Patient Specific Systems for Computer Assisted Robotic Surgery Simulation, Planning, and Navigation

The evolving scenario of surgery: starting from modern surgery, to the birth of medical imaging and the introduction of minimally invasive techniques, has seen in these last years the advent of surgical robotics. These systems, making possible to get through the difficulties of endoscopic surgery, allow an improved surgical performance and a better quality of the intervention. Information technology contributed to this evolution since the beginning of the digital revolution: providing innovative medical imaging devices and computer assisted surgical systems. Afterwards, the progresses in computer graphics brought innovative visualization modalities for medical datasets, and later the birth virtual reality has paved the way for virtual surgery. Although many surgical simulators already exist, there are no patient specific solutions. This thesis presents the development of patient specific software systems for preoperative planning, simulation and intraoperative assistance, designed for robotic surgery: in particular for bimanual robots that are becoming the future of single port interventions. The first software application is a virtual reality simulator for this kind of surgical robots. The system has been designed to validate the initial port placement and the operative workspace for the potential application of this surgical device. Given a bimanual robot with its own geometry and kinematics, and a patient specific 3D virtual anatomy, the surgical simulator allows the surgeon to choose the optimal positioning of the robot and the access port in the abdominal wall. Additionally, it makes possible to evaluate in a virtual environment if a dexterous movability of the robot is achievable, avoiding unwanted collisions with the surrounding anatomy to prevent potential damages in the real surgical procedure. Even if the software has been designed for a specific bimanual surgical robot, it supports any open kinematic chain structure: as far as it can be described in our custom format. The robot capabilities to accomplish specific tasks can be virtually tested using the deformable models: interacting directly with the target virtual organs, trying to avoid unwanted collisions with the surrounding anatomy not involved in the intervention. Moreover, the surgical simulator has been enhanced with algorithms and data structures to integrate biomechanical parameters into virtual deformable models (based on mass-spring-damper network) of target solid organs, in order to properly reproduce the physical behaviour of the patient anatomy during the interactions. The main biomechanical parameters (Young's modulus and density) have been integrated, allowing the automatic tuning of some model network elements, such as: the node mass and the spring stiffness. The spring damping coefficient has been modeled using the Rayleigh approach. Furthermore, the developed method automatically detect the external layer, allowing the usage of both the surface and internal Young's moduli, in order to model the main parts of dense organs: the stroma and the parenchyma. Finally the model can be manually tuned to represent lesion with specific biomechanical properties. Additionally, some software modules of the simulator have been properly extended to be integrated in a patient specific computer guidance system for intraoperative navigation and assistance in robotic single port interventions. This application provides guidance functionalities working in three different modalities: passive as a surgical navigator, assistive as a guide for the single port placement and active as a tutor preventing unwanted collision during the intervention. The simulation system has beed tested by five surgeons: simulating the robot access port placemen, and evaluating the robot movability and workspace inside the patient abdomen. The tested functionalities, rated by expert surgeons, have shown good quality and performance of the simulation. Moreover, the integration of biomechanical parameters into deformable models has beed tested with various material samples. The results have shown a good visual realism ensuring the performance required by an interactive simulation. Finally, the intraoperative navigator has been tested performing a cholecystectomy on a synthetic patient mannequin, in order to evaluate: the intraoperative navigation accuracy, the network communications latency and the overall usability of the system. The tests performed demonstrated the effectiveness and the usability of the software systems developed: encouraging the introduction of the proposed solution in the clinical practice, and the implementation of further improvements. Surgical robotics will be enhanced by an advanced integration of medical images into software systems: allowing the detailed planning of surgical interventions by means of virtual surgery simulation based on patient specific biomechanical parameters. Furthermore, the advanced functionalities offered by these systems, enable surgical robots to improve the intraoperative surgical assistance: benefitting of the knowledge of the virtual patient anatomy