Integrating robot-assisted ultrasound tracking and 3D needle shape prediction for real-time tracking of the needle tip in needle steering procedures.

BACKGROUND Needle insertions have been used in several minimally invasive procedures for diagnostic and therapeutic purposes. Real-time position of the needle tip is an important information in needle steering systems. METHODS This work introduces a robot-assisted ultrasound tracking (R-AUST) system integrated with a needle shape prediction method to provide 3D position of the needle tip. The tracking system is evaluated in phantom and ex vivo beef liver tissues. RESULTS An average error of 0.60 mm was found for needle insertion tests inside the phantom tissue. The R-AUST integrated with shape prediction in the beef liver tissue was able to track the needle tip with an average and maximum error of 0.37 and 0.67 mm, respectively. The average error reported in this work is within the mean allowable needle placement error (<2.7 mm) in targeted procedures. CONCLUSIONS Integration of R-AUST tracking method with needle shape prediction results in a reasonably accurate real-time tracking suitable for ultrasound-guided needle insertions.