Currently submitted to: JMIR Medical Informatics
Date Submitted: Nov 5, 2026
Open Peer Review Period: Nov 16, 2026 - Jan 11, 2027
(currently open for review)
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Pseudonymization and Standardization of Heterogeneous Multicenter Ophthalmic Images in a Secure Environment: Implementation Report of the FREDD Image Pipeline
ABSTRACT
Background:
FREDD (French Rare Eye Disease Database) is a national health data warehouse dedicated to rare eye diseases, currently in its pilot phase. It collects images (fundus, autofluorescence, optical coherence tomography [OCT]) from various participating hospitals and clinical centers. Due to the heterogeneous nature of the exports (nested archives, duplicated transfers of DICOM/E2E files) and the potential presence of identifying data, all images must be processed in a secure environment.
Objective:
To describe the implementation of a pipeline for the inventory, classification, deduplication, pseudonymization, and standardization of images, with reconciliation against the information collected in FREDD (the electronic case report form, eCRF).
Methods:
The pipeline consists of five steps: inventory, content analysis without decoding pixels, sorting with a single status per source file, pixel preprocessing involving the creation of a new DICOM file with whitelisted metadata, and reportings. Independent safeguards, including a pixel text detector (without character recognition) and quarantine, protect the pseudonymization process. Execution runs in batches and can resume. Five indicators were tracked: sorting completeness, redundancy, quarantine rate, residual text at visual check, and agreement with the eCRF.
Results:
The final run inventoried 34,158 files in 41.1 hours. Of these, 18,515 (54.2%) were exact duplicates, as some deliveries had been inadvertently placed in several archives. It produced 25,767 pseudonymized DICOM files (15,784 OCT, 5,609 autofluorescence, 4,374 color fundus); 2 images were quarantined. Across the four centers with images, 23,372 of 27,860 comparisons of declared and found images (83.9%) agreed with the eCRF. A visual inspection of 354 images, stratified by device and modality, revealed no embedded text (upper bound of the residual rate, 4% per stratum and 0.8% pooled).
Conclusions:
. Heterogeneous multicenter ophthalmic exports can be standardized and pseudonymized in a secure environment, with a traced status for each file. Hash-based deduplication and eCRF reconciliation also serve as data quality tools. The visual check mitigates the residual risk of identifying text without fully eliminating it, and the usefulness of the images for downstream research remains to be evaluated.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.