A Meshfree Method for Deformation Field Reconstruction of Soft Tissue in Needle Insertion

Abstract

Objective: The deformation field inside the soft tissue is useful to predict and track the specific target of needle insertion. Finite element (FE) provides a sensorless way to reconstruct the deformation field inside soft tissue. However, the time-consuming model meshing makes it difficult to automate the reconstruction during needle insertion operation. The purpose of this work is to present a numerical method that can automatically reconstruction of deformation field of large-deformed soft tissue during needle insertion. Methods: Reproducing kernel particle method (RKPM) was used to reconstruct the deformation and stress field of soft tissue with real-time acquired displacement and force boundary conditions. The tissue crack was simulated by employing a node split mechanism. The validation experiment involves puncturing a silicone phantom with a robotic arm integrated with a needle. Results: The reconstructed displacements approach the experimental measurements with the average error of 0.15mm, 0.30mm, 0.63mm, and 0.55mm respectively at 12mm, 24mm, 36mm, and 40mm insertion depths. The reconstructed data have respectively 88.9%, 50%, 16.7%, and 27.8% nodes with an absolute error of less than 0.3mm (2 pixels). The stress relaxation of the silicon model has been revealed and be used to qualitatively explain the reconstruction error. Von-mises stress field has been also presented and registered into the X-ray image. Conclusion: The proposed meshfree-based method has acceptable accuracy for reconstructing the deformation field inside the large-deformed organ.