In keeping with the theme of this year's e-Science All Hands Meeting--past, present and future--we consider the motivation for, the current status of, and the future directions for, the technologies developed within the GIMI (Generic Infrastructure for Medical Informatics) project. This analysis provides insights into how some key problems in data federation may be addressed. GIMI was funded by the UK's Technology Strategy Board with the intention of developing a service-oriented framework to facilitate the secure sharing and aggregation of heterogeneous data from disparate sources to support a range of healthcare applications. The project, which was led by the University of Oxford, involved collaboration from the National Cancer Research Institute Informatics Initiative, Loughborough University, University College London, t+ Medical, Siemens Molecular Imaging and IBM UK.

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.

BACKGROUND:The UK Medical Devices Agency has suggested that ophthalmic practitioners should, where practicable and not compromising clinical outcome, restrict corneal contact devices to single patient use to minimise a remote theoretical risk of transmission of new variant Creutzfeldt-Jakob disease (vCJD). This study reports on a modified technique of ultrasound A-scan biometry that complies with the MDA recommendations.METHODS:The right eyes of 37 consecutive hospital patients had a series of biometry readings taken with a Humphrey 820 A-scan instrument with a plane wave transducer use d conventionally and with the addition of a disposable latex cover.RESULTS:Intrasessional repeatability of axial length measurements was similar for conventional readings--mean difference 0.027 mm, 95% confidence intervals (CI) +/- 0.44 mm and those taken with a disposable cover (0.028 mm, CI +/- 0.38). Intersessional repeatability was equivalent with (0.002 mm, CI +.- 0.51) and without a cover (0.03 mm, CI +/- 0.51). Readings with a cover were not significantly different from those without (paired t test; p >0.05), but tended to be greater (mean difference 0.085 mm, CI +/- 0.60).CONCLUSIONS:These findings suggest that corneal contact biometry with a disposable cover is a viable and theoretically safer alternative to the conventional technique.

Aims - To determine the accuracy of intraocular lens (IOL) power calculation in a group of pseudophakic children.Methods - A retrospective analysis of biometric and refractive data was performed on 52 eyes of 40 infants and children, who successfully underwent cataract extraction and IOL implantation. The following parameters were included: age at the time of surgery, keratometry, axial length, estimated refraction, and the power of IOL implanted. The postoperative refractive outcome was taken as the spherical equivalent of the refraction at 3 months after surgery. The prediction error was taken as the absolute difference between the estimated and actual postoperative refraction. The data were analysed to assess the effects of age at the time of surgery, keratometry, and axial length on the accuracy of calculation of IOL power.Results - For the overall group the mean and median prediction errors were 1.40 D and 0.84 D (SD 1.60). The mean and median prediction errors in eyes with axial lengths greater than or equal to 20 mm were 1.07 D and 0.71 D (SD 0.98) and in eyes < 20 mm. were 2.63 D and 2.61 D (SD 2.65). The mean and median prediction errors in eyes in children aged greater than or equal to 36 months were 1.06 D and 0.68 D (SD 1.02) and in children aged < 36 months was 2.56 D and 2.29 D (SD 2.50). The differences between the prediction errors for both axial length and age were statistically significant (p <0.05).Conclusions - For the overall group IOL, power calculation is satisfactory. In eyes with axial lengths less than 20 mm. and in children less than 36 months of age larger errors can arise. This study demonstrates the need for an IOL formula specifically designed for paediatric use.