A peer-to-peer environment to share medical images and diagnoses providing context-based searching

The enormous amount of radiological data obtained nowadays comprises a huge quantity of clinical knowledge that it is not efficiently exploited. Although many peer-reviewed radiological databases are available, the number of cases and the variety of pathologies are not enough for research. The most of the information remains locally in hospitals and clinics without being shared. Sharing medical studies is not trivial. The vast amount of data that image studies involve, the difficulty on indexing the information and the lack of adequate tools prevent radiologists from sharing information at a larger scale. This paper presents a peer-to-peer (P2P) environment that assists image diagnosis apprenticeship and research, enabling radiologists to share DICOM studies and its corresponding diagnoses. The P2P environment presented in the paper proposes a solution to ease searching and securely sharing medical data. Automatic anonymisation of data is done, including blurring of clinically relevant data. Searching is performed by context (based on the mutual information function or relative entropy) and using different filter expressions, based on DICOM metadata. Since the amount of clinical data is huge, storage and searching is performed distributed in a P2P manner. The system presented is a collaborative environment designed for sharing DICOM images and diagnoses that constitutes a distributed virtual database; accessible by the radiologists´ community enabling distributed context-based searching on DICOM images.