Novel "late potential map" algorithm: Abnormal potentials and scar channels detection for ventricular tachycardia ablation

Background: Automated systems for substrate mapping in the context of ventricular tachycardia (VT) ablation may annotate far-field rather than near-field signals, rendering the resulting maps hard to interpret. Additionally, quantitative assessment of local conduction velocity (LCV) remains an unmet need in clinical practice. We evaluate whether a new late potential map (LPM) algorithm can provide an automatic and reliable annotation and localized bipolar voltage measurement of ventricular electrograms (EGMs) and if LCV analysis allows recognizing intrascar conduction corridors acting as VT isthmuses.

Methods: In 16 patients referred for scar-related VT ablation, 8 VT activation maps and 29 high-resolution substrate maps from different activation wavefronts were obtained. In offline analysis, the LPM algorithm was compared to manually annotated substrate maps. Locations of the VT isthmuses were compared with the corresponding substrate maps in regard to LCV.

Results: The LPM algorithm had an overall/local abnormal ventricular activity (LAVA) annotation accuracy of 94.5%/81.1%, which compares to 83.7%/23.9% for the previous wavefront algorithm. The resultant maps presented a spatial concordance of 88.1% in delineating regions displaying LAVA. LAVA median localized bipolar voltage was 0.22 mV, but voltage amplitude assessment had modest accuracy in distinguishing LAVA from other abnormal EGMs (area under the curve: 0.676; p < .001). LCV analysis in high-density substrate maps identified a median of two intrascar conduction corridors per patient (interquartile range: 2-3), including the one acting as VT isthmus in all cases.

Conclusion: The new LPM algorithm and LCV analysis may enhance substrate characterization in scar-related VT.