Optimization of Lead Placement in the Right Ventricle During Cardiac Resynchronization Therapy. A Simulation Study

Abstract

Patients suffering from heart failure and left bundle branch block show electrical ventricular dyssynchrony causing an abnormal blood pumping. Cardiac resynchronization therapy (CRT) is recommended for these patients. Patients with positive therapy response normally present QRS shortening and an increased left ventricle (LV) ejection fraction. However, around one third do not respond favorably. Therefore, optimal location of pacing leads, timing delays between leads and/or choosing related biomarkers is crucial to achieve the best possible degree of ventricular synchrony during CRT application. In this study, computational modeling is used to predict the optimal location and delay of pacing leads to improve CRT response. We use a 3D electrophysiological computational model of the heart and torso to get insight into the changes in the activation patterns obtained when the heart is paced from different regions and for different atrioventricular and interventricular delays. The model represents a heart with left bundle branch block and heart failure, and allows a detailed and accurate analysis of the electrical changes observed simultaneously in the myocardium and in the QRS complex computed in the precordial leads. Computational simulations were performed using a modified version of the O'Hara et al. action potential model, the most recent mathematical model developed for human ventricular electrophysiology. The optimal location for the pacing leads was determined by QRS maximal reduction. Additionally, the influence of Purkinje system on CRT response was assessed and correlation analysis between several parameters of the QRS was made. Simulation results showed that the right ventricle (RV) upper septum near the outflow tract is an alternative location to the RV apical lead. Furthermore, LV endocardial pacing provided better results as compared to epicardial stimulation. Finally, the time to reach the 90% of the QRS area was a good predictor of the instant at which 90% of the ventricular tissue was activated. Thus, the time to reach the 90% of the QRS area is suggested as an additional index to assess CRT effectiveness to improve biventricular synchrony.

Keywords: LBBB; QRS duration; cardiac resynchronization therapy; computational modeling; heart failure; optimization.

Figure 1

Anatomical model. (A) Biventricular hexahedral mesh of a segmented human heart. (B) Model color-coded to show the assignment of the elements to the different cellular model in order to model the transmural heterogeneity: endocardial cells (blue), midmyocardial cells (green) and epicardial cells (red). (C) Arrows indicating the principal myofiber orientation of epicardial (red) and midmyocardial (green) cells. (D) Purkinje System (PS), including three main LV branches (posterior, septal, anterior) and RV main brunches (septal and anterior). Purkinje-Junctions are represented as magenta spheres. His Bundle, and the location of the LBBB are labeled in the model. (E) PS (black) coupled to the biventricular model. (F) Torso model with the biventricular mesh embedded (red) and precordial leads location (white).

Figure 2

Heart subdivisions and stimulation points for CRT protocol. (A) RV septal endocardial stimulation points tested (green). (B) Left ventricular (LV) free wall region divided into three regions: posterior (yellow), anterior (brown), and lateral (green). (C) Subdivisions of the three LV free wall regions into nine segments. Epicardial stimulation points tested in the middle of each segment (red dots). (D) Endocardial stimulation points tested in the LV free wall (blue dots).

Figure 3

Model validation. (A) Cross section of biventricular model showing color coded local activation maps of a healthy (left) and pathological heartbeat (right). (B) Precordial leads signals recorded on torso surface.

Figure 4

Precordial leads signals on CRT. QRS complexes in the precordial leads under HF + LBBB conditions, before (red trace) and after (green trace) the application of the best CRT configurations (shorter QRSd). Three different locations for the RV pacing lead were tested: RV apex with epicardial (A) and endocardial (B) LV lead stimulation; RV mid septum with epicardial (C) and endocardial (D) LV lead stimulation; and RV upper septum with epicardial (E) and endocardial (F) LV lead stimulation. Stimulation points are shown in light green inside the insets for the RV lead, and in blue and red for the LV endocardial and epicardial lead, respectively.

Figure 5

Cumulative frequency histograms of the normalized percentage of activated tissue. The curves correspond to healthy (black), HF + LBBB (red) and CRT (green) scenarios. The best CRT configurations (shortest QRSd) for the three locations of the RV lead were tested: RV apex with epicardial (A) and endocardial (B) LV lead stimulation; RV mid septum with epicardial (C) and endocardial (D) LV lead stimulation; and RV upper septum with epicardial (E) and endocardial (F) LV lead stimulation.

Figure 6

Time to 90% of ventricular activation for the different CRT configuration delays assessed. (A–C) Epicardial LV lead stimulation for the three RV lead location tested: (A) RV apex, (B) RV mid septum and (C) RV upper septum. (D–F) Endocardial LV lead stimulation for the three RV lead location tested: (D) RV apex, (E) RV mid septum and (F) RV upper septum. The three LV regions (anterior, lateral and posterior walls) are shown in different color brightness (red, blue and yellow). The values for healthy and HF + LBBB configurations are depicted in black and red lines respectively.

Figure 7

Correlation between ventricular activation and QRS. (A) Correlation between QRS duration and TAT (red circles show LV epicardial leads, blue circles show LV endocardial leads). (B) Correlation between the QRS area and TAT. (C) Correlation between t₉₀QRSa and t₉₀.