Estimating Needle Tip Deflection In Biological Tissue From A Single Transverse Ultrasound Image Application To Brachytherapy


This paper proposes a method to predict the deflection of a flexible needle inserted into soft tissue based on the observation of deflection at a single point along the needle shaft.


We model the needle-tissue as a discretized structure composed of several virtual, weightless, rigid links connected by virtual helical springs whose stiffness coefficient is found using a pattern search algorithm that only requires the force applied at the needle tip during insertion and the needle deflection measured at an arbitrary insertion depth. Needle tip deflections can then be predicted for different insertion depths.


Verification of the proposed method in synthetic and biological tissue shows a deflection estimation error of <2 mm for images acquired at 35 % or more of the maximum insertion depth, and decreases to 1 mm for images acquired closer to the final insertion depth. We also demonstrate the utility of the model for prostate brachytherapy, where in vivo needle deflection measurements obtained during early stages of insertion are used to predict the needle deflection further along the insertion process.


The method can predict needle deflection based on the observation of deflection at a single point. The ultrasound probe can be maintained at the same position during insertion of the needle, which avoids complications of tissue deformation caused by the motion of the ultrasound probe.