Unipolar voltage mapping in right ventricular cardiomyopathy: pitfalls, solutions and advantages

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. Background Endocardial unipolar and bipolar voltage mapping (UVM/BVM) of the right ventricle (RV) are used for transmural substrate delineation. Epicardial BVM is considered less accurate due to the impact of fat thickness (FT). Data on epicardial UVM are sparse. Far-field electrograms and electrogram changes due to injury current may influence automatically generated UVM, which rely on the largest peak-to-peak amplitude within the window-of-interest (WOI). Purpose To assess the influence of manually corrected WOI and the potential role of epicardial UVM in RV cardiomyopathies. Methods Consecutive patients who underwent endo-epicardial RV mapping with computed-tomography (CT) integration were included. Mapping points were superimposed on short-axis CT-slices and correlated with local FT. All points were manually re-analyzed and the WOI was adjusted to correct for false high UV. Per point-pair endo-epicardial BV and UV were correlated for different FT categories. Results A total of 4225 point-pairs of 33 patients were analyzed. In 71% of endocardial points and 65% of epicardial points the window-of-interest needed to be adjusted, due to inclusion of far-field electrograms, injury current components or RV-pacing artifacts (Figure 1). The median absolute difference between epi- and endocardial UV was -0.21mV (IQR:-1.05-0.36) for FT<1mm, compared to -0.60mV (IQR:-2.00-0.30) for FT≥2.8mm. This difference was larger for BV, with a median of -0.36mV (IQR:-2.03-0.23) and -1.29mV (IQR:-3.36-0.05), respectively (p-value epi-endo difference UV vs. BV in all fat categories: <0.001). The Pearson correlation between corrected endo-epicardial BV and UV was lower for point-pairs with greater FT; however, this correlation was stronger and less influenced by fat for UV (coefficient for UV 0.696 for FT<1mm and 0.544 for FT≥2.8mm; for BV 0.437 and 0.155, respectively; Figure 2). In a subanalysis of point-pairs with distance<5mm and FT<1mm, the Pearson coefficient was 0.789 for UV and 0.525 for BV. Conclusions At the majority of mapping sites the window-of-interest needs to be manually adjusted for correct unipolar voltage mapping. Unadjusted UVM underestimates low UV regions. UV seems to be less influenced by epicardial fat and endo-epicardial wavefront, suggesting a promising role for UVM in epicardial substrate delineation.