Electrocardiographic imaging-based recognition of possible induced bundle branch blocks during transcatheter aortic valve implantations

Aims Conventional electrocardiogram (ECG)-based diagnosis of left bundle branch block (LBBB) in patients with left ventricular hypertrophy (LVH) is ambiguous. Left ventricular hypertrophy is often seen in patients with severe aortic stenosis in which a transcatheter aortic valve implantation (TAVI) frequently results in a LBBB due to the mechanical interaction of the artificial valve and the conduction system. In this feasibility study, we propose and evaluate the sensitivity of a new electrocardiographic imaging tool; the cardiac isochrone positioning system (CIPS), visualizing the cardiac activation to detect interventricular conduction patterns pre- and post-TAVI. Methods and results The CIPS translates standard 12-lead ECG into ventricular isochrones, representing the activation sequence. It requires a patient-specific model integrating heart, lungs, and other thoracic structures derived from multi-slice computed tomography. The fastest route-based algorithm was used to estimate the activation isochrones and the results were compared with standard ECG analysis. In 10 patients the CIPS was used to analyse 20 ECGs, 10 pre- and 10 post-TAVI. In 11 cases the CIPS results were in agreement with the ECG-based diagnosis. In two cases there was partial agreement and in seven cases there was disagreement. In four of these cases, the clinical history of the patients favoured interpretation as assessed by CIPS, for the remaining three, it is unknown which method correctly classified the activation. Conclusion This feasibility study applying the CIPS shows promising results to classify conduction disorders originating from the left anterior or posterior ventricular wall, or the septum. The visualization of the activation isochrones as well as ventricular model-derived features might support TAVI procedures and the therapy selection afterwards.

• View at Publisher
• View at Scopus
• View at Web of Science