The aim of my lifetime research quest is the modeling of the entire cardiovascular system in the human body. This interdisciplinary project requires a complex interplay of mechanisms that occur across a number of time and length scales, from 0D to 7D. The strong coupling of these mechanisms is partly still unknown and some interactions are quite a challenge to investigate in-vivo. This work brings together high level knowledge of human physiology, molecular biochemistry, material properties and design, biosensors, computational fluid dynamics, medical image processing, signal processing, algorithms, numerical methods, cardiovascular pressure wave analysis, lumped-parameter models, and mathematical models.
The physiology of the human cardiovascular system is split into components: the heart, the valves, and the vasculature. Designing new physiological sensors and finding improved high-speed data acquisition methods increase the accuracy of the provided clinical data. Mathematical and computational models for each cardiovascular component can be accurately designed by analysing the clinical data. The final step is to find methods that bring together the components into one model of the whole cardiovascular system.
The human nervous system represents a set of complex neural pathways that may be controlled by hormones in order to vary the cardiac muscle contraction, the heart rate, the vasodilation, the peripheral resistance and the blood pulse pressure. The cardiac tissue information is mapped in both spatial and temporal direction using computed tomography, contrast angiography, or echocardiography. Magnetic resonance imaging is used primarily as problem-solving tool: it provides, as with X-ray CT, high resolution anatomic structure but, in addition to the information that the CT scan offers, it provides high contrast between different soft tissues. The physical basis of soft tissue contrast and the enhancement mechanism with exogenous contrast materials are different for the two imaging modalities. Both CT and MRI have the ability to change the imaging plane without moving the patient, but it takes longer to acquire an MRI scan than with CT and the MRI screening is more susceptible to patient motion.
Flexibility is both strength and weakness of MRI. The number of ways for MRI screening of the chest is virtually unlimited, but not all imaging sequences can be applied to every patient. The protocol of MRI acquisition, with respiratory and cardiac motion compensation, needs to be designed to answer the project’s specific clinical question. Gadolinium-enhanced 3D MR angiography volume scan depicts best the origins and direction of branch vessels. In-plane or through-plane 2D+time phase-contrast velocity- encoded data captures the flow development in the aortic arch. The left ventricular outflow tract is captured in 1-3 slices of ‘white blood’ 2D+time steady-state free precessing in sagittal or coronal plane. The same modality, but with slices orientated transverse to the aortic valve, is known to capture the leaflets of the aortic valve. The single-phase ‘black blood’ 2D spin-echo scan, gated to systole to achieve the optimal black-blood to tissue contrast, captures the aortic arch and the thoracic aorta, in- and through-plane.
The fluid-structure interaction is a main topic in computational cardiovascular modeling. It investigates the strong coupling between the unsteady hemodynamics and the structure of the cardiovascular system components: the vessel walls, the valves, and the blood cells. Blood is either modeled as a set of flexible structures suspended in a fluid, or at larger scales simply as a continuum fluid with “non-Newtonian” behaviour. Incompressible fluid mechanics is the most accurate choice to model the system.
The pressure waveform of the blood passing through the vessels, along the vessel centreline, varies according to position. The morphology changes as a result of wave reflexion at bifurcations, and due to vessel tapering. The changes in vessel diameter are captured in medical images, by having an appropriate voxel size in the CT image. There are three wave analysis methods: the augmentation index, the wave intensity analysis in time domain and the frequency domain analysis. Accuracy in computing pulse pressure is affecting the computation of the value for the aortic compliance and characteristic impedance at difference measurement sites.
The model of the cardiovascular system is of great need to be integrated in the current clinical practice in order to provide a risk-free non-invasive analysis and the ability to understand key physiological mechanisms. The main requirements of the model is to be accurate and reliable. This can be achieved only after a large cohort of anonymised patient-specific clinical data has passed though the processing workflow that proves to have reliable repeatability of testing.
Improvements in the ability of computational fluid-structure interaction modeling and that of electrical modeling to provide results that are validated with accurate clinical data increase the feasibility of integrating these methods into diagnostic tools used in day-to-day clinical practice.
