Profile-based sharing of medical imaging data

Abstract

Over the past few years, the healthcare industry had been relying on Information Technologies more than ever. Particularly in the medical imaging departments, the information flow and services available for medical practice had a huge increase. Consequently, also medical imaging repositories are becoming bigger and hard to manage and scale. Institutions still deal with this using the traditional approach based on a single installation of medical imaging repository serving the entire organization and are continuously improving the infrastructure to efficiently support the new service reality. Some institutions have already outsourced the storage service to the cloud. However, several issues associated with the specificity of medical imaging are still hard to solve. For instance, a universal service with the geographic distribution of repositories is very complex due to the volume of data that can reach the gigabyte scale per sample in some scenarios. So, current approaches lack support for advanced usage scenarios and the challenges are yet to be overcome. It is becoming urgent to explore new methods to manage the big data that is being produced on a daily basis, with a high level of abstraction and simplified integration with nowadays systems since it is imperative to not disrupt the established medical practice methods and support new usage scenario like, for instance, research. This thesis researched and proposes a distributed architecture for the Web storage and sharing of medical imaging, in a secure and scalable way. It makes use of standard formats and network protocols to ensure transparent integration with market equipment. Proposed methods ensure a service layer abstraction and intend to turn the medical imaging content discovery and retrieval more efficient and performant. To validate the solution, the performance of architectural components was evaluated and implemented in two advanced scenarios: research and digital pathology.