Key result
Adding QRS rotation to single-lead ECG boosts LBBB positive predictive value by ~13%.
Why the study?
Single-lead ECGs are widely used in wearables for rhythm interpretation, but morphological ECG interpretation generally requires multiple leads.
Does QRS waveform decomposition and rotation improve the prediction of LBBB from single-lead ECGs?
Does QRS waveform decomposition and rotation improve the prediction of LBBB from single-lead ECGs?
Rotating the QRS waveform of a single-lead ECG to estimate different viewpoints significantly improves the morphological interpretation and detection of left bundle branch block.
May enhance morphological interpretation from single-lead wearables; leaves open prospective validation before clinical adoption.
OBJECTIVE: The single-lead electrocardiogram (ECG) is widely used in watches and other wearable devices, with proven efficacy in cardiac rhythm interpretation. However, for morphological ECG interpretation, multiple viewpoints (leads) are generally necessary. APPROACH: This contribution introduces a novel methodology to rotate the QRS waveform of a single-lead ECG for estimating QRSs of different viewpoints that may be more favorable for morphological interpretation. This rotation can be achieved by recognizing that the three-dimensional QRS loop is approximately planar, and that the shape of any one-dimensional projection of a planar loop can be determined entirely by its component on the QRS plane. In particular, this methodology models a single-lead QRS as a composition of an "intrinsic component" specific to the individual and an "angular component" represented as a simple sinusoid defined by the "phase" of the QRS along the continuum between monophasic and biphasic waveforms. Indeed, for any single-lead QRS of an arbitrary viewpoint on the QRS plane, its phase can be computed based on its isoelectric zero-crossing(s). Rotation of a QRS to a different (target) viewpoint can then be achieved by substitution of the angular component of the QRS with the angular component of the target phase. MAIN RESULTS: The performance of this methodology is assessed in this work with the large PhysioNet PTB-XL database, as a proof-of-concept application, where single-lead QRS complexes and their respective monophasic rotations are evaluated for single-lead prediction of human-annotated 12-lead ECG interpretation of left bundle branch block. Results for 224,871 single-lead predictions of LBBB via incorporating QRS rotation demonstrate a 12.6% gain in positive predictive value and a 8.9% gain in F1 score for corresponding sensitivities of at least 0.8. SIGNIFICANCE: Towards facilitating morphological interpretation, the QRS waveform of a single-lead ECG can be augmented to provide more favorable viewpoints of its QRS complex.
No takes yet. Share an insight, caveat, or question.
Chiu et al. (2026) studied Left bundle branch block (LBBB) (n=18,869). QRS waveform decomposition and rotation vs. Original single-lead QRS waveform was evaluated on Single-lead prediction of human-annotated 12-lead ECG interpretation of left bundle branch block (12.6% gain in positive predictive value and 8.9% gain in F1 score). Incorporating QRS rotation for single-lead prediction of left bundle branch block demonstrated a 12.6% gain in positive predictive value and an 8.9% gain in F1 score for sensitivities of at least 0.8.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: