Present (in BOLD):

• NASA/JPL: Dan Crichton, Sean Kelly, Heather Kincaid, Ashish Mahabal
• Arizona State University: Ji Qiu
• Boston University: Jennifer Beane
• EVMS/Old Dominion: Julius Nyalwidhe
• DMCC: Jackie Dahlgren, Chad He, Royce Malnik, Stephanie Page-Lester, Suzanna Reid
• Johns Hopkins: Zhen Zhang
• Moffitt: Matt Schabath, John Heine, Yoga Balagurunathan, Radka Stoyanova
• NCI: Sidney Fu, Guillermo Marquez, Christos Patriotis, Juan Miguel Villanueva
• PNNL: Tao Liu
• University of California: William Hsu
• University of North Carolina: Kristen Anton
• University of Washington: Savannah Partridge 

Current Action Items:

1. ONGOING: PIs are asked to review their data in LabCAS and let JPL know of any issues. Update:  Dan Crichton reminded PIs to let the know about any new data they want to send to JPL and to make sure their data is updated in LabCAS.
2. ONGOING: Discuss roadmap for additional hackathons and workshops.
3. DONE: Invite Elucidata to present at December call.  

Agenda/Discussion:

Update on DICOM Validation: Heather Kincaid provided an update.  

• GOAL: To define a minimal set of required DICOM header fields to make EDRN DICOM imaging data reusable and AI-ready, while minimizing site burden.  This will also support standardization and guide SOP development for EDRN studies.
• The EDRN studies driving this work are:
• Prostate MRI Reference Set
• LTP2 Reference Set
• Collaborating with subgroup of EDRN DICOM Researchers:
• Drafted required Core Header Tags from P-MRI and LTP2 samples in LabCAS
• Finalized tags to update post-ingest
• Implemented a validation script currently running on P-MRI and LTP2 images in LabCAS
• Drafts in process:
• Draft EDRN DICOM de-identification process
• Draft EDRN DICOM Core Header Tags

Heather Kincaid reviewed the proposed workflow.  After the site uploads their images, JPL runs a validation process. If they detect any missing tags or PHI/PII, the images are sent back to the sites to re-upload. If the upload passes validation, JPL updates the tags using the API that they are getting from the DMCC such as the blinded participant site, blinded participant ID, visit code and Event ID.  After the images are published in LabCAS, the site is asked to log in and verify the images.

Heather then reviewed the DICOM validation script: the validation script checks for all the required tags that was drafted with the DICOM subgroup plus a set of CT-like extension of CT-specific tags or MRI tags—that is what would show as validation errors on a report.  It also looks for PHI/PII using DICOM PS.15 Appendix E, which checks for burned pixels in the actual image and header. The script helps ensure the accuracy of the metadata and establishes consistency across the imaging data sets. It also helps make the images AI ready and able to support analysis. Algorithms can be plugged in to update headers so the sites won’t manually update this—they just need to order their folders under Event IDs and integrate the DMCC information.

For the existing images:

• Step 1-verify
• Run DICOM verification on all P-MRI and LTP2 images
• If PHI/PII is detected: delete and request new images from the site
• Review verification reports with the DICOM subgroup to decide per tag to:
• Remove from Draft Core
• Waive for existing images (keep for future)
• Request additional images from sites if necessary
• Step 2-Update Headers
• Apply header updates to P-MRI/LTP2
• Transfer images to approved research sites (Dr. Stoyanova, Dr. Wilson)

It was noted that if there is any PHI leakage then then the images will go back to the site, but JPL will do the updates shen they can. The difficulty that sites have scrubbing PHI was discussed. Jackie Dahlgren said that both P-MRI and LTP2 have IRB approval to have complete real imaging dates stored in our database but the institutions have institutional guidelines that don’t allow the real date to be shared. So the institutions must have date shifting and must inform the DMCC of the date shifting rule the used, which must be the standard of the institution, so they can’t use a different date shifting rule for each patient. The DMCC documents the date shifting rules.

Heather Kincaid asked if the DMCC could share scrubbing scripts—Jackie Dahlgren said that the DMCC is willing but IRB approval is needed from CalTech’s IRB for JPL to share this information. The institutions involved would also need to approve the sharing of this data.

New Image Uploads: verification is run when uploaded:

• If issues are detected—notify site with summary and detailed report
• If images pass verification:
• Run header update script
• Publish to LabCAS
• Notify site they are ready for review

Next Steps:

• Schedule DICOM Subgroup review or pMRI/LTP2 Validation Reports
• Revise draft DICOM Core + CT/MR extensions
• Rerun validation after updates
• Send site-specific reports for any needed re-uploads
• Run header updates.

Update on LabCAS Holdings: Dan Crichton said that the Quarterly Data Holdings Report has been submitted to NCI. He added that JPL continues to work to making sure that the new data follows their standards and plan to update the old data based on the priority and resources when they are available. He discussed how JPL is upgrading LabCAS metadata. They are enhancing LabCAS collections to make EDRN data more FAIR and reusable by upgrading metadata and adding standardized README files. 

• Prioritized new and reference set collections first (working on BBD and DCIS collections right now)
• Request PIs, NCI, DMCC and Data Custodians to review, edit, and complete a Metadata upgrade sheet.

William Hsu asked about data cards, and JPL follows the standard README file.  William Hsu has links regarding the other standards that he will forward to JPL.

Federated Learning Talk on December 15th: Elucidata is presenting in December.  There are several EDRN projects using federated learning and these investigators will be invited to this call.

Next Call: Monday, December 15, 2025 at 1pm Eastern/10 am Pacific