Personalized biventricular mechanics and sensitivity to model morphology

Area of Science:

• Computational biology
• Biomedical engineering
• Cardiovascular research

Background:

• Patient-specific computational models are crucial for understanding cardiac mechanics.
• Standard clinical data (ECG, blood pressure, CTA) offer potential for personalized cardiac modeling.

Purpose of the Study:

• To create a computational framework for patient-specific cardiac mechanics models using standard clinical data.
• To compare the accuracy of a full BiV model against simplified t-BiV and LV models.

Main Methods:

• Developed a computational framework integrating ECG, blood pressure, and CTA data.
• Coupled a BiV model to a closed-loop LPN circulatory model with rule-based fiber architecture.
• Personalized model parameters via sequential tuning of circulatory dynamics, passive mechanics, and active contraction.
• Compared BiV model performance against truncated BiV (t-BiV) and left ventricle-only (LV) models.

Main Results:

• The personalized BiV model accurately reproduced clinical pressure, volume, and myocardial deformation data.
• LV-only models showed similar global pressure/volume behavior but moderate regional deformation differences.
• Truncated BiV (t-BiV) models demonstrated significant discrepancies in both global function and local mechanics.

Conclusions:

• Patient-specific BiV models derived from clinical data provide accurate cardiac mechanics.
• LV-only models may suffice for global biomechanical studies.
• Truncating the BiV model at the basal plane significantly impacts results and requires cautious application.