Determining the relevant field-of-view of unipolar electrodes from in-vivo clinical imaging
S K Allampalli, U Rohrer, I Nazarov, A Rashid, F Campos, L Azzolin, A Neic, G Plank, D Hodson, R Rajani, U Tedrow, M Bishop, J WhitakerAbstract
Background
Electroanatomical voltage mapping (EAVM) is central to the identification of arrhythmogenic substrates in ventricular tachycardia (VT) ablation. However, the relevant field-of-view (FOV) of electrodes used in EAVM is not well defined. Recent studies used volume of viable myocardium (VM) derived from Cardiac Magnetic Resonance (CMR) imaging to determine the relevant FOV of electrodes pre-clinically, but not yet with clinical in-vivo data.
Purpose
To determine the relevant FOV of 1mm electrodes from clinical data and evaluate the ability of CMR-derived VM and Cardiac Computed Tomography (CCT)-derived Extracellular Volume (ECV) in predicting unipolar electrogram voltage.
Methods
14 patients undergoing VT ablation with pre-procedural Late Gadolinium Enhanced (LGE) CMR and CCT imaging were recruited. CT Angiography (CTA) scans were segmented to create a biventricular (BiV) finite element mesh, and LGE-CMR images were segmented to create a left ventricular (LV) mesh with normalised LGE intensities. LGE values were mapped to the BiV mesh (Figure 1a) using Universal Ventricular Co-ordinates. The volume of tetrahedra with an LGE<0.5 defined the volume of VM. ECV was extracted by applying LV myocardium and blood pool masks from CTA segmentations onto late iodine enhanced (LIE) and baseline (non-contrast) CT scans to calculate the change in Hounsfield units in each region (Figure 1b). Local comparison of LGE and ECV within 1mm spheres was conducted. LV endocardial EAVM was performed with a Decanav catheter (1 mm electrodes). Peak to peak unipolar electrogram amplitudes and their measurement positions were exported from CARTO3. Volume-weighted ECV and volume of VM at different radii from each electrode position were compared to the unipolar voltages. Using scatter plots of VM and ECV vs. unipolar voltage at each FOV radius, correlation coefficients were calculated and plotted against FOV radius (Figure 1d, 1e). The radius at which the graph plateaus was defined as the relevant FOV.
Results
Regional comparison of LGE and ECV found that ECV averages were higher in regions of LGE-defined scar than in healthy myocardium. 8 patients had LV endocardial EAVM and were included in the FOV analysis. As the FOV radius increased, the correlation coefficients between VM and unipolar voltage increased. A plateau occurred at 13mm, indicating this as the relevant FOV (Figure 1d), with a maximum correlation coefficient of 0.583 (p<0.001). Volume-weighted ECV showed no distinct plateau and obtained a lower maximum correlation coefficient (Figure 1e) of -0.351 (p<0.001).
Conclusion
Volume of VM used to assess the FOV of 1mm electrodes from clinical data achieved similar accuracy to pre-clinical data and determined a relevant FOV of 13mm, marginally larger than seen in pre-clinical work. The observed correlation is lower than is often assumed, highlighting the importance of synergistic electrophysiological and imaging assessment during VT ablation.