Portal de Periódicos da CAPES

A DICOM Architecture for Clinicians and Researchers

2012; IOS Press; Linguagem: Inglês

10.3233/978-1-61499-101-4-539

1879-8365

Michael D. Onken,

Radiomics and Machine Learning in Medical Imaging

Over the last years there has been a strong trend of publishing health data in anonymized format in order to make it available for research. This is also true for medical imaging where the DICOM standard is the predominant data format and network protocol. This paper proposes an extension to any DICOM networking infrastructure that permits sharing of medical images in an anonymized way. Standard DICOM software is utilized on client and server side. While offering researchers access to all images in anonymous format, the architecture enables authorized clinicians to access the same images including their original patient information (name, institution, etc.). Identifying parts and anonymous parts of the image data are stored to geologically different databases. Together with sophisticated network protocols, patient privacy is fully preserved.