Image data sharing for biomedical research--meeting HIPAA requirements for De-identification.

Area of Science:

• Biomedical Imaging
• Data Science
• Health Informatics

Background:

• Data sharing is vital for cross-disciplinary research and scientific validity.
• Existing policies encourage data sharing, yet clinical data sharing remains limited due to privacy concerns, particularly HIPAA.
• De-identifying complex clinical data, especially medical images, presents significant challenges.

Purpose of the Study:

• To describe the development of an open-source software suite for de-identifying clinical imaging data.
• To implement DICOM Supplement 142 for enhanced privacy protection in medical images.
• To share lessons learned from acquiring over 30 million images across 20 collections.

Main Methods:

• Development of an open-source software suite integrated with the National Biomedical Imaging Archive (NBIA).
• Implementation of DICOM Supplement 142 standards for automated image de-identification.
• Acquisition and management of large-scale imaging datasets.

Main Results:

• Successful development and deployment of a software suite for DICOM image de-identification.
• Facilitation of secure clinical data sharing through adherence to privacy regulations.
• Accumulation of a substantial imaging data repository exceeding 30 million images.

Conclusions:

• The release of DICOM Supplement 142 significantly reduces barriers to clinical image data sharing.
• Open-source tools implementing these standards are essential for promoting secure biomedical data sharing.
• The NBIA platform, with its de-identification capabilities, supports large-scale biomedical research initiatives.