ITKsnap, http://www.itksnap.org

3D Slicer, http://www.slicer.org/

ImageJ, http://imagej.nih.gov/ij/

Fiji (is ImageJ), http://fiji.sc/Fiji

TurtleSeg, http://www.turtleseg.org/

To my knowledge, such software does not exist.

Hounsfield units are not sufficient to properly segment various organs. Skin, lungs and bone can be done, but I guess that's not what you're looking for.

A CT in an operating theater is even worse; I mean, a diagnostic CT scan has much better homogeneity than a C-arc.

This being said, surgeons will be helped a lot by just providing an easy interface to adjust level/window, zoom/pan and to switch between axial/coronal/ sagittal reconstructions.

Hello Lambert, thanks.

As you said, many of the segmentation software (I use 3D Slicer and 3D Doctor) are able to reconstruct 3D images for tissues with high contrast (bone or circulatory / digestive systems using contrast agents).

But, for soft tissues, it's imposible to use an automatic procedure. Do you know anything about technique based on atlas images?

Atlas based needs a large training dataset of segmented volumes (segmented by experts/volunteers; delineation by hand).

In our clinic (radiotherapy; prostates) the success rate is still poor...

Maybe automatic segmentation on MR and subsequent registration with the 'live' C-arc CT is more feasible.

Have you ever seen a robotic intervention in prostate brachytherapy?

No, never seen it; I don't know much about brachytherapy.

Of cause those irradiation seeds are highly visible in a C-arc CT; possibly also on single X-ray images. But I have no idea how the actual location of the prostate is established.

Maybe it is assumed that the location is constant with respect to bone in the pelvic area of that particular patient?

Or ultrasound guidance?

Well, this is the problem. Prostate must be placed precisely in a 3D environment in which the robot works, taking into account the other organs. Checking needles paths can be done with ultrasound.

If you know a programmer, a custom made solution can be made by C++ or MATLAB etc. TurtleSeg is also an option, but like 3Dslicer and 3Ddoctor, very general.

Thanks, Jorgen. The first sentence of the TurtleSeg software is: "Accurate and automatic 3D medical image segmentation remains an elusive goal and manual intervention is often unavoidable." I'm afraid that is true...

I'm afraid that's true for CT. With MRI, a lot of automatic segmentation can be done because the soft-tissue contrast is much better.

The challenge with MRI, is that the pixels do not have a calibrated scale (like Hounsfield units (HU) for CT), so additional calibration for each MRI image is needed. This makes automated segmentation more complex for MRI, in my point of view. Contrast agents and/or dual X-ray scanning with different kV levels may help the image quality.

I agree Jorgen, but at least different tissues are highly visible in MRI.

There are a lot of among these, you find TurtleSeg and 3Dslicer but

I prefer work with matlab or Python

ITKsnap, http://www.itksnap.org

3D Slicer, http://www.slicer.org/

ImageJ, http://imagej.nih.gov/ij/

Fiji (is ImageJ), http://fiji.sc/Fiji

TurtleSeg, http://www.turtleseg.org/