AIM:To demonstrate the use of grid technology to produce a database of mammograms and supporting patient data, specifically using breast density as a biomarker of risk for breast cancer, for epidemiological purposes. METHOD:The cohort comprised 1737 women from the UK and Italy, aged 28-87 years, mean 54.7 years, who underwent mammography after giving consent to the use of their data in the project. Information regarding height, weight, and exposure data (mAs and kV) was recorded. The computer program Generate-SMF was applied to all films in the database to measure breast volume, dense breast volume, and thereby percentage density. Visual readings of density using a six-category classification system were also available for 596 women. RESULTS:The UK and Italian participants were similar in height, but the UK women were significantly heavier with a slightly higher body mass index (BMI), despite being younger. Both absolute and percentage breast density were significantly higher in the Udine cohort. Images from the medio-lateral projection (MLO) give a significantly lower percentage density than cranio-caudal (CC) images (p<0.0001). Total breast volume is negatively associated with percentage density, as are BMI and age (p<0.0001 for all), although 80% of the variability in percentage density remains unexplained. CONCLUSION:The study offers proof of principle that confederated databases generated using Grid technology provide a useful and adaptable environment for large quantities of image, numerical, and qualitative data suitable for epidemiological research using the example of mammographic density as a biomarker of risk for breast cancer.

This paper describes the prototype for a Europe-wide distributed database of mammograms entitled MammoGrid, which was developed as part of an EU-funded project. The MammoGrid database appears to the user to be a single database, but the mammograms that comprise it are in fact retained and curated in the centres that generated them. Linked to each image is a potentially large and expandable set of patient information, known as metadata. Transmission of mammograms and metadata is secure, and a data acquisition system has been developed to upload and download mammograms from the distributed database, and then annotate them, rewriting the annotations to the database. The user can be anywhere in the world, but access rights can be applied. The paper aims to raise awareness among radiologists of the potential of emerging "grid" technology ("the second-generation Internet"). (C) 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

Medical conditions such as breast cancer, and mammograms as images, are extremely complex with many degrees of variability across the population. An effective solution for the management of disparate mammogram data sources that provides sufficient statistics for complex epidemiological study is a federation of autonomous multi-centre sites which transcends national boundaries. Grid-based technologies are emerging as open-source standards-based solutions for managing and collaborating distributed resources. In the light of these new computing solutions, the MammoGrid project, as one example of a HealthGrid, is developing a Grid-aware medical application which manages a European-wide database of mammograms. The MammoGrid solution utilizes the grid technologies in seamlessly integrating distributed data sets and is investigating the potential of the Grid to support effective co-working among mammogram analysts throughout the EU.

The MammoGrid project has recently delivered its first proof-of-concept prototype using a Service-Oriented Architecture (SOA)-based Grid application to enable computing spanning national borders. The underlying AliEn Grid infrastructure has been selected because of its practicality and its emergence as a potential open source standards-based solution for managing and coordinating distributed resources. The resultant prototype is expected to harness the large amounts of medical image data needed to perform epidemiological studies, advanced image processing and ultimately tele-diagnosis over communities of 'virtual organizations'. This paper outlines the MammoGrid approach in managing a federation of Grid-connected mammography databases in the context of the recently delivered prototype and describes the next phase of prototyping.

The MammoGrid project aims to prove that grid infrastructures can be used for collaborative clinical analysis of database-resident but geographically distributed medical images. This requires: a) the provision of a clinician-facing front-end workstation and b) the ability to service real-world clinician queries across a distributed and federated database. The MammoGrid project will prove the viability of the grid by harnessing its power to enable radiologists from geographically dispersed hospitals to share standardized mammograms, to compare diagnoses (with and without computer aided detection of tumours) and to perform sophisticated epidemiological studies across national boundaries. This work outlines the approach taken in MammoGrid to seamlessly connect radiologist workstations across a grid using an information infrastructure and a DICOM-compliant object model residing in multiple distributed data stores in Italy and the UK.