Localization of the ventricular pacing site from BSPM and standard 12-lead ECG: a comparison study

Figures

• 

  Figure 1. The overview of the used inverse ECG method (i-Stim) to localize right and left ventricular stimulation sites as well as the intrinsic ventricular activation. (a) X-ray image of a patient with implanted CRT device, and ECG electrodes attached. (b) The CT from which the patient specific model of heart and torso are reconstructed. (c) The frontal subset of the 99 body-surface-potential map (BSPM) electrodes. Red circles indicate the used 12-lead ECG positions, where V5/V6 were placed on the back. (d) Example of a BSPM with left ventricular (LV) stimulation. (e) The used gold standard: the LV and RV CRT lead tip can be localized from the CT, also shown in the heart model as a red dot. (f) the output of i-Stim for the localization of the LV stim site (red dot), with the gold standard in yellow.

• 

  Figure 2. Activation sites for the initial activation. In the left panel the sites identified as potential His-Purkinje excitation sites. The right panels show the right and left ventricular activation sites as identified by the QRS axis. The right ventricular activation site can be either located on the right septal wall (not visible) or on the right free wall.

• 

  Figure 3. Example of the localized RV and LV lead of patient 4. A geodesic error of 0 mm means that the identified lead location was on the triangle to which the CT derived lead location was projected. For instance, see initial estimate RV stim location. The dark red point is the vertex from which the simulated activation started; the yellow dot is the gold standard lead location derived from CT.

• 

  Figure 4. comparison between 12-lead ECG derived activation and BSPM of patient 1 and 7. Patient 1 shows early activation on the left septum with late activation on the left lateral wall, whereas patient 7 shows a typical left bundle block pattern with only initial activation on the right free wall. The optimization of the activation sequence changed limited when the 12-lead ECG was used, slightly more changes were found in the estimations from BSPMs.

• 

  Figure 5. Localization errors and correlations between different localizations. (Panel a) Geodesic localization error for the initial (green) and optimized (blue) activation sequence (panel b). A geodesic error of 0 mm means that the identified lead location was on the triangle to which the CT derived lead location was projected. (Panel c) shows the correlation for the normal sinus rhythm activation maps derived from the 12-lead ECG and the BSPM for the initial sequence (dark yellow) and after the final optimization (light yellow). In (Panel d) the correlation of the activation sequence before and after optimization in case the 12-lead ECG was used (light gray) and when the BSPM was used (dark gray). The found correlations were in most cases well above 90% , i.e. the initial and optimized map were similar.