Related Experiment Video
Updated: Jul 18, 2026
06:18
Pioneering Patient-Specific Approaches for Precision Surgery Using Imaging and Virtual Reality
Published on: April 5, 2024
Pseudonymization of radiology data for research purposes
Rita Noumeir1, Alain Lemay, Jean-Marc Lina
1Ecole de Technologie Supérieure, 1100 Notre-Dame West, Montreal, QC, Canada H3C 1K3. noumeir@ele.etsmtl.ca
Journal of Digital Imaging
|December 28, 2006
Summary
Researchers need validated clinical data for medical AI. This study proposes a secure, pseudonymous imaging database architecture to enable longitudinal data collection for research while protecting patient privacy.
Area of Science:
- Medical image analysis
- Health informatics
- Data privacy
Background:
- Medical image processing algorithms require validation using real-world clinical data.
- Electronic health records (EHR) offer valuable, large-scale data but raise privacy concerns.
- Longitudinal patient data, including disease progression, is crucial for research.
Purpose of the Study:
- To develop a secure, incrementable research database for medical imaging data.
- To enable the use of pseudonymous clinical data for research while ensuring patient confidentiality.
- To explore and evaluate software architectures for longitudinal data management.
Main Methods:
- Exploration of various software architectures for imaging research databases.
- Evaluation of security measures and potential pitfalls.
- Proposal of a de-identification scheme aligned with Digital Imaging and Communications in Medicine (DICOM) standards.
- Implementation of pseudonymization for longitudinal data tracking.
Main Results:
- Identified and evaluated different software architectures for building incrementable imaging research databases.
- Assessed the security implications of proposed architectures.
- Developed a DICOM-compliant de-identification strategy for pseudonymization.
- Demonstrated a method for securely updating research databases with new patient data over time.
Conclusions:
- A robust software architecture and a DICOM-aligned de-identification scheme are essential for creating secure, incrementable imaging research databases.
- Pseudonymization effectively balances the need for longitudinal data with patient privacy.
- This approach facilitates secondary use of EHR data for research and public health surveillance.
