Web-Based CMR Teaching Atlas
The AMRG’s CMR teaching atlas, hosted by the Society for Cardiovascular Magnetic Resonance, was created in 1999 and recently updated in 2007. It comprises a comprehensive range of CMR images of a single healthy volunteer, including SSFP cine function images, myocardial tagging images, T1 anatomical images, and phase contrast flow images. This site has been extremely popular and is extensively used for education and training purposes, and is among the top Google hits for “Cardiac MRI”. The objective of the CAP is to establish a web-accessible probabilistic atlas comprising data from thousands of asymptomatic and symptomatic individuals.
The Cardiome Project
The ABI has developed accurate mathematical models of heart geometry, microstructure, material properties, stress, perfusion, cellular electromechanics and activation. Cellular mechanisms, including membrane channel characteristics, excitation-contraction coupling and cross-bridge cycling dynamics, are incorporated into a continuum description of the whole organ. This model now forms part of a world-wide effort to model the cellular, structural, mechanical and electrical properties of the heart in order to provide a framework to predict and understand the effect of gene expression, pathophysiological processes and treatment strategies on cardiac function – the Cardiome Project. The ABI also has extensive experience in design of ontologies (i.e. taxonomies together with sets of domain-specific rules for linking objects) for the classification and web-based manipulation of mathematical models.
The metadata framework in the Physiome language system is the basis for defining relations between raw data, mathematical models, and biological ontologies. The Physiome project provides general standards for the development of domain specific schema and has chosen the web ontology language OWL to define its metadata frameworks. The schema that describe the metadata specifications are a way of capturing the rules to which metadata need to conform. As an example, CellML is one of Physiome's structured markup languages developed for representing mathematical models of cellular systems.
In the wider biological community, a number of biological ontologies have been developed for the archival and classification of data. These aim to develop “innovative technology and methods that allow scientists to create, disseminate, and manage biomedical information and knowledge in machine-processable form”. The Semantic Web is a key working group, which aims to develop ontologies using common formats and principles. In order to classify, assemble and curate CMR examinations, domain specific schema for describing semantics of the cardiac image data, analyses and derived models must be developed.
Center for Computational Biology Computational Brain Atlas
The CCB has established the necessary computational infrastructure, mapping tools, and database management facilities required to facilitate and develop a computational probabilistic atlas of the brain. This infrastructure facilitates multicenter efforts to pool imaging and clinical data from a wide variety of sources and fuse them into a coherent framework. Central to their work is the development of a computational atlas of the brain: a database-like infrastructure which allows information from a variety of maps to be registered and fused into a common framework. A map is defined to be a set of relationships between points located within a coordinate space and associated features or annotations. The atlas therefore allows detailed quantitative mapping of brain structure and function, using data from a variety of sources. MRI has also been extremely useful in the brain mapping project, and there are areas of commonality between the MR brain image archive hosted by the CCB and the CMR image database to be established in the proposed research. Software has been developed for the automated registration of 3D and 2D images, classification of tissue types, graphical and image data format conversion, visualization and analysis of data. Using these tools researchers can collaboratively combine, edit, browse and query scientific information in an intuitive and efficient manner. The CAP would make use of these tools and expertise, particularly in the areas of web databases, searching, morphometrics and web based discussion and project tracking tools.
Cardiac Image Modeling Software
The PI and colleagues in the AMRG have developed functional analysis software (Cardiac Image Modeler, or CIM), which enables quantification of left ventricular mass and volumes throughout the cardiac cycle in less than five minutes. The software exploits a four dimensional (3D space and time) mathematical model of the LV which is interactively fitted to the MR images using guide point modeling. Regional as well as global Parameters of cardiac shape and function can be extracted from the model, which is intrinsically registered to the anatomy of the heart. These morphometric analysis tools lend themselves to the creation of statistical models of heart shape and function.
Parametric Distribution Models
By customizing mathematical models of the anatomy and function of the heart to individual cases, it is possible to construct parameter variation models describing the distribution of regional cardiac shape and function across patient subgroups. Cootes et al. pioneered the application of Point Distribution Models in computer vision problems. Homologous landmarks (i.e. the points which are aligned to match corresponding features in the shape) were used to characterize shape and shape variations with the aid of a principal component analysis. A major advantage of the CAP will be the availability of a large number of examinations and derived contours, which can be used to in conjunction with finite element modeling methods to develop statistical evaluation and classification tools.
