# CTA vessel segmentation: petrous ICA vs bone overlap problem

**URL:** <https://discourse.slicer.org/t/cta-vessel-segmentation-petrous-ica-vs-bone-overlap-problem/46803>\
**Category:** Support\
**Created:** [April 21, 2026, 8:49pm UTC](https://discourse.slicer.org/t/cta-vessel-segmentation-petrous-ica-vs-bone-overlap-problem/46803 "2026-04-21T20:49:31Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![sandcastle](https://avatars.discourse-cdn.com/v4/letter/s/59ef9b/32.png) [@sandcastle](https://discourse.slicer.org/u/sandcastle)\
**Post date:** [April 21, 2026, 8:49pm UTC](https://discourse.slicer.org/t/cta-vessel-segmentation-petrous-ica-vs-bone-overlap-problem/46803/1 "2026-04-21T20:49:31Z")

</div>

Hi all,

I’ve been working on segmenting the circle of Willis and proximal carotids from CTA data. This has been straightforward with 3DRA and TOF-MRA, but I’m running into difficulty with the petrous ICA being inseparable from adjacent bone on CTA.

The main issue seems to be that contrast-enhanced vessels and cortical bone have overlapping intensities, so:

- manual thresholding fails

- Grow from Seeds leaks into bone

- pre-processing with Swiss Skull Stripper has not helped

Does anyone have recommendations for separating vessels from bone in this region? For example:

- specific Slicer workflows or modules

- preprocessing techniques

- or alternative segmentation strategies beyond intensity-based methods

I’m relatively new to 3D Slicer, so any guidance on specific workflows or modules would be especially helpful. Thanks in advance!

---

<div class="post-metadata">

**Author:** ![pieper](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/pieper/32/8_2.png) [@pieper](https://discourse.slicer.org/u/pieper)\
**Post date:** [April 22, 2026, 1:19pm UTC](https://discourse.slicer.org/t/cta-vessel-segmentation-petrous-ica-vs-bone-overlap-problem/46803/2 "2026-04-22T13:19:57Z")

</div>

A lot depends on the quality of your CTA. I had some luck on exactly this problem using nnInteractive.

---

<div class="post-metadata">

**Author:** ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)\
**Post date:** [April 22, 2026, 7:26pm UTC](https://discourse.slicer.org/t/cta-vessel-segmentation-petrous-ica-vs-bone-overlap-problem/46803/3 "2026-04-22T19:26:03Z")

</div>

You can use the `Guided artery segmentation` and the `Guided vein segmentation` modules in the [SlicerVMTK](https://github.com/vmtk/SlicerExtension-VMTK) extension. A sample result is below.

It’s indeed not straightforward inside a bone. Segmenting the carotids at the neck level is done in a breeze (except in very very tight stenosis). Segmenting the circle of Willis is even more tricky.

Since you are a beginner, I would advise to crop your volume to the target region of study to save time. Resampling to isotropic voxel size gives better results that are more good looking in addition.

 ![IntraPetrousICAGAS](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/7/0/704dc6ea353b6e3670cd0ddbfb832921c042d634.jpeg)

 ![IntraPetrousICAGVS](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/6/d/6d007df61d0a95a7876ee9b766c46ade8b6503c7.jpeg)