Approved research

Construction of Multi-dimensional Statistical Shape Model for a heart

Lay summary

Statistical Shape Model (SSM) of a target organ quantitatively describes the average and variance of the organ among individuals. Since the SSM enables to estimate the whole shape of a target organ from its partial shape data, SSM is used in various computer-aided treatment and diagnosis systems. Conventional SSMs describe the shape variation of motionless organs (e.g. bone). On the other hand, few description methods consider the deformation of organs (e.g. heart). The aim of our research is to construct a new multi-dimensional SSM for the heart that describes not only inter-individual differences of cardiac shape but also the cardiac deformation. One potential application using the multi-dimensional SSM for the heart is to estimate the movement of the patient heart in a cardiac cycle from the partial shape of the patient heart at a certain time. Accordingly incorporating the multi-dimensional SSM for the heart into a computer-aided system leads to the improvement of the efficient and highly accurate treatment / diagnosis for hearts.

Scope extension:

Statistical Shape Model (SSM) of a target organ quantitatively describes individual differences of the organ shapes using their average and variance. Since the SSM enables to estimate the whole shape of a target organ from its partial shape data, SSM is used in various computer-aided treatment and diagnosis systems.

 Although conventional SSM constructions have focused on only motionless organs such as bones, there are few SSMs for a heart that describe the deformation of the heart. Moreover, hearts have an internal structure. Therefore, the SSM for the heart needs to describe the deformation of both the entire heart and the internal structure in a cardiac cycle. These factors make it more difficult to construct the SSM for heart. The SSM for the heart has the potential of estimating the movement of the patient heart in a cardiac cycle from the partial shape of the patient heart at a certain time. Incorporating the SSM for the heart into a computer-aided system leads to the efficient and highly accurate treatment/diagnosis for hearts.

Our research aim is to construct a multi-dimensional SSM for the heart that considers not only shape differences among individuals but also sequential deformation within the cardiac cycle. Moreover, by analyzing correlations among the cardiac motion obtained from the SSM, lifestyle, and genetic information, we aim for constructing a correlation model which is helpful for radiogenomics.

The results of previous experiments have implicated that to improve the accuracy of radiogenomics, not only lifestyle and genetic information but also various patient metadata including cholesterol levels, ECGs and family history are required. We therefore extend our research objectives to the construction of a correlation model that includes these metadata.