ICA annotation training

ICA annotation training  

  By: LeonhardRist on Sept. 8, 2023, 1:53 p.m.

Dear organizers,

I have a question regarding the ICA annotations. Normally the annotations for the other labels cover all vessel voxels included in the region of interest. For the ICA however, probably due to its curved nature, the annotation sometimes stops even though it could be traced further. Other examples for incomplete annotation occur when the ICA reappears at the ROI bottom. Examples can be seen in Patient 1 or prominently in Patient 12. While the basilaris is traced down to the bottom, the last slices for the ICA are not unannotated.

Is there any rule behind that annotation protocol?

Thank you very much in advance!

 Last edited by: LeonhardRist on Sept. 18, 2023, 12:17 p.m., edited 1 time in total.

Re: ICA annotation training  

  By: petersergeykeynes on Sept. 9, 2023, 12:15 p.m.

Dear Leonhard,

For CT angiography (CTA) modality, the ICAs are not labelled through the anterior clinoid and sphenoid (the bone parts). For CTA, the ICA is labeled starting from the C7 segment that is not surrounded by sphenoid and clinoid bone regions.

For MRA, we do label the entire curvature of the ICA in the ROI region. If you check the ROI cropped annotations, the ICAs are fully labelled in the ROI. For example, Patient 1 and Patient 12's MR annotation in the ROI:

Maybe you were referring to the CTA annotations where you see the C7 segments onwards? For MRA annotations, we label the ICA in the ROI region fully even for those in the sphenoid and clinoid bone regions.

Hope this helps!

Best, Kaiyuan