Personalized Hemodynamic Analysis Based on Multimodal Medical Data and a Cardiovascular Heat Transfer Model

Non-invasive measurement of basic physiological parameters, such as blood flow and temperature, has long been crucial for daily healthcare and the early screening of cardiovascular diseases. This study aims to integrate multimodal human data with a one-dimensional arterial tree model and a lumped-parameter tissue heat transfer model to assess personalized cardiovascular hemodynamics. By matching the measured radial pressure and flow velocity waveforms of the subjects, we obtained personalized peripheral impedance and pulse wave velocities, and first predicted personalized central aortic pressure for validation of the one-dimensional (1D) cardiovascular model. The 1D cardiovascular model was further coupled with a lumped-parameter heat transfer model to investigate the interplay between blood flow and skin temperature. Skin blood perfusion and tissue layer thickness in the facial and upper limb regions were predicted based on the integrated thermo-blood flow model and the measured skin temperatures of the subject's nose tip and fingertip. We finally explored variations in cardiovascular blood flow under local heating and drowsy conditions using the developed model, based on the measured peripheral skin temperatures. The results demonstrate that the coupled cardiovascular-tissue heat transfer model effectively simulates the subjects' blood flow waveforms and skin temperature, supporting the use of non-invasive pulse wave collection devices to predict and calibrate central aortic pressure. This approach enhances our understanding of the mechanisms by which skin temperature reflects peripheral blood flow dynamics.