Presentation Authors: Francesco Porpiglia, Enrico Checcucci, Daniele Amparore, Federico Piramide, Pietro Piazzolla, Andrea Bellin, Cristian Fiori*, Orbassano, Italy
Introduction: Recently, the indication to partial nephrectomy (PN) in increasingly complex renal masses is spreading, thanks to the introduction of robotic surgery. Intrarenal tumours represent a technical challenge because their borders are not visible on kidneyâ€™s surface. Moreover, the posteriorly located lesions requires a kidney rotation to correctly expose the tumor. We already demonstrated the importance of hyper-accuracy three-dimensional (HA3D) reconstruction of kidney and tumor anatomy based on preoperative CT images, to facilitate the surgical planning and performance. In the present study we expanded this indication and developed a dedicated system to overlap virtual 3D data on the endoscopic video to perform robotic Augmented Reality (AR) PN.
Methods: From 01/2017 to 07/2018 we prospective enrolled all patients with complex renal masses (PADUA nephrometric score > 10). Demographics and perioperative data were collected. Specifically for the study, on the basis of CT-images the HA3D virtual reconstructions were performed. The models were loaded by the rViewer application, a specifically developed software using the Unity platform to improve modelâ€™s navigation. All transformations were applied starting from kidneyâ€™s vascular pedicle to accurately reproduce movements and rotations of the real kidney during surgery. To maximize surgeonâ€™s awareness the software allowed to isolate specific parts of the model, modifying their transparency to give a flexible control of the surfaces. Finally, the video rendered by the rViewer application was fused with the one taken by the endoscopic camera using a video-mixer application, and then, the obtained images were sent back to Da Vinci console by using the Tile-pro. For posteriorly located masses we developed the 3D Elastic Augmented Reality (AR) system. The application of non-linear parametric deformations made possible to stretch and bend the organ during its rotations to expose the posterior face.
Results: 39 patients underwent AR-RAPN. 12 lesions were completely endophytic, and the overlap of 3D AR images correctly identified their location as proven by the intraoperative ultrasound. For the posteriorly located lesions the 3D elastic AR image correctly identified and simulated the tissue deformation during the kidney mobilization.
Conclusions: Our findings suggest that HA3D virtual model and real-time superimposed imaging allow to perform an effective AR-RAPN. This technology can be potentially useful in case of complex renal masses. Moreover, the evolution of the software, allows to perform a 3D elastic AR procedure, and correctly simulate the tissue deformation in case of posterior renal masses.