Integrating Imaging and Invasive Pressure Data into a Multiscale Whole-Heart Model

Area of Science:

• Biomedical Engineering
• Computational Biology
• Cardiovascular Physiology

Background:

• Cardiovascular diseases are a leading cause of mortality, with treatment decisions hindered by complex clinical data integration.
• Personalized physics-based models offer potential but face calibration and validation challenges.

Purpose of the Study:

• To present a novel, systematic calibration method for a whole-heart, multi-scale, electromechanics model.
• To validate the model's ability to replicate cardiac function and response to therapy using clinical data.

Main Methods:

• Employed emulators, sensitivity analysis, and history matching for systematic model calibration.
• Utilized ECG-gated CT and invasive LV pressure data to calibrate 25 model parameters.
• Validated against CT-derived motion, chamber configurations, and hemodynamic response to biventricular pacing.

Main Results:

• Achieved precise calibration, fitting key cardiac features within 0.8-10.8% of target values and 1.4 standard deviations.
• Demonstrated accurate replication of atrioventricular plane displacement and end-diastolic/end-systolic configurations.
• Successfully simulated the hemodynamic response to biventricular pacing, closely matching clinical measurements.

Conclusions:

• The developed systematic calibration method enables robust integration of clinical data into whole-heart electromechanics models.
• The validated model accurately captures local heart motion and therapeutic responses, showing promise for clinical decision support in cardiovascular disease management.