Institute of High Performance Computing


Powerding Discoveries!



Computing Science (CS)


Patient-Specific Computational Cardiology

The human heart is a highly complex biological system. The understanding of each component of the heart, as well as its whole system, represents a grand challenge for the scientific, research and medical communities. Our motivation for the study of the heart is driven not only by natural curiosity but also the many benefits that would result in health care and medical practice. Working closely with medical practitioners, researchers in IHPC strive to realise a synergistic confluence between the biomedical and computational sciences, and to revolutionise our understanding of the human heart and our ability to influence quality of life.

Functional Assessment of Left Ventricle

The main function of the heart is to pump blood throughout the circulatory system. In particular, the left ventricle (LV) contracts and pumps oxygenated blood to distant parts of the body at high pressure during systole, and is often considered as the most important chamber of the heart. Numerous efforts have been dedicated to the analysis and understanding of LV mechanism and function. However, it remains a challenging problem with much to be explored.

Collaborating with doctors from the National Heart Centre, Singapore, we devise a method to compute 3D regional LV surface curvature descriptors, expressed in term of local curvedness.

A framework was developed for the analysis of LV regional shape as applied to normal subjects and patients with dilated cardiomyopathy. The methodology is based on local surface fitting and differential geometry techniques to interrogate LV geometry at end-diastole and end-systole. These geometrical models were then used to derive regional curvedness and the corresponding variation coefficient.

Comparison of percent curvedness change (C%) between diseased and normal heart

We have observed that LVs with dilated cardiomyopathy have smaller curvedness value and lower percentage curvedness change as compared with normal subjects. These encouraging initial results provide impetus for further studies to analyse the relationship between quantitative shape parameters and regional function, as well as to determine the sensitivity and specificity of our method compared to traditional wall motion analysis measurements.

Haemodynamic Study

Working with another group of doctors from the National University Hospital, Singapore and researchers from Swansea University, United Kingdom, we seek to understand the complex transport mechanism in the left heart, which might potentially facilitate the diagnosis of cardiac dysfunction. To perform a comprehensive investigation of the intraventricular flow dynamics, we use a combination of Cardiac Magnetic Resonance (CMR) imaging techniques, computational 4D geometrical modelling techniques and Computational Fluid Dynamics (CFD) simulation.

CFD being an effective means of studying complex flow dynamics, can provide detailed haemodynamic information that is not acquirable from direct flow imaging techniques. Unlike models with simplified shape, accurate endocardial geometries will be derived from image-based patient-specific data using Magnetic Resonance Imaging (MRI). In most existing LV flow simulations, structural entities such as the mitral valves, aorta and atrium are often not modelled accurately. However the intricate valvular shape and inflow boundary conditions are well recognized to influence the accuracy of simulations. Also, atrioventricular plane displacements are often used as important indices of LV function. Therefore, our research includes details of these important anatomical structures to achieve accurate simulations.

The goal is to realise a framework for fully automatic computer-aided simulation and analysis in the context of cardiological study. This involves the development, implementation and validation of various technologies that serve as building blocks to the framework.

Reconstruction of detailed mesh model of left heart


Back to Research Highlight

This page is last updated at: 26-MAY-2009