalograms, integrating temporal and spectrotemporal representations through feature-level fusion. The framework was evaluated on primary valve lesion classification in 94 patients, with an additional exploratory analysis of multi-valvular, aortic-stenosis-dominant disease. Under patient-level stratified group cross-validation, the fused model achieved the highest mean accuracy (65.3%; 95% CI, 59.8-70.7%), exceeding both unimodal encoders and the majority-class (35.2%) and covariate-only (39.1%) baselines. The same mean performance ordering was preserved when the image backbone was replaced, although the confidence intervals overlap at this cohort size. Under the segment-level protocol used in prior studies, the framework achieved 99.63% accuracy, with faster training convergence and more compact, separable feature clusters than either unimodal encoder. These findings support the value of combining temporal and spectrotemporal SCG representations and demonstrate its potential for ECG-free HVD classification under patient-independent evaluation.
No takes yet. Share an insight, caveat, or question.
Kim et al. (2026) conducted an observational in Valvular heart disease (n=94). ECG-free dual-domain deep learning framework combining 1D convolutional-recurrent and 2D EfficientNet encoders of SCG signals vs. Unimodal encoders, majority-class baseline, and covariate-only baseline was evaluated on Mean accuracy of primary valve lesion classification under patient-level stratified group cross-validation (95% CI 59.8-70.7). Dual-domain deep learning fusion of SCG signals achieved 65.3% mean accuracy (95% CI, 59.8-70.7%) for primary valve lesion classification, exceeding majority-class (35.2%) and covariate (39.1%) baselines.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: