Robotic assistance for percutaneous needle insertion in the kidney: preclinical proof on a swine animal model

Background: We evaluated the accuracy, safety, and feasibility of a computed tomography (CT)-guided robotic assistance system for percutaneous needle placement in the kidney. Methods: Fiducials surgically implanted into the kidneys of two pigs were used as targets for subsequent robotically-assisted needle insertion. Robotically-assisted needle insertions and CT acquisitions were coordinated using respiratory monitoring. An initial scan volume data set was used for needle insertion planning defining skin entry and target point. Then, needle insertion was performed according to robot positioning. The accuracy of needle placement was evaluated upon the distance between the needle tip and the predefined target on a post needle insertion scan. A delayed contrast-enhanced CT scan was acquired to assess safety. Results: Eight needle trajectories were performed with a median procedural time measured from turning on the robotic system to post needle insertion CT scan of 21 min (interquartile range 15.5−26.5 min). Blind review of needle placement accuracy was 2.3 ± 1.2 mm (mean ± standard deviation) in lateral deviation, 0.7 ± 1.7 mm in depth deviation, and 2.8 ± 1.3 mm in three-dimensional Euclidian deviation. All needles were inserted on the first attempt, which determined 100% feasibility, without needle readjustment. The angulation and length of the trajectory did not impact on the needle placement accuracy. Two minor procedure-related complications were encountered: 2 subcapsular haematomas (13 × 6 mm and 35 × 6 mm) in the same animal. Conclusions: Robotically-assisted needle insertion was shown feasible, safe and accurate in a swine kidney model. Further larger studies are needed.