부산대학교 도서관

The interaction between an asymmetric (bevel) tipped needle and the surrounding tissue in image-guided percutaneous interventions, such as in prostate needle biopsy, can result in the deflection of the needle tip and cause significant targeting error. Researchers have extensively investigated different aspects of the needle-tissue interaction and proposed various solutions to mitigate this issue. While there are several promising approaches, most require complex software or hardware which makes them difficult to be deployed for clinical use. In this paper, we present a predictive model-based approach to passively compensate for undesired deflection of the needle tip prior to the initial insertion into the tissue. In this approach, an approximation of the initial deflection angle of the needle tip and the natural curvature of the needle path are utilized to simulate the insertion and predict the tip deflection prior to insertion. The model then calculates a modified needle entry point which accounts for the predicted tip deflection during the insertion. To create the model, we first collected a set of needle insertion data utilizing nonhomogeneous gelatin phantoms and ex vivobovine tissue, using an electromagnetic (EM) tracking system and a needle with an EM sensor embedded at its tip. The collected data were then used to find the model parameters, namely, the initial deflection angle and the needle path curvature. After creating the model, a validation study consisting of two sets of insertions, one with and another without compensation, was carried out to evaluate the performance of our proposed model in compensating for the tip deflection error. The results demonstrate an average of 77% targeting accuracy improvement for insertions with modified entry points based on our model’s prediction compared with the uncompensated insertions.