Watershed Topographic Map 2.0 simulates the true boundaries of pulmonary vascular basins using artificial intelligence algorithms based on personalized lung anatomical variation patterns. Principle of watershed topographic map: Use digital simulation of the watershed boundary of the pulmonary artery and vein watershed to simulate the real boundary, block the pulmonary branch arteries - the watershed blood dries up; Blocking pulmonary branch veins - venous hypertension leads to blood stasis in the watershed, slow entry of peripheral circulating ICG into the watershed, and clear visualization of watershed boundaries within 30s to 30min after injection of fluorescent ICG into the peripheral vein. Guide doctors in preoperative planning and real-time intraoperative navigation, shorten surgical time, and reduce unexpected risks such as surgical bleeding.
This sounds like a great idea but I don’t think we have this currently. The models for extracting the pulmonary vasculature are getting very good though and it should be possible to implement this feature.
The watershed segmentation method is available in the Segment Editor module (in Watershed effect, provided by SegmentEditorExtraEffects extension). You can also use the competitive region growing method implemented in Grow from seeds effect to compute regions that are closest to each vessel branch.
That said, currently, clinicians plan sublobar resection in Slicer by segmenting the arteries and veins (AI segmentation of lobes and vessels, followed by arteries/veins separation by Grow from seeds effect) and then defining the segments by a few cuts in 3D (using Scissors effect). After this the lesion is segmented and based on all these segmentation results the clinician can make a judgment about what region to remove. The tedious, time-consuming step is the artery and vein segmentation, so I’m not sure if computation of the watershed would be particularly helpful. It seems to be more impactful to focus the efforts in automating the artery and vein segmentation.
Maybe @Eserval can provide some additional useful insights.