GEMSS |
The Grid-enabled radiosurgery application uses RAPT as a front end to the EGS Monte Carlo engine to model ionising radiation transport through the head of the patient. The simulation requires:
The results archive is pulled from the Grid in encrypted form and unzipped on the client within the working Output directory. This data is accessible to the user as a set of text files, but a visualization application has been provided to enable the user to explore the distribution of radiation dose within the treatment volume. The dose data can be represented as contours on specified planes through the volume mesh or visualised as isosurfaces in 3D. The job submission process is outlined in this figure. |
Treatment set-up Bubble head set-up Computed dose contours |
The primary role of the Grid in this application is to provide sufficient accessible computing power for the radiosurgery simulation to be completed in a fraction of an hour. However, in addition to compute issues, the GEMSS Grid has also implemented a means of costing and billing the user in accordance with the degree of Grid utilisation. Radiosurgery planning is a time critical application and therefore the process of job submission includes specification of the deadline by which the solution is to be completed, accompanied by suitable guarantees from the Grid provider (accommodated by the Quality of Service (QoS) component of the GEMSS middleware). Further requirements are associated with integrity and security of data transfer. These are themes that are common to other Grid applications and are key goals that have been addressed by the GEMSS consortium (e.g. adoption of X.509). In this application, the anticipated end-user is the clinical radiosurgery department, with the RAPT software operated by a technical expert (physicist). The Monte Carlo simulation is used to verify the treatment doses calculated by the more traditional (and more approximate) methods of GammaPlan. In order that the application be accessible, the Grid enabled application has been designed to address user-interface, accessibility, bandwidth, processor scalability and storage issues in such a way that it provides an effective means of solving real world problems in the desired timescale. |