Robotic ultrasound guidance for real-time motion compensated radiation therapy (RobUST)
March 1, 2016
April 30, 2020
The project will analyse the possibility of cooperative use of two robotic systems in radiation therapy. An advantage of the clinically used CyerKnife system is the large number of degrees of freedom when selecting the directions of the beams used for treatment and its capability of compensating for respiratory motion by synchronously moving the radiation source. This allows for creating very precise and homogeneous dose distributions with steep gradients. Until now, the clinical state of the art is motion tracking with externally measured surrogate signals and infrequent X-ray or CBCT imaging. The goal of the project is the evaluation of real-time target tracking using 3D ultrasound, ultimately allowing treatment without the need for surrogate signals. Additionally, this approach will also be able to deal with non-respiratory motion (like bowel motion, changes in bladder filling, etc.). To achieve this, a 3D ultrasonic transducer will be attached to a robot arm. Radiosurgical treatment requires the delivery of multiple beams from many different directions and that visibility of the target needs to be ensured throughout the whole treatment. This directly gives rise to the core questions of the project: how can the transducer be placed safely and kinematically adequate and how can occlusion of possible beam paths by the robot and the transducer be incorporated into treatment planning. To allow high-quality imaging while occluding as few beam paths as possible, the determination of possible view ports based on pre-treatment CT or MRI data will be evaluated, as will be methods to bring the transducer to these view ports using a kinematically redundant system. This redundancy is an essential requirement to coordinate motion of the two robots, the beam delivery system and the imaging system. Additionally, image quality strongly depends on the pressure exerted by the probe, requiring adequate force-sensing capabilities. Finally, the optimization problem for inverse treatment planning will be extended to incorporate possible positions and configurations of the ultrasound robot. This means that, in addition to the beams used for treatment, the position and configuration of the second robot needs to be determined for each beam to be delivered. Based on multiple clinical scenarios, this method will then be used to evaluate the influence of transducer position, beam paths and robot configuration on treatment plan quality.