Incorrect rendering of DICOM SEG

[fedorov]

See problematic sample below. You can download the DICOM study in question using the following:

$ pip install idc-index --upgrade
$ idc download 1.2.840.113654.2.55.192012426995727721871016249335309434385

https://volview.kitware.app/?urls=[s3://idc-open-data/46350591-eaf4-4f53-8790-645d72087aa2,s3://idc-open-data/3a84b4b8-b9c1-45e5-99db-d778fd5218f8,s3://idc-open-data/4c74c61b-48b1-48f4-af20-ac92b0a8a221,s3://idc-open-data/dc93f58c-2c3c-4dce-bed3-fdb291a40cee,s3://idc-open-data/9e7dc056-400c-4ac1-b01c-8e8441486f3c]

This is how it looks in OHIF v3:

https://viewer.imaging.datacommons.cancer.gov/v3/viewer/?StudyInstanceUIDs=1.2.840.113654.2.55.192012426995727721871016249335309434385&SeriesInstanceUIDs=1.2.840.113654.2.55.305538394446738410906709753576946604022,1.2.276.0.7230010.3.1.3.313263360.15787.1706310178.804490

And Slicer:

[fedorov]

this is what I see if I open the CT series, and add SEG as a layer

[floryst]

@PaulHax looks like the SEG layer is not being properly associated with the underlying image.

[PaulHax]

Thank for the files! For the SEG file, our WASM loader computes incorrect spacing. Also, the pixel array looks like each frame along the axial has a single segment, rather than merging multple segments together for each frame. Just the tip of the issues probably.

Related:
InsightSoftwareConsortium/ITK-Wasm#991
InsightSoftwareConsortium/ITK-Wasm#1165

[jadh4v]

The ITK-Wasm PR InsightSoftwareConsortium/ITK-Wasm#1165 was recently merged. The @itk-wasm/dicom package was released (https://www.npmjs.com/package/@itk-wasm/dicom/v/7.2.0). I will test out the new reader in Volview. Hopefully its easily swappable/upgradable.

[fedorov]

@floryst when should we expect the production app in volview.kitware.app to be updated with this?

[floryst]

I'll update that deployment this week. I'm targeting EOD Wednesday.

[floryst]

@fedorov I forgot to mention that volview.kitware.app is now updated to the latest dev version!

[fedorov]

@floryst it is working now, thank you!

One observation is that SEG thumbnails give infinite spinner - maybe you should just disable rendering of thumbnails for SEG series?