Loading IDC converted Visible Human DICOM images into Slicer

Hi there,

I recently stumbled across this IDC post about the original medical imaging from the Visible Human Project (VPH) being converted into DICOM files.

Following the post guidelines, I was able to successfully download the files using s5cmd, however when I try to load them into 3D slicer, it seems as though I’m only pulling in single files, opposed to the batch of files, as shown in the image attached.

Can you provide further guidance on the best way to load these files as a volume into 3D slicer? When loading in the files, I’ve ensured to deselect the “single file” checkbox to no avail.

I appreciate any insight you may have!

All the best,

Alexandria

Alexandria,

what may be happening is that you are opening an individual file with a single slice, while the entire volume you want to see consists of multiple slices/files.

You can import the entire directory into Slicer: click button pointed to by the red arrow in the screenshot below to open DICOM Browser, and then green arrow to import.

Next you can open the DICOM series from DICOM Browser, as shown below.

Alternatively, you can install SlicerIDCBrowser extension and load the entire study or individual series directly, as shown in the video in this post: SlicerIDCBrowser extension released!

Sorry for the delay in replying. Please let me know if your question is addressed!

Andrey

Andrey,

I appreciate the guidance. I was able to successfully import the entire study using the IDC extension in slicer.

Although I’m able to open the individual series for the T1 axial and T1 coronal knee, loading all other data through the dicom database (CT or MR) results in viewing of a single slice.

If you’ve got any additional advice to view these other series as volumes, I’d greatly appreciate it.

Please let me know if you think this is an appropriate request to bring to IDC office hours, in which case I’ll do so on Wednesday!

All the best,

Alexandria

···

From: Andrey Fedorov via Imaging Data Commons notifications@canceridc.discoursemail.com
Sent: Thursday, May 2, 2024 10:42 AM
To: Alexandria Podolski apodolski@ara.com
Subject: [Imaging Data Commons] [PM] Loading IDC converted DICOM images into Slicer

fedorov May 2 Alexandria, what may be happening is that you are opening an individual file with a single slice, while the entire volume you want to see consists of multiple slices/files. You can import the entire directory into Slicer: click







fedorov
May 2

Alexandria,

what may be happening is that you are opening an individual file with a single slice, while the entire volume you want to see consists of multiple slices/files.

You can import the entire directory into Slicer: click button pointed to by the red arrow in the screenshot below to open DICOM Browser, and then green arrow to import.

image

Next you can open the DICOM series from DICOM Browser, as shown below.

image

Alternatively, you can install SlicerIDCBrowser extension and load the entire study or individual series directly, as shown in the video in this post: SlicerIDCBrowser extension released!

Sorry for the delay in replying. Please let me know if your question is addressed!

Andrey


Visit Message or reply to this email to respond to IDC support email triage group (6).

To unsubscribe from these emails, click here.

@apodolski I will investigate and reply - there are some known issues with some of the VHD images, it may be that!

Would you have any objections if I move this thread into the public Support category of the forum, so that other users could benefit from this discussion as well? Since you sent your question via email, by default it is not public, and can only be seen by you and the members of the IDC team.

It’s possible, but if that were the case, I’d anticipate that the rest of the MR imaging would be easily loaded.

I appreciate your willingness to investigate! I don’t mind the thread being moved into the public Support category. Will I still receive email updates, or will I need to monitor the forum directly?

All the best,

Alexandria

···

From: Andrey Fedorov via Imaging Data Commons notifications@canceridc.discoursemail.com
Sent: Monday, May 6, 2024 3:45 PM
To: Alexandria Podolski apodolski@ara.com
Subject: [Imaging Data Commons] [PM] Loading IDC converted DICOM images into Slicer

fedorov May 6 @apodolski I will investigate and reply - there are some known issues with some of the VHD images, it may be that! Would you have any objections if I move this thread into the public Support category of the forum, so that other







fedorov
May 6

@apodolski I will investigate and reply - there are some known issues with some of the VHD images, it may be that!

Would you have any objections if I move this thread into the public Support category of the forum, so that other users could benefit from this discussion as well?


Visit Message or reply to this email to respond to IDC support email triage group (8).

To unsubscribe from these emails, click here.

1 Like

I agree! Let me investigate first!

You should be getting email updates I believe, but let me know if you don’t - we may need to look into settings!

@fedorov I’m guessing @apodolski is seeing this issue: Correct Visible Human MRI SpacingBetweenSlices (only visible to the IDC team). It looks like it will be easy to resolve in the original data, but if you need to load this sooner then a small python script could be used to patch the files.

Steve, most likely yes, but what I do not know is whether that issue covers all of the problems that exist in the dataset. We should also document here in this public issue the problems we have discovered.

@apodolski I am sorry for the delay in responding.

In short, and unfortunately, it is indeed the case that a significant portion of the CT/MR images in the nlm_visible_human_project collection is not loadable in 3D Slicer due to the issues that reconstruction of the corresponding DICOM series into 3D volumes.

Thank you for raising this issue. We did not test 3d reconstruction for this dataset prior to its release, and some issues went unnoticed.

If you are interested in specifics, most of the problematic series are MR, and cannot be reconstructed due to the SpacingBetweenSlices tag set to 0 (for those interested, and perhaps even more for my own records, at the bottom of this post you will find the SQL query that selects the series affected by this issue).

Some of the CT series (specifically, 1.3.6.1.4.1.5962.1.3.1174.2.1672334394.26545 and 1.3.6.1.4.1.5962.1.3.1174.5.1672334394.26545) have other issues (AcquisitionNumber set to varying values within the series, missing slices and possibly inconsistent orientation).

We are investigating this issue and will try to remedy it for the next IDC release v19. I will try to remember to update this thread when we have a resolution.

Thank you again for reaching out with your report!


The query below will select all of the series from the nlm_visible_human_project collection that contain more than 1 slice and have SpacingBetweenSlices set to 0.

WITH
  slices_counted AS (
  SELECT
    SeriesInstanceUID,
    COUNT(DISTINCT(SOPInstanceUID)) AS num_slices
  FROM
    `bigquery-public-data.idc_current.dicom_all`
  WHERE
    collection_id = "nlm_visible_human_project"
    AND Modality IN ("MR",
      "CT")
  GROUP BY
    SeriesInstanceUID
  HAVING
    num_slices > 1 )
SELECT
  DISTINCT(Modality),
  PatientID,
  SeriesNumber,
  SeriesDescription,
  StudyInstanceUID,
  SeriesInstanceUID
FROM
  `bigquery-public-data.idc_current.idc_v18`
WHERE
  collection_id = "nlm_visible_human_project"
  AND SAFE_CAST(SpacingBetweenSlices AS float64) = 0
  AND SeriesInstanceUID IN (
  SELECT
    SeriesInstanceUID
  FROM
    slices_counted)
ORDER BY
  SAFE_CAST(SeriesNumber AS INT) ASC

@apodolski it took us a while to get v19 out of the door, but it is live as of last week.

We fixed all of the issues that we could, and it is now possible to load most of the CT/MR series as volumes in Slicer. Please try again. Note that you will need to update SlicerIDCBrowser (or idc-index for access from Python or command line) to make sure you use IDC v19.

To access the specific CT volume in the below, you can run the following commands:

pip install --upgrade idc-index
idc download 1.3.6.1.4.1.5962.1.3.370.2.1672334394.26545

There are still some series that have missing slices. Some series will have inconsistent geometry (like the one that is shown in the above), but those issues are native to the original data collected by the project.