Hi, I've opened hundreds of CBCT scans with RadiAnt viewer (version 2021.2.2, 64-bit) without issue. They are all opened from .zip files so I haven't been paying attention if they were all multiple file sets (single-frame) or single file (multi-frame) but looking back, they all seem to be multiple file sets (single-frame).

Recently, a familiar provider exported the scans differently and sent me single file scans (multi-frame). When I open them, the "Multiplanar Reconstructions" and the "3D Volume Rendering" buttons are greyed out and nonfunctional.

Using the forum search function, I can see a post on Oct 1, 2017 saying "Improved VR/3D MPR support for multi-frame series". Is RadiAnt supposed to handle these multi-frame files? What might I need to do to get these to work?

Thanks, I have loved and used your product for years.

I forgot to mention that with these multi-frame files, I can still scroll through the axial slices normally in the viewer. I just can't open the MPR or 3D VR windows.

Thanks again.

I tried the just released beta (RadiAnt DICOM Viewer 2022.1 BETA #10100) with one of the multi-frame files. I experienced the same result with nonfunctional MPR and 3DVR buttons.

I tried using the dcm4che tool emf2sf to split the multi-frame file but I got an error that the volume was the wrong media storage type. The Media Storage SOP Class UID is 1.2.840.10008.5.1.4.1.1.2 which is "CT Image Storage". The emf2sf tool is made for enhanced multi-frame files which if CT are Class: 1.2.840.10008.5.1.4.1.1.2.1 (Enhanced CT Image Storage).

I understand that multi-frame files have been around for many years before the "enhanced" classes were defined but does RadiAnt (and maybe dcm4che) expect an enhanced class to deal with a multi-frame file?

I also tried to split the multi-frame file with an old tool called dicom2 (2007) from Sebastien Barre which gave me an error "can not save a specific frame in a multiframe encapsulated/compressed stream".

There is probably some spatial information missing in the DICOM file.
RadiAnt can only display 3D reconstructions when all required data is included in the DICOM file. You can copy the DICOM tags and send us via a contact form so that we can check it.
https://www.radiantviewer.com/contact/

For anyone else coming across this thread, here is further information.

RadiAnt reviewed my DICOM tags and found that "(0018,0088) Spacing Between Slices" was missing. These multi-frame scans were from an i-Cat (Imaging Sciences International).

I've read that distance between slices (different from slice thickness) can be obtained by two methods. Explicitly From "(0018,0088) Spacing Between Slices" and implicitly by comparing "(0020,0032) Image Position (Patient)" x, y, z of each slice/frame. The Z-position difference should tell you the distance between slices.

If I understand correctly, because this is a multi-frame file with only ONE set of DICOM tags, there is only ONE place to record "(0020,0032) Image Position (Patient)" and thus only ONE set of x-y-z co-ordinates are listed. There is no way to get a differential between the slices. I don't think that the multiple frames listed in "(7FE0,0010) Pixel Data" have any position data. So, in a multi-frame file, "(0018,0088) Spacing Between Slices" is necessary.

However, I was able to open these multi-frame files with multiple (I tried 3) other programs and none of them complained or had reduced functionality like RadiAnt did. I suspect that these programs are making assumptions when the spatial information is missing. I do have "(0028,0030) Pixel Spacing" data which gives x and y spacing; maybe the programs assume a voxel ratio of 1:1:1 without overlap or empty space and thus set z = x or y. I don't know.

It would be nice if RadiAnt would ask for input if this spatial data is missing or find out exactly how other programs are confidently processing the images without the spatial data but I understand not wanting to accommodate poor standard compliance (if that is what is happening here).

Thanks for your help RadiAnt and thanks for the great product.

There are a couple other tags from which I think distance between slices could be determined.

"Slice Location (0020,1041)"
I've heard this is optional and not a common tag but if each slice is marked with a relative position, the distance between them can be found. This, again, doesn't help in a multi-frame file where this tag only has ONE instance.

"Frame Increment Pointer (0028,0009)"
If I understand correctly, this tag is made for multi-frame files and points to another tag that should have frame information. For example, you could point "Frame Increment Pointer" to "Slice Location Vector (0018,2005)" which is an array of relative slice locations.

I'm very inexperienced with DICOM tags and the vector algebra needed to interpret 3D data. But maybe this information helps others with their data sets that they may be having trouble getting RadiAnt to work on.

Yes, slice thickness does not necessarily means the distance between slices. Hence we don't want to make such assumptions. We will certainly add in future versions some way to enter missing data so that such ambiguous datasets can be rendered.

Note that the multi-frame DICOM images can also have image position defined specifically for each frame.

"Note that the multi-frame DICOM images can also have image position defined specifically for each frame."

Oh! That's interesting. Where is this data stored? In "Pixel Data"; from the "Frame Increment Pointer"? Somewhere else? I've read many pages of information on DICOM tags but haven't found this information.

Thanks

dicom.nema.org/medical/dicom/current/output/chtml/part03/sect_C.7.6.16.html