Virtual pacing of a patient’s digital twin to predict left ventricular reverse remodelling after cardiac resynchronization therapy

Figures

• 

  Figure 1. Overview of the input and output of the digital twin generation algorithm.12 Left ventricular 18-segment longitudinal strain, LVEDV, and LVESV measurements of the patient (upper left panel) were used to automatically estimate cardiac output, global LV mechanical activation duration, regional LV myocardial constitutive properties, and regional LV and RV onset of mechanical activation (right panel). As a result, a personalized model (DT) was obtained, which resembled the patient (lower left panel). DT, digital twin; LV, left ventricular; LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume; RV, right ventricular.

• 

  Figure 2. Bullseye plots of RV and LV regional mechanical activation at baseline and during virtual biventricular pacing in the DT. Starting from the baseline DT with automatically estimated onset of activation in the RV wall and LV wall segments (Baseline), a personalized pacing configuration was simulated in which activation was initiated at the location of the RV and patient-specific LV lead, indicated by the black and white asterisks, respectively (Pacing). From the pacing lead locations, activation propagated radially with an intersegmental delay of 15 ms. The patient illustrated here had a L LV lead location, indicated by the circled asterisk. Furthermore, patient-specific AV delay was set to 130 ms based on the pacing protocol of the original study. Note that definition of myocardial segment location differs slightly from that of LV lead location. A, anterior; AL, anterolateral; AS, anteroseptal; I, inferior; IS, inferoseptal; L, lateral; LV, left ventricular; P, posterior; PL, posterolateral; RV, right ventricular.

• 

  Figure 3. The association between septal-to-lateral work imbalance reduction by virtual pacing in the DT and long-term clinical LV reverse remodelling after CRT device implantation, illustrated for a single patient with lateral LV lead location. The DT of this patient demonstrated a relatively large pacing-induced reduction of septal-to-lateral myocardial work difference (ΔMWLW-S,DT), which corresponded with a relatively large reduction of LV end-systolic volume (ΔLVESV) in the patient at clinical follow-up. CRT, cardiac resynchronization therapy; LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume; LW, lateral wall; S, septum.

• 

  Figure 4. Correlation plots of relative LV volume change 6 months after CRT in the patient (ΔLVESV) vs. baseline DT septal-to-lateral myocardial work difference (MWLW-S,DT, A) and indices derived from virtual pacing of the DT: absolute change in MWLW-S,DT (ΔMWLW-S,DT, B), and absolute change of maximum rate of LV systolic pressure rise (ΔdP/dtmax,DT, C). The blue dots (n = 37) represent patients with a Class I indication for CRT, while the red dots (n = 8) represent patients with a Class IIa or IIb indication for CRT, according to the Strauss criteria for LBBB.17 Indices MWLW-S,DT and ΔMWLW-S,DT were significantly associated with ΔLVESV (r = −0.60, P < 0.001 and r = 0.62, P < 0.001, respectively), while ΔdP/dtmax,DT correlated considerably weaker with ΔLVESV (r = −0.34, P = 0.02).