Medical image computing raises new challenges related to the scale and complexity of the required analysis, for example in studies that require the federation of large data sets or in complex models and processing. Grid technology are addressing problems related to large data sets manipulation over wide computing networks, providing tools for exchanging data and computing power and additionally serving as a vector for structuring the user communities as they enable cross-enterprises collaborations. In the medical imaging area, grids provide a foundational layer that can be exploited e.g., to build patient-specific models, reduce computing time to meet clinical practice constraints, algorithms validation and optimization, or collaborative studies on rare diseases. Specific grids initiatives are emerging worldwide, demonstrating a growing interest from the health community for such infrastructures and impacting the way to conduct medical research. However, deploying medical image analysis applications on grid infrastructures requires a proper understanding of the specific needs in this area.
The Workshop on Medical imaging on grids: achievements and perspectives (MICCAI-Grid) was organized with the goals of providing an opportunity to demonstrate the current achievements of grid technologies within the medical imaging community; identifying the fundamental problems limiting the adoption of existing systems and methods; and stimulating the community to build new collaborations by taking advantage of the sharing capabilities of grids. The workshop was organized in conjunction with the MICCAI conference in New York, USA, in 2008. Seventeen papers were submitted to the workshop, representing research developed in 10 countries around the globe. Each paper has been reviewed by three independent reviewers from an international scientific committee with representatives from Europe, North and South America, and Asia.
The final program consists of seven papers selected for full presentations, three papers selected for short presentations, and additionally two invited talks. The invited talks by Stephan G. Erberich and David Manset illustrate successful stories from projects in the USA and Europe that are applying grid technologies for medical imaging applications, namely Globus MEDICUS, Health-e-Child and neuGRID.
The papers discuss a large variety of topics, including frameworks for sharing imaging data and algorithms [Tromans et al.], workflows for medical image analysis [Glatard et al., Krefting et al.], and aids to run and manage medical imaging software on grids [Ruiz et al., Acher et al.]. The presented work additionally illustrates how the large computing capacity of grids have enabled medical imaging studies that would not have been possible to accomplish otherwise [Pernod et al., Aoun et al., Soleman et al.]. Finally, the papers also report about the deployment of grid systems in clinical environments [Niinimaki et al.] and explore the potential of grid-enabled simulators of imaging devices for virtual radiology [Camarasu et al.].
This interesting program represents the end of a journey that was possible due to the collaboration of many (organizers, reviewers, invited speakers, authors, sponsors). The MICCAI 2008 organization provided shelter to this workshop in this important conference. The members of the program committee promptly and enthusiastically embraced the idea, prepared high-quality reviews of the papers and (!) returned them sharp on time. The invited speakers were willing to leave their busy routines and share their expertise at the workshop in NY. The authors trusted the workshop, submitted their research work, sending materials in the indicated format, within the page limit, at the requested time (!). And last but not least, the workshop sponsors (Virtual Laboratory for e-Sciences project, NL and GDR ASR, CNRS, FR) provided funds to facilitate the trip of the invited speakers.
We wish the MICCAI-Grid workshop will leave up to this promising start.
New York, 6 September 2008
Sílvia D. Olabarriaga
Academic Medical Center
University of Amsterdam
Johan Montagnat and Diane Lingrand
CNRS / Univ. of Nice-Sophia Antipolis