Medical imaging
AvailableDICOM De-identifier
A desktop tool that strips identifying data from DICOM studies before they leave your network. Start from a standard profile, then remove, keep or replace any tag, private tags included.
How a run works
- 01
Source study
Files or folders, read in place
- 02
Profile
Default rules plus your changes
- 03
De-identified copy
Written to a separate folder
- 04
Run log
What changed, per file and tag
What it does
Profiles
- A default profile covering the identifying attributes named in the DICOM Basic Application Level Confidentiality Profile (PS3.15 Annex E).
- Per-tag actions: remove, empty, keep, or replace with a value you choose.
- Custom tags added by group and element, including private tags.
- Profiles saved and reused, so every export from a project is treated the same way.
Mapping values
- Map a tag to a fixed value, such as a project or site code.
- Replace patient identifiers with pseudonyms that stay the same for the same patient across a run.
- Regenerate UIDs consistently, so series and studies still reference each other afterwards.
- Replace dates three ways: a completely random date, a fixed value, or a random value inside a range you set in your own mapping.
Working with studies
- Process a single file, a folder, or a whole study tree in one run.
- Source files are never modified; output goes to the folder you choose.
- A log of every change, by file and by tag, kept with the output.
At a glance
- Runs on
- A desktop workstation. Tell us your operating system and we will confirm the build.
- Reads
- DICOM Part 10 files, single or in folders
- Built around
- DICOM PS3.15 Annex E, PS3.6 data dictionary
- Data
- Processed on your machine. Nothing is uploaded.
- Access
- By request. We send a download link and its terms of use.
Check the output
De-identification is a judgement as well as a process. Review the result against your own privacy policy before any data leaves your organization, and test new profiles on non-production data first.
Coming next
What we are adding, roughly in order. Tell us in your request which of these matters to you, and it moves up.
- 01Detection and removal of text burned into pixel data, common in ultrasound and screen captures. In progress.
- 02Alignment with the IHE IT Infrastructure De-Identification Handbook: profiles chosen by use case, each data element classified, and a written record of the risk assessment.
- 03The full set of Annex E options, with the method used recorded in the output (De-identification Method Code Sequence).
- 04An embedded tag viewer with a before-and-after comparison of every file.
- 05Sample study generation for US, CT, MR and OCT, with known identifiers planted so the de-identifier can be tested against them.
- 06A scan for identifiers left behind in text values after a run.
Questions
Not in the current release, where pixel data passes through unchanged. Detecting and removing burned-in text is in progress and first under Coming next. Until it ships, check ultrasound and screen captures separately.
Each date tag can be kept, set to a fixed value, replaced with a completely random date, or replaced with a random value inside a range you define in a mapping you provide.
Only if you keep the mapping between original and replacement values, and that is your decision. Store it separately from the de-identified data, with its own access controls.
Ask us. Access terms are sent with the download link, and they depend on how you plan to use it.
Request information
Tell us what you are working on and what you would like to know. If the tool is a fit, the reply includes a download link.
