Software for Practical Annotation and Exchange of Virtual Anatomy

点击翻译按钮获取中文摘要

Virtual representations of anatomy are ubiquitous for interpretation of physiological function in health and disease. While computational description of anatomy have relied on a diverse array of image-based and CAD formats, mesh-based approaches have become increasingly important to assist clinical care, medical device development, additive manufacturing, and computational simulations. Anatomical geometry alone is not necessarily useful without appropriate annotation. Annotation enables descriptions of anatomical relationships and physiological interactions between organs, and encapsulation of metadata. It can also capture disease related information for diagnosis and prognosis. Properties can be assigned, i.e., the physiological state, as a field variable across the object’s spatial domain, can be stored. An emerging need for virtual anatomy is software to unify general purpose and specialized annotation of computational representations of anatomical objects. The software should include features for automated annotation, i.e., relying on taxonomy of anatomical relationships, and for interactive annotation, both of which increase users’ efficiency to markup anatomical objects. The software should also facilitate exchange between formats to empower the biomedical community with reuse and further annotation of virtual anatomies. At the moment, formats to store and exchange anatomical data are widely fragmented. Many are designed for specific purposes and accommodate only a subset of annotation capacity. There are a limited number of standards proposals, which are mostly specialized and awaiting traction in the community. To fulfill these unmet needs, we propose a novel software suite for pragmatic annotation and exchange of virtual anatomy by leveraging mesh-based representations and supporting image- and CAD-based formats. The project will be accomplished through three specific aims. First, a multi-platform software library will be developed as a foundation. Second, a high level application interface, a graphical user interface, and a plug-in infrastructure will be developed. These components will increase the usability potential of the software and will also provide the capacity for extensions using plug-ins. Various plug- ins will be developed by the investigators to support a large variety of anatomical data exchange formats. Finally, use cases from musculoskeletal, cardiovascular, and neurological domains will demonstrate the software’s general applicability in a wide range of biomedical disciplines and curate sample data sets to guide prospective annotation and exchange conventions. The Cleveland Clinic investigators will act as the application scientist team to provide specifications for software development and to implement case studies. The Kitware, Inc. group will utilize an agile software development workflow to deliver software components. Free and open source implementation will ensure utmost engagement with the prospective user and developer communities. In a nutshell, the project will result in a functionally evaluated and scientifically validated software package to efficiently work on virtual anatomy for a diverse range of biomedical applications.