Importing DICOM-RT from Eclipse to 3D Slicer

[Ronald Bogdasarian]

Dear Plastimatch,

My name is Ron Bogdasarian and I am a pre-medical student working on a
project that investigates the education potential of CT/MRI contouring
programs. I have used the contouring tools of the Varian Eclipse
program here at the UMass Medical School to outline anatomical
structures. I can then compare my drawings to those of others and
gain a better understanding of how the cross-sectional anatomy relates
to the 3D anatomy. While contouring might be an unrealistic method
for medical students to study anatomy, they could study pre-contoured
"pro-sections" and relate them to the 3D anatomy.

We are hoping to expand our project to see if other students can also
benefit from studying anatomy in this fashion. The Varian Eclipse
contouring tools are excellent but the program is simply too
expensive. We began looking for new programs with similar
capabilities as Eclipse about a month ago and found 3D Slicer and
Plastimatch. We hope they will allow us to import my work (DICOM-RT)
into 3D Slicer. Ideally I would then be able to manipulate or edit
the images using the 3D Slicer program. This would allow me to send
my work from Eclipse to other students and allow them to study and
contour using the free 3D Slicer program.

Do you know if 3D Slicer and Plastimatch are capable of importing
DICOM RT from Eclipse? If so, can a user edit the images using the
contouring tools in 3D Slicer?

Thanks for your help,
Ron Bogdasarian

[Gregory Sharp]

Hi Ron,

Sounds like a very interesting project. The short answer is "yes",
and "yes."

Caveat #1. In 3D Slicer, anatomic structures are described
using labelmaps. In a labelmap, an image voxel can only belong to one
structure. This is in conflict with DICOM RT, where a voxel could be
contained within many structures. You can workaround, such as
creating multiple 3D Slicer labelmaps, but I guess you should
consider this limitation before you proceed.

Caveat #2. When importing DICOM RT into 3D slicer, your
structure colors and names will be lost. If you need to
preserve these, for example re-typing them is infeasible, I can
help you work around this problem.

Greg
--
Greg Sharp, PhD
Department of Radiation Oncology
Massachusetts General Hospital

The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.

[Marta Peroni]

Hi Ron!
This sound like a very interesting project to me too and honestly
speaking most of the (poor) anatomy knowledge that I have came exactly
from working every day with medical data and looking both at
cross-sectional and 3D rendering.
Anyway, Plastimatch can definitely deal with Varian Aria/Eclipse
DicomRT/DicomDose. They can be imported and converted to any other
format supported from Plastimatch (e.g. labelmaps). In brief, you can
generate either one volume only with all the structures or one image
volume per structure.
Please take into account the remark Dr Sharp has made in his previous
email, about multiple labels per voxel.
I do also have an additional question for you. How do you plan to
compare the different drawings?
thanks
Marta

--
*******************************************************************************************
Marta Peroni
PhD Student - Bioengineering department - Politecnico di Milano (IT)

contacts:
mail: marta....@mail.polimi.it , mpe...@partners.org , mpe...@csail.mit.edu
mobile: +393488202136 (ITA) +16178186365 (USA)
office: +39 02 2399 9022 (ITA) +16177266329 (USA)

*******************************************************************************************

[Ronald Bogdasarian]

Thanks everyone for your quick replies,

I'm glad to hear that 3D Slicer has the capabilities that I'm looking
for and I look forward to solving the issues with converting. The
fact that labelmaps do not allow for multiple structures in a single
voxel (I interpret that as I cannot overlap with my drawings) will be
a problem, but not a huge one. I do have overlaps in my drawings
(e.g. where tendon connects to bone, bone on bone contact), but these
areas are somewhat subtle and are not the focus of my project.

Greg, Caveat #2 might be more of an issue. I have a head/neck case, a
male thorax/abdomen/pelvis case, and a female pelvis case. I would
guess that each of these 3 cases contains an average of 200
structures. Therefore, I would rather not have to retype names and
redo colors. Is it difficult to retain names and colors? I know that
Varian Eclipse has a wealth of colors and I would not expect 3D slicer
to have the same amount.

Marta, thanks for your interest in my project. Viewing CT scans and
3D rendering is a great way to learn anatomy! I'm sure your knowledge
is great.
I will certainly look into the two options that you wrote about - 1
volume or multiple volumes. I think that it might be best for me to
do multiple volumes, because it is important for students to be able
to "turn on and off" individual slices.
To answer your question about comparing drawings: In Varian Eclipse I
am able to "fuse" two identical CT scans together. Eclipse allows the
user to then scroll between the two CT image sets. One of these sets
is the "pro-section" with my drawings. The other CT image set is
blank, and can be drawn on by the student. I don't know if 3D Slicer
has this feature, but if not a student will hopefully be able to open
the 3D Slicer program twice and view a "pro-section" CT images set and
a blank CT image set side by side.

If you have any other questions feel free to ask. Now, I suppose it
is time to ask how I go about converting Eclipse DICOM-RT to 3D Slicer
labelmaps.

Thanks again and have a good weekend,
Ron

On Sep 8, 4:49 pm, Marta Peroni <m.per...@gmail.com> wrote:
> Hi Ron!
> This sound like a very interesting project to me too and honestly
> speaking most of the (poor) anatomy knowledge that I have came exactly
> from working every day with medical data and looking both at
> cross-sectional and 3D rendering.
> Anyway, Plastimatch can definitely deal with Varian Aria/Eclipse
> DicomRT/DicomDose. They can be imported and converted to any other
> format supported from Plastimatch (e.g. labelmaps). In brief, you can
> generate either one volume only with all the structures or one image
> volume per structure.
> Please take into account the remark Dr Sharp has made in his previous
> email, about multiple labels per voxel.
> I do also have an additional question for you. How do you plan to
> compare the different drawings?
> thanks
> Marta
>
> On 09/08/2010 10:30 PM, Gregory Sharp wrote:
>
>
>
>
>
> > On Wed, 8 Sep 2010 11:35:51 -0700 (PDT)

> > Ronald Bogdasarian<rbogdasar...@gmail.com>  wrote:

> ***************************************************************************­****************

> Marta Peroni
> PhD Student - Bioengineering department - Politecnico di Milano (IT)
>
> contacts:

> mail: marta.per...@mail.polimi.it , mper...@partners.org , mper...@csail.mit.edu

> mobile: +393488202136 (ITA) +16178186365 (USA)
> office: +39 02 2399 9022 (ITA) +16177266329 (USA)
>

> ***************************************************************************­****************- Hide quoted text -
>
> - Show quoted text -

[Greg Sharp]

Hi Ron,

How comfortable are you in compiling the program and running the
command line?

I just added the capability to export to 3D slicer color table,
which will preserve the colors and structure names. You will
need to update to the latest source code from svn.

Assuming the rtss and ct volume are in the same directory,
you may try the following command:

plastimatch convert --input your_dicom_directory --output-labelmap
labelmap.nrrd --output-colormap colormap.ctbl

-Greg

[D'angelo]

Hi Marta,

I am currently encountering the same problem and will really appreciate if you could be of any help. I am actually trying to calculate the hausdorff distance between the tumor contour and an isodose line using the slicerRT software .However while it has worked for some patients it is not working for the majority and i cant figure out why. Can you please please give me the great outlines for importing data from eclipse to 3D slicer. What should be done to ensure a proper working of the system

Thanks,

Ange 

[Marta Peroni]

Hi Ange,

I have too few information to actually being able to properly help you, but if I were you I would ask myself what are the peculiarities of the patients that work? are there differences with the ones that don't?

Also take into account the following caveats from Dr Sharp:

"
Caveat #1.  In 3D Slicer, anatomic structures are described
using labelmaps.  In a labelmap, an image voxel can only belong to one
structure.  This is in conflict with DICOM RT, where a voxel could be
contained within many structures.  You can workaround, such as
creating multiple 3D Slicer labelmaps, but I guess you should
consider this limitation before you proceed.

Caveat #2.  When importing DICOM RT into 3D slicer, your
structure colors and names will be lost.  If you need to
preserve these, for example re-typing them is infeasible, I can
help you work around this problem.
"

Maybe the dataset that do not work have multiple labels per voxel?

Hope this helps you a little bit...

Thanks

Regards
Marta

--
You received this message because you are subscribed to the Google Groups "Plastimatch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plastimatch...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
*******************************************************************
Marta Peroni, PhD

contacts:
mobile: +393488202136 (IT)

********************************************************************

[Sharp, Gregory C.]

Hi Marta,

 

Good to hear from you.  SlicerRT properly maintains the names and colors

of structures.  You should upgrade your version.  :)

 

Dear Ange,

 

What do you mean “not working”?  You might look in the Error Log.

Also, please tell me which version of Slicer/SlicerRT.

 

Greg

The information in this e-mail is intended only for the person to whom it is

[Elizabeth Watt]

Hi everyone,

I also have a question along a similar thread to those previously posted.  I am using 3D Slicer with the RT extension to import structure sets from Eclipse and generate STL files, such that these structures can be imported into SolidWorks and ultimately 3D printed/milled.  I've been noticing, however, that the volumes of the structures changes when I import them from Eclipse, and the factor by which the values change is inconsistent between structures within the same patient and between patients.  Any advice on how to combat this issue?

Thanks very much,

Elizabeth

[Sharp, Gregory C.]

Hi Elizabeth,

 

It is my impression that the conversion from DICOM-RT to STL is a “lossy” operation.

Do you think this could be the cause?  A screenshot or two would help.

                                                                  

I also cc to Slicer users email list who might know.

 

Greg

 

From: Elizabeth Watt [mailto:elizabet...@gmail.com]
Sent: Tuesday, September 02, 2014 5:04 PM
To: plast...@googlegroups.com
Cc: Sharp, Gregory C.
Subject: Re: [Plastimatch] Importing DICOM-RT from Eclipse to 3D Slicer

 

Hi everyone,

 

I also have a question along a similar thread to those previously posted.  I am using 3D Slicer with the RT extension to import structure sets from Eclipse and generate STL files, such that these structures can be imported into SolidWorks and ultimately 3D printed/milled.  I've been noticing, however, that the volumes of the structures changes when I import them from Eclipse, and the factor by which the values change is inconsistent between structures within the same patient and between patients.  Any advice on how to combat this issue?

 

Thanks very much,

Elizabeth

[Elizabeth Watt]

Thank you all for your quick replies!  I have noticed that it is definitely considerably worse for small structures, but fine at larger structures.  I look forward to the full evaluation report capability!

Elizabeth

On Tue, Sep 2, 2014 at 5:08 PM, Csaba Pinter <csaba....@queensu.ca> wrote:

Hi Elizabeth,

 

Indeed, this process is lossy, but even worse, it is different in every TPS and research tool. We are trying to address this issue, see [1] for further info and tracking its progress.

We are also improving the SlicerRT importer, to avoid as much data loss as possible, see [2].

 

In a nutshell, I think a few percent mismatch should be OK for most structures (you might get considerable differences in small ones such as optic organs), as TPSs have got FDA with significantly different importers and rasterizers, but also you should see a big improvement in the SlicerRT importer soon, hopefully decreasing the mismatch, and a full evaluation report in the medium term too (MSc project about it starts about now).

 

csaba

 

[1] https://www.assembla.com/spaces/slicerrt/tickets/521

[2] https://www.assembla.com/spaces/slicerrt/tickets/103

[silvia...@gmail.com]

Hi everyone,

I am a Medical Physicist in training, I'm tryng to convert a DICOM RT structure into a .stl file. Apparently 3D slicer could complete this operation, but it doesn't. After some efforts I could create a label map from my contour, but after it was impossible to create the 3D model.

I have some error messages, this one for example:

ERROR: In ..\..\..\VTKv6\Common\ExecutionModel\vtkDemandDrivenPipeline.cxx, line 710

vtkCompositeDataPipeline (0000000024487880): Input port 0 of algorithm vtkImageMapToColors(00000000117970F0) has 0 connections but is not optional.

Do you have any suggestion?

Thank you

     Silvia

[Sharp, Gregory C.]

Hi Silvia

 

Welcome to medical physics.  :)

 

This issue is better discussed on the slicer user list.

I have not seen this error, but maybe others have.

 

Meanwhile, I would ask you to try creating a simple test labelmap

using the editor, and then generating the STL from that.

Do you get an error in that case?

 

Also, I ask you to tell us which Slicer version you are using.

I recommend to use a recent nightly build.

 

Greg

--

You received this message because you are subscribed to the Google Groups "Plastimatch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plastimatch...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

The information in this e-mail is intended only for the person to whom it is

[Silvia Calusi]

Thank you Gregory,

it does work with labelmaps created from a CT study.

I am using the 4.4.0 stable version of slicer.

     Silvia 

[Sharp, Gregory C.]

Hi Csaba,

 

Just to be clear, one cannot do the following workflow

 

Contour -> Labelmap -> Model -> STL

 

?

 

Greg

 

From: Csaba Pinter [mailto:csaba....@queensu.ca]
Sent: Friday, January 30, 2015 9:56 AM
To: Sharp, Gregory C.; slicer...@bwh.harvard.edu
Cc: plast...@googlegroups.com
Subject: RE: [slicer-users] [Plastimatch] Re: Importing DICOM-RT from Eclipse to 3D Slicer

 

Hi Silvia,

 

Labelmaps cannot be saved to STL, models on the other hand can (STL is a vertex-based format, not voxel-based).

You'll need to convert your contours to contain closed surface representation, see

http://www.slicer.org/slicerWiki/index.php/Documentation/4.4/Modules/Contours

 

Then (this part is a bit sketchy, because we don't have GUI for this operation yet) you'll need to extract the models from the contours, using the python interactor, like this (please change the node names to match yours):

 

contour = getNode('Bladder_Contour')

surface = contour.GetClosedSurfacePolyData()

surfaceCopy = vtk.vtkPolyData()

surfaceCopy.DeepCopy(surface)

surfaceCopy.GetNumberOfPoints()

modelNode = slicer.vtkMRMLModelNode()

modelNode.SetName(contour.GetStructureName())

modelNode = slicer.mrmlScene.AddNode(modelNode)

polyDataProducer = vtk.vtkTrivialProducer()

polyDataProducer.SetOutput(surfaceCopy)

modelNode.SetPolyDataConnection(polyDataProducer.GetOutputPort())

displayNode = slicer.vtkMRMLModelDisplayNode()

slicer.mrmlScene.AddNode(displayNode)

displayNode.VisibilityOn()

modelNode.SetAndObserveDisplayNodeID(displayNode.GetID())

 

This script uses a contour node with a specific name. If you want to do this for all contours, then just iterate over the vtkMRMLContourNode nodes in the scene, like this:

 

allContours = slicer.util.getNodes('vtkMRMLContourNode*')

for contour in allContours.values():

  surface = contour.GetClosedSurfacePolyData()

  ...

 

I wanted to defer implementing this feature until the new implementation is ready (we're working on a completely new version for contour handling), but this is requested so often nowadays that I should probably just do it.

 

Let me know how it goes!

 

csaba

[Silvia Calusi]

Thank you Csaba,

I tried creating the labelmap from the contour and then I used the pyton interactor and a model was created. When I open the "save window" and I choose the directory and the file extension apprently everything works correclty, but then I cannot find tha model file in the directory.....

     Silvia

2015-01-30 16:07 GMT+01:00 Csaba Pinter <csaba....@queensu.ca>:

It should be possible, yes, but then I am not the right person to help, as I'm not familiar with the labelmap->model converter internals, only the Contour internals.

 

csaba

_______________________________________________
slicer-users mailing list
slicer...@bwh.harvard.edu
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to slicer-use...@bwh.harvard.edu with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ