Sensitivity analysis and optimization of a cardiovascular lumped parameter model for patient-specific modelling

Parameter estimation poses a significant challenge in developing patient-specific cardiovascular models. This study presents a framework that enhances parameter estimation in lumped parameter cardiovascular models by combining sensitivity analysis for parameter selection with multi-objective genetic algorithm optimization. Four key parameters were identified as the most influential and subsequently optimized. Model outputs, specifically mean arterial pressure (MAP), were validated against clinical values from a public database. The optimized model's MAP demonstrated a strong correlation with clinical MAP (<i>r</i> = 0.99997, <i>p</i> < 0.001), and a t-test analysis (<i>p</i> = 0.752) confirmed statistical equivalence with clinical data. This approach highlights the potential of sensitivity analysis and genetic algorithms to improve accuracy in patient-specific cardiovascular modelling.