Tamas Hauer

Experiences of engineering Grid-based medical software

OBJECTIVES Grid-based technologies are emerging as potential solutions for managing and collaborating distributed resources in the biomedical domain. Few examples exist, however, of successful implementations of Grid-enabled medical systems and even fewer have been deployed for evaluation in practice. The objective of this paper is to evaluate the use in clinical practice of a Grid-based imaging prototype and to establish directions for engineering future medical Grid developments and their subsequent deployment. METHOD The MammoGrid project has deployed a prototype system for clinicians using the Grid as its information infrastructure. To assist in the specification of the system requirements (and for the first time in healthgrid applications), use-case modelling has been carried out in close collaboration with clinicians and radiologists who had no prior experience of this modelling technique. A critical qualitative and, where possible, quantitative analysis of the MammoGrid prototype is presented leading to a set of recommendations from the delivery of the first deployed Grid-based medical imaging application. RESULTS We report critically on the application of software engineering techniques in the specification and implementation of the MammoGrid project and show that use-case modelling is a suitable vehicle for representing medical requirements and for communicating effectively with the clinical community. This paper also discusses the practical advantages and limitations of applying the Grid to real-life clinical applications and presents the consequent lessons learned. CONCLUSIONS The work presented in this paper demonstrates that given suitable commitment from collaborating radiologists it is practical to deploy in practice medical imaging analysis applications using the Grid but that standardization in and stability of the Grid software is a necessary pre-requisite for successful healthgrids. The MammoGrid prototype has therefore paved the way for further advanced Grid-based deployments in the medical and biomedical domains.

Semantic Visualization of Patient Information

Clinical practice and research rely increasingly on analytic approaches to patient data. Visualization enables the comparative exploration of similar patients, a key requirement in certain clinical decision support systems. Patient data is complex and heterogeneous, may have different formats, reside in various structures and carry different semantics. This makes the comparison and analysis of clinical data a challenging task. Most medical applications visualize patient data without integrating additional semantic information to structure the analysis. Our objective is to map patient data onto relevant fragments of ontologies and inferred ontological structures as a basis for improved patient data visualization, comparison, and analysis. Two visualization scenarios that we have implemented using the patient data acquired in the Health-e-Child project will be presented and their clinical evaluation will be provided.

The Requirements for Ontologies in Medical Data Integration: A Case Study

Evidence-based medicine is critically dependent on three sources of information: a medical knowledge base, the patients medical record and knowledge of available resources, including where appropriate, clinical protocols. Patient data is often scattered in a variety of databases and may, in a distributed model, be held across several disparate repositories. Consequently addressing the needs of an evidence- based medicine community presents issues of biomedical data integration, clinical interpretation and knowledge management. This paper outlines how the Health-e-Child project has approached the challenge of requirements specification for (bio-) medical data integration, from the level of cellular data, through disease to that of patient and population. The approach is illuminated through the requirements elicitation and analysis of Juvenile Idiopathic Arthritis (JIA), one of three diseases being studied in the EC-funded Health- e-Child project.

Medical Data Integration and the Semantic Annotation of Medical Protocols

The integration of heterogeneous biomedical information is one important step towards providing the level of personalization required in the next generation of healthcare provision. In order to provide the computer-based decision support systems needed to access this integrated healthcare information it will be necessary to handle the semantics of (amongst other things) medical protocols. The EC FP6 Health-e-Child project aims to develop an integrated healthcare platform for European paediatrics and decision support tools to access personalized health information. This paper introduces both the integrated data model in the Health-e-Child project and through a case study using the brain tumour protocols it demonstrates the semantic annotation of patient data acquired in the project using UMLS as the primary source of semantic data.

An Architecture for Semantic Navigation and Reasoning with Patient Data - Experiences of the Health-e-Child Project

Medical ontologies have become the standard means of recording and accessing conceptualized biological and medical knowledge. The expressivity of these ontologies goes from simple concept lists through taxonomies to formal logical theories. In the context of patient information, their application is primarily annotation of medical (instance) data. To exploit higher expressivity, we propose an architecture which allows for reasoning on patient data using OWL DL ontologies. The implementation is carried out as part of the Health-e-Child platform prototype. We discuss the use case where ontologies establish a hierarchical classification of patients which in turn is used to aid the visualization of patient data. We briefly discuss the treemap-based patient viewer which has been evaluated in the Health-e-Child project.

Semantic Matching Using the UMLS

Traditional ontology alignment techniques enable equivalence relationships to be established between concepts in two ontologies with some confidence value. With semantic matching, however, it is possible to identify not only equivalence (***) relationships between concepts, but less general (

Health-e-Child: A Grid Platform for European Paediatrics

\sqsubseteq

A Grid Information Infrastructure for Medical Image Analysis

) and more general relationships (

Grid databases for shared image analysis in the MammoGrid project

\sqsupseteq

A data model for integrating heterogeneous medical data in the health-e-child project

). This is beneficial since more expressive relationships can be discovered between ontologies thus helping us to resolve heterogeneity between differing semantic representations at a finer level of granularity. This work concerns the application of semantic matching to the medical domain. We have extended the SMatch algorithm to function in the medical domain with the use of the UMLS metathesaurus as the background resource, hence removing its previous reliance on WordNet, which does not cover the medical domain in a satisfactory manner. We describe the steps required to extend the SMatch algorithm to the medical domain for use with UMLS. We test the accuracy of our approach on subsets of the FMA and MeSH ontologies, with both precision and recall showing the accuracy and coverage of different versions of our algorithm on each dataset.