PaLM Conf Minutes 2021-October-13

From IHE Wiki
Jump to navigation Jump to search


Name Email
Markus Herrmann
David Clunie
James McNulty
Alessandro Sulis
Dan Rutz
Raj Dash
Megumi Kondo
Jim Harrison
Ralf Herzog
JD Nolen
Francesca Frexia
Mary Kennedy
Gunter Haroske
Riki Merrick
Gianluca Pavan Arsenal.IT
Alex Goel
David Beckman
Mandel Mickley
Craig Sayers NHS/NPIC
Hynek Kruzik

Next meeting: November 10, 2021


  • Announcements:
  • Markus Hermann discussed the pre-release of pathology collections on the National Cancer Institute’s Imaging Data Commons (IDC)
    • His presentation is attached:
    • Data commons collocate data, storage, and computing infrastructure with core services and commonly used tools and applications for managing, analyzing, and sharing data to create an interoperable resource for the research community.
    • Link to the publication mentioned in the slides:
    • Questions
      • Image data commons – is that available internationally / any cost associated with this?
        • Intended to be globally available
        • Mission is to provide data to cloud computing tools (google / AWS), which is free with any google account
        • You will have to pay for the CPU cycles used for analysis of the data
          • STRIDES program provides low-cost computing resources and cloud credits, which is also not restricted to US people (no idea if credits are available to corporations)
      • Could use the SET profile to capture the events that occur to a specimen during processing
      • Digital Pathology evidence creation profile to be developed should help with some of this
      • Some things can be sorted out in DICOM, but need to have conversation with LIS vendors to decide on required metadata
      • How do we include the gross pictures related to the sections in slides (and also the Ultrasound) – to help with AI use of the pathologist interpretation?
        • Gross images can be included in DICOM
        • Secondary capture is also an option
        • In DICOM information about the specimen makes assumption that it is on a slide
        • Have talked about it in DICOM WG26 – need to examine the workflow of how the person can make annotations to identify the areas from where the slides are taken– this still needs to have template created
        • Radiology – slides – that is what the Concordance profile is for; DICOM has tools for describing coordinates and correlations.
      • IDC slide:
        • Is based entirely on DICOM standards
        • Viewer interacts with Google API (this is a standard DICOM based API) – some also have V2 and FHIR interfaces
        • IDT architecture is intended to be vendor agnostic
      • Challenges with working with GCP and web assembly
        • GCP – DICOM web part worked very well – found a few inconsistencies in regard to the DICOM standards, but they have adjusted, when discovered
        • Web assembly is used for de-coding of compressed images; there were a few issues with old SVS files (not transforming into other color spaces that web assembly wasn’t able to detect this) – that’s when DICOM meta-data to properly decompose some
        • Jpeg and jpeg2000 work well, but older images have non-standard compression choices – have found ways to work around this
  • DPIA – Comments have been reviewed and document will be sent to Mary Jungers by Raj
  • SET Public Comments review. There were no comments. Alessandro will ask Mary Jungers to proceed with publication.
  • Update on new projects
    • Digital Pathology Aid for Telemedicine (DP-AT) – no update
    • Representation of laboratory results in patient summaries – up update
    • Digital Pathology and Structured Reporting (SR) – Per Gunter, there will be an implementation guide by the end of the year.
  • Next call November 10, 2021