EDRN DICOM Image Headers Subgroup Call Minutes Meeting 02/13/2026
(This meeting is not online.)
Attendees
- Radka Stoyanova, University of Miami
- John Wei, University of Michigan
- Indu Kohaar, NCI
- Guillermo Marquez, NCI
- Jackie Dahlgren, DMCC
- Suzanna Reid, DMCC
- Stephanie Page-Lester, DMCC
- Chad He, DMCC
- Savannah Partridge, University of Washington SOM
- Will Hsu, UCLA
- Matt Schabath, Moffitt
- Sean Kelly, JPL
- Dan Crichton, JPL
- Heather Kincaid, JPL
Agenda
- Quick recap and updates
- Review updates to the validation logic for DICOM Tags and PHI
- Review potential validation errors and PHI in PMRI and LTP2 Images
- Review and agree on validation rules
- Next steps
Goals
- Define a standard tag set that sites can adopt now for both existing and new imaging studies, with future iterations based on feedback.
- Review and agree on validation rules that have been implemented.
Key Discussion Points and Decisions. Validation Logic and Enforcement.
- Discussion: Reviewed validation logic and the implementation of profiles and routing logic for tags to be Required, Optional, or Not Applicable (NA). Walked through main validation issues at the summary level to get feedback on enforcement rules (e.g., require sites to fix and reupload, relax for existing data but enforce for new).
Issue 1: Over-anonymization of Images (Missing SOP Class UID).
- Discussion: Some images are missing the SOP Class UID, indicating over-anonymization.
- Decision: Rerun this site assuming a standard CT profile to help the site determine requirements and identify other potential missing information.
Issue 2: Image Type (0008,0008) = "LOCALIZER".
- Discussion: Validation errors related to the "LOCALIZER" value for Image Type.
- Decision: Provide counts to the subgroup for this validation error. After review, the core requirement for MR will likely be adjusted to Allow single-token "LOCALIZER". Keep the CT requirement as is.
Issue 3. Dates and Times
- Discussion: Look at MR images that are DCE to see if time is missing. Work with Radka Stoyanova to investigate these.
- Decision (Date): Do not validate dates coming from sites. Instead, use the DMCC provided Date Offset to update date tags in headers.
- Decision (Time): Add a requirement for Time for MR DCE images. All other timestamps are to be kept as is. (LTP2): Waive for existing LTP2 tags.
- How update applied in header:
- We will use the Shifted Image Visit Date provided by the DMCC API.
- For Standard Date tags: Apply this shifted date to all standard Date tags (Study, Series, Acquisition, Content, etc.).
- For Date/Time tags: Modify only the date component; the time component (and UTC offset) must remain original.
- For Time Tags: Do not modify any Time tags.
- Private Tags: Do not apply shifting to private tags unless specifically requested.
Issue 4. Missing Required Provenance Tags
- Discussion: Missing tags include Manufacturer, Manufacturer Model Name, and Software Versions.
- Decision: Waive these tags for existing data; Require them for new data.
Issue 5. Missing Required Series Description
- Discussion: This tag is used to help determine MR DCE images.
- Decision (General): Run a report to determine affected images and totals. Work with Radka Stoyanova to send her examples of images where this is missing, then decide if MR sites need to reupload.
- Action: Ask Radka how these are being tagged for MRI.
- Decision (LTP2): Waive for existing LTP2 tags; Require for new LTP2 data.
Issue 6. Missing Window Center
- Discussion: JPL needs to do further investigation on these images.
Other Discussion Points
- Warnings on secondary capture do not necessarily need to be shown, but can be kept for now until reports are finalized.
- Send PHI/PII to sites for them to confirm.
- Small files are showing low-res images. Show some of these images to the subgroup to get feedback on whether they are from each study.
Action Items
- Update and rerun summary reports to include profile, so that investigators can see totals and percentages of missing tags/validation errors.
- Determine the percentage of PHIs that have failed.
- Send out core tags and profile routing to get agreement.
- Send totals (number of patients/eventids) to Radka Stoyanova, John Wei, and Indu Kohaar.
- Rerun LTP2 site that is missing SOP Class UID assuming a standard CT profile to help the site determine requirements and identify other potential missing information.
- Work with Radka Stoyanova to investigate MR DCE images missing time (Dates and Times).
- Send Radka examples of images where Series Description is missing.
- JPL to do further investigation on images missing Window Center.
- Send PHI/PII to sites for them to confirm.
- Show some low-res images (small files) to the subgroup to get feedback on whether they are from each study.
| DICOM Tag | DICOM Tag Name | Source | |
|---|---|---|---|
| (0010,0020) | Patient ID | DMCC API - Blinded ParticipantID | Blinded ParticipantID |
| (0020,0010) | Study ID | DMCC API - EventID | EventID |
| (0012,0050) | Clinical Trial Time Point ID | DMCC API - VisitCode | VisitCode |
| Date Tags | DMCC-API - ImagingDate? | ImageDate | |
| Date/Time tags: Modify only the date component; the time component (and UTC offset) must remain original. | DMCC-API - ImagingDate? | ImageDate | |
| For pMRI Only: | |||
| (0018,9087) | Diffusion b-value Attribute | Python Script (Radka’s group) | Run high-B value algorithm |
When
- Los Angeles
- Feb. 13, 2026, 11:47 a.m.
- Denver
- Feb. 13, 2026, 12:47 p.m.
- Chicago
- Feb. 13, 2026, 1:47 p.m.
- New York
- Feb. 13, 2026, 2:47 p.m.
- UTC
- Feb. 13, 2026, 7:47 p.m.