An integrated computational workflow for myocardial perfusion combining the multiple-network poroelastic theory and subject-specific imaging data

Microvascular dysfunction plays an early and potentially independent role in coronary artery disease, yet remains difficult to assess due to limitations in current imaging resolutions and diagnostic tools. To address this gap, we present an integrated computational workflow for subject-specific myocardial perfusion modeling based on the multiple-network poroelastic theory (MPET). The myocardium is represented as a poroelastic medium containing three interacting fluid networks: arterial, arteriole/capillary, and venous, to capture the multiscale dynamics of myocardial perfusion. Specifically, the workflow integrates realistic left ventricular (LV) morphology and physiologically informed coronary inflow patterns derived from routine magnetic resonance (MR) imaging into the MPET framework. Sensitivity analysis guided parameter calibration to ensure physiological relevance of the myocardial perfusion model. The simulation results showed that the model reproduced physiologically realistic myocardial blood flow patterns and transmural perfusion gradients, in agreement with clinical MR perfusion imaging. Notably, the model demonstrated sensitivity to myocardial wall thickness, highlighting its potential for assessing structural-functional relationships.