Research On The Insertion Model In The Inhomogeneous Medium And The Real-time Curve Monitoring Of A Bevel-tip Flexible Needle

As one of the main treatments of the early cancers’ diagnosis and therapy, thepercutaneous needle insertion is widely used in many therapeutic applications e.g.biopsy, ablation, brachytherapy and drug delivery. However, the rigid needlescurrently used in the clinical procedures can only implement the straight insertionpath, but cannot fulfill the curved paths which are needed to avoid obstacles in thecomplicated anatomical structures. Therefore, the flexible needle with anasymmetric tip is gaining extensive attentions from the engineering researchers andthe clinicians since it enables curved insertion paths.Recently, with the further research on the applications, some limitations of theflexible needle insertion occur, e.g. a simple “needle manipulation” always results inthe tissue slicing, and the status feedback cannot be implemented in a path-planningdependent insertion. Based on the challenges above, the following topics arefocused and discussed in this thesis, i.e. the “needle-tissue” compound manipulationmethod and the real-time curve monitoring of a flexible needle without medicalimaging systems.The “needle-tissue” compound manipulation methodology includes two phases,i.e., the target lesion will be “pushed” or “squeezed” through an open insertiontrajectory by using the “tissue manipulation”, and then an insertion operation will beperformed by using the “needle steering”. In the “needle-tissue” compoundmanipulation, the tissue turns into a typically inhomogeneous medium, so it is thefundamental to investigate the motion principle of the flexible needle within aninhomogeneous medium. The force condition is analyzed for a needle insertionprocess, in which the lateral force component of the tip is obtained as a key inputparameter of the mechanics model. In addition, the inhomogeneous tissue is dividedinto several equivalent homogeneous medium and the effect that the tissue appliesagainst the flexible needle is modeled as a series of springs just like elasticfoundations. Finally, the stiffness of the springs can be solved by usingRayleigh-Ritz method and the minimum potential energy principle, which is thesecond key parameter in the mechanics model. The deflected needle shape can becalculated and solved by substituting the two parameters into the mechanics model.The conventional medical imaging devices cannot ensure the accuracy of theflexible needle trajectory sensing and the feasibility of the imaging systemintegration. Therefore, the real-time monitoring of the needle curve without imagingdevices mentioned above will facilitate the closed-loop control in the robot assistedneedle insertion system. In this thesis, the strain gauges are proposed to be utilized to detect the strain information of the flexible needle body, via which the needlecurve can be re-constructed. Moreover, the measurement uncertainty is analyzed andthe accuracy evaluation method is also proposed. Finally, the layout of the sensorson the needle surface is configured out based on the consideration of the accuracyand the cost.The fundamental research works are validated on the developed prototype inthe last section. In the needle insertion tests in inhomogeneous medium, the two keyparameters are acquired experimentally and the validation is performed based on themodel curve and the testing curve respectively. In the needle monitoring tests, theproposed methodology is also validated based on a given fitting error boundary, inwhich the planar single-circular curve, the planar double-circular curve, the spatialcircular curve and the spatial non-circular curve are tested respectively. Theexperiments verified the effectiveness of the proposed theoretical works, and themethodology outlined in this thesis provides a new solution for the bevel tip flexibleneedle insertion operation.