A Custom Medical Image De-identification System Based on Data Privacy

doi:10.11871/jfdc.issn.2096-742X.2025.03.010

Abstract

Abstract:

【Objective】Medical imaging data has great value, but it contains a significant amount of sensitive information about patients. At present, laws and regulations regarding to the de-identification of medical imaging data are not clearly defined around the world. This study aims to develop a tool that meets compliance-driven desensitization requirements tailored to diverse research needs. 【Methods】To enhance the security of medical image data, we designed and implemented a DICOM format medical image de-identification system on the Windows operating system. 【Results】Our custom de-identification system is adaptable to the legal standards of different countries and can accommodate specific research demands. The system offers both web-based online and desktop offline de-identification capabilities, enabling customization of de-identification rules and facilitating batch processing to improve efficiency. 【Conclusions】This medical image de-identification system robustly strengthens the stewardship of sensitive medical data, aligning with data security protection requirements while facilitating the sharing and utilization of medical image data. This approach unlocks the intrinsic value inherent in such datasets.

Key words: de-identification system, medical image, data privacy, DICOM, data sharing

ZHANG Jingchen, WANG Jiayang, ZHAO Yuanzhi, ZHOU Wei, LUO Wei, QIAN Qing. A Custom Medical Image De-identification System Based on Data Privacy[J]. Frontiers of Data and Computing, 2025, 7(3): 122-135, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2025.03.010.

Figures/Tables 5

Fig.1

Fig.2

Fig.3

Fig.4

Table 1

References 27

[1]	WANG S, BONOMI L, DAI W, et al. Big data privacy in biomedical research[J]. IEEE Transactions on big Data, 2016, 6(2): 296-308
[2]	PHILLIPS M. International data-sharing norms: from the OECD to the General Data Protection Regulation (GDPR)[J]. Human Genetics, 2018, 137(8): 575-582. doi: 10.1007/s00439-018-1919-7 pmid: 30069638
[3]	KELS C G. HIPAA in the Era of Data Sharing[J]. Jama-journal Of The American Medical Association, 2020, 323(5): 476-477.
[4]	Central People’s Government of the People’s Republic of China. The Civil Code of the People’s Republic of China[EB/OL]. [2023-12-27]. https://www.gov.cn/xinwen/2020-06/01/content_5516649.htm.
[5]	National People’s Congress of the People’s Republic of China. Personal Information Protection Law of the People’s Republic of China[EB/OL]. [2023-12-27]. http://en.npc.gov.cn.cdurl.cn/2021-12/29/c_694559.htm.
[6]	ROBINSON J D. Beyond the DICOM Header: Additional Issues in Deidentification[J]. American Journal of Roentgenology, 2014, 203(6): W658-W664.
[7]	MOORE S M, MAFFITT D R, SMITH K E, et al. De-identification of Medical Images with Retention of Scientific Research Value[J]. Radiographics, 2015, 35(3): 727-735. doi: 10.1148/rg.2015140244 pmid: 25969931
[8]	SHAHID A, BAZARGANI M H, BANAHAN P, et al. A Two-Stage De-Identification Process for Privacy-Preserving Medical Image Analysis[J]. Healthcare (Basel), 2022, 10(5): 755-755.
[9]	BOFFA D J, NELSON H, SHULMAN L N. The Growth of Medical Knowledge and Data Sharing[J]. Jama-Journal of the American Medical Association, 2023, 329(15): 1314-1315.
[10]	KARPEN S R, WHITE J K, MULLIN A P, et al. Effective Data Sharing as a Conduit for Advancing Medical Product Development[J]. Therapeutic Innovation & Regulatory Science, 2021, 55(3): 591-600.
[11]	ALNAZER I, BOURDON P, URRUTY T, et al. Recent advances in medical image processing for the evaluation of chronic kidney disease[J]. Medical Image Analysis, 2021, 69: 101960.
[12]	NIESSEN W J. MR brain image analysis in dementia: From quantitative imaging biomarkers to ageing brain models and imaging genetics[J]. Medical Image Analysis, 2016, 33: 107-113. doi: S1361-8415(16)30086-X pmid: 27344937
[13]	SUINESIAPUTRA A, MCCULLOCH A D, NASH M P, et al. Cardiac image modelling: Breadth and depth in heart disease[J]. Medical Image Analysis, 2016, 33: 38-43. doi: S1361-8415(16)30105-0 pmid: 27349830
[14]	LAROBINA M. Thirty Years of the DICOM Standard[J]. Tomography, 2023, 9(5): 1829-1838. doi: 10.3390/tomography9050145 pmid: 37888737
[15]	CLUNIE D A. DICOM Format and Protocol Standardization-A Core Requirement for Digital Pathology Success[J]. Toxicologic pathology, 2021, 49(4): 738-749. doi: 10.1177/0192623320965893 pmid: 33063645
[16]	WIRTH F N, MEURERS T, JOHNS M, et al. Privacy-preserving data sharing infrastructures for medical research: systematization and comparison[J]. BMC Medical Informatics and Decision Making, 2021, 21(1): 242-242. doi: 10.1186/s12911-021-01602-x pmid: 34384406
[17]	COHEN I G, AMARASINGHAM R, SHAH A, et al. The Legal And Ethical Concerns That Arise From Using Complex Predictive Analytics In Health Care[J]. Health Affairs, 2014, 33(7): 1139-1147.
[18]	GLENN T, MONTEITH S. Privacy in the Digital World: Medical and Health Data Outside of HIPAA Protections[J]. Current Psychiatry Reports, 2014, 16(11): 494. doi: 10.1007/s11920-014-0494-4 pmid: 25218603
[19]	NEWHAUSER W, JONES T, SWERDLOFF S, et al. Anonymization of DICOM electronic medical records for radiation therapy[J]. Computers in Biology and Medicine, 2014, 53: 134-140. doi: 10.1016/j.compbiomed.2014.07.010 pmid: 25147130
[20]	ARYANTO K Y E, BROEKEMA A, LANGENHUYSEN R G, et al. A Web-Based Institutional DICOM Distribution System with the Integration of the Clinical Trial Processor (CTP)[J]. Journal Of Medical Systems, 2015, 39(5): 45. doi: 10.1007/s10916-014-0186-y pmid: 25732073
[21]	National Cancer Institute. National Biomedical Imaging Archive[EB/OL]. [2023-12-27]. https://wiki.nci.nih.gov/display/NBIA.
[22]	DICOM Library, Free online medical knowledge exchange portal-DICOM Library[EB/OL]. [2023-12-27]. https://www.dicomlibrary.com/about/.
[23]	ARYANTO K Y E, OUDKERK M, van OOIJEN P M A. Free DICOM de-identification tools in clinical research: functioning and safety of patient privacy[J]. European Radiology, 2015, 25(12): 3685-3695. doi: 10.1007/s00330-015-3794-0 pmid: 26037716
[24]	PixelMed Publishing. DICOM Cleaner[EB/OL]. [2023-12-27]. https://www.dclunie.com/pixelmed/software/webstart/DicomCleanerUsage.html.
[25]	KUNDU S, CHAKRABORTY S, CHATTERJEE S, et al. A Corrigendum on “De-Identification of Radiomics Data Retaining Longitudinal Temporal Information”[J]. Journal of medical systems, 2020, 44(9):166-166.
[26]	MUSCHELLI J. Recommendations for Processing Head CT Data[J]. Frontiers in Neuroinformatics, 2019, 13(1): 61.
[27]	GUJJARY V A, SAXENA A. A neural network approach for data masking[J]. Neurocomputing, 2011, 74(9): 1497-1501.

Content	Description	Results
The system needs to adopt light- weight WEB service models	The runtime environment framework is the Microsoft ASP.NET framework, which is a lightweight WEB service framework	PASS
The system should be capable of processing over 50,000 DICOM files within 5 seconds	In the testing environment, configured on a single server, with 1,000 files per group and a total of 10 groups tested, the average processing time for 1,000 files was 5 seconds. Based on the highest configuration of a WINDOWS SERVER 2016 cluster with 64 servers, the system could theoretically process 64,000 files in 5 seconds, exceeding the testing requirements by 28%	PASS
The system must employ a stable underlying technical architecture to ensure system stability	Set up an ASP.NET environment and utilize.NET DICOM-related components for testing. The testing is conducted with groups of 1,000 files each, totaling 20 groups, with 0 unknown errors occurring	PASS