A new robotic system for visually controlled percutaneous interventions under CT fluoroscopy

Abstract

Minimally invasive CT-guided interventions are an attractive option for diagnostic biopsy and localized therapy delivery. This paper describes the concept of a new prototypical robotic tool, developed in a preliminary study for radiological image-guided interventions. Its very compact and light design is optimized for usage inside a CT-gantry, where a bulky robot is inappropriate, especially together with a stout patient and long stiff instruments like biopsy needles or a trocar. Additionally, a new automatic image-guided control based on “visual servoing” is presented for automatic and uncalibrated needle placement under CT-fluoroscopy. Visual servoing is well established in the field of industrial robotics, when using CCD cameras. We adapted this approach and optimized it for CT-fluoroscopy-guided interventions. It is a simple and accurate method which requires no prior calibration or registration. Therefore, no additional sensors (infrared, laser, ultrasound, etc), no stereotactic frame and no additional calibration phantom is needed. Our technique provides accurate 3D alignment of the needle with respect to an anatomic target. A first evaluation of the robot using CT fluoroscopy showed an accuracy in needle placement of ±0.4 mm (principle accuracy) and ±1.6 mm in a small pig study. These first promising results present our method as a possible alternative to other needle placement techniques requiring cumbersome and time consuming calibration procedures.