Toward the improvement of 3D-printed vessels' anatomical models for robotic surgery training

Multi-Detector Computed Tomography is nowadays the gold standard for the pre-operative imaging for several surgical interventions, thanks to its excellent morphological definition. As for vascular structures, only the blood flowing inside vessels can be highlighted, while vessels' wall remains mostly invisible. Image segmentation and three-dimensional-printing technology can be used to create physical replica of patient-specific anatomy, to be used for the training of novice surgeons in robotic surgery. To this aim, it is fundamental that the model correctly resembles the morphological properties of the structure of interest, especially concerning vessels on which crucial operations are performed during the intervention. To reach the goal, vessels' actual size must be restored, including information on their wall. Starting from the correlation between vessels' lumen diameter and their wall thickness, we developed a semi-automatic approach to compute the local vessels' wall, bringing the vascular structures as close as possible to their actual size. The optimized virtual models are suitable for manufacturing by means of three-dimensional-printing technology to build patient-specific phantoms for the surgical simulation of robotic abdominal interventions. The proposed approach can effectively lead to the generation of vascular models of optimized thickness wall. The feasibility of the approach is also tested on a selection of clinical cases in abdominal surgery, on which the robotic surgery is performed on the three-dimensional-printed replica before the actual intervention.