CHAPTER VI ROBOT-ASSISTED FLEXIBLE NEEDLE INSERTION
6.1 Introduction
Many modern clinical and therapeutic procedures in minimally invasive surgery require percutaneous insertion of long thin flexible needles into the deformable human tissue for a wide variety of diagnostic and interventional procedures including biopsies [120], regional anesthesia, neurosurgery [121], brachytherapy [122], drug injection and thermal ablation, among many others. In such procedures it is critical to position the needle tip precisely at the desired target. The effectiveness of needle tip placement and the success rate of the diagnosis are highly dependent on the accuracy of needle tip placement to the site of the lesion [123]. However, precise targeting is challenging because of needle deformation during insertion, registration error, tissue heterogeneity, tissue deformation, target mobility, patient movement and poor maneuverability of the needle, among other factors. While image-guided robotic needle alignment systems can improve accuracy [124], they cannot completely eliminate the above sources of error, because they cannot steer the needle to compensate for unmodeled effects.
To address some aspects of this, recent research has focused on imaging techniques, surgical planning and haptic feedback to accurately position the needle tip at the desired target. Real-time imaging techniques (e.g., CT-scan, ultrasound, fluoroscopy and MRI etc.) can increase the performance of the surgeon in navigating the tool and tracking the target [125]. Haptic based force feedback is also introduced in the surgical procedure that
can reduce the human errors due to hand tremor and problem in hand/eye coordination and which finally can contribute to reduction in targeting error [126]. Various surgical planning and simulators have been developed for specific surgical applications that can model the human anatomy accurately which is advantageous for training medical residents, predicting the surgical outcome of complex procedures, and practicing new procedures. These simulators would reduce the need for animals, cadavers and anatomical phantoms as training objects and medical residents would have unlimited opportunity to practice before performing an operation on a patient.
Much emphasis has been placed on guiding the straight rigid needles inside the deformable tissue [124]. A significant problem of inserting the stiff needle is the tissue deformation that occurs when it penetrates the tissue and thus, many times needle might miss the target [127]. Alteroviz et al. [128] proposed a way to predict rigid-needle placement error in prostate brachytherapy procedures and to correct for this error by choosing an alternative insertion point.
Recently, a number of researchers have explored an alternative approach to ensuring the success of percutaneous procedures by employing a thin bevel-tip flexible needle.
Noteworthy work in this area has been performed by Webster et al. [17]. There are numerous advantages of using such thin steerable needle. According to [129], less serious complications occur with fine biopsy needles than with a standard coarse needle. Thinner the diameter of the needle causes less damage to the tissue and reduces the chance of postdural puncture headache (PDPH) in spinal anesthesia [130]. In addition, flexible steerable needles offer the potential to “turn corners” during insertion, thereby avoiding obstacles like sensitive tissues (e.g., nerves, blood vessels etc.), reducing tip placement
error, and enabling less-invasive access to challenging anatomical locations. However, there are significant challenges of controlling a thin flexible needle inside the soft tissue.
It requires an automated computerized robotic system that can plan and perform the needle insertion safely.
In this research we have developed a robotic system that can hold a flexible bevel-tip needle stably during insertion and can position the needle tip to the target site precisely.
When the needle is inserted into the tissue, asymmetric tip forces cause it to bend in a curved path as it cuts through the tissue. The direction of bending can be controlled by axially rotating the tip. In this way, the insertion speed and the angular rotation of the needle shaft can be varied easily in the automatic robotic insertion procedure for precise needle placement. We derive a controller that will drive the tip of a bevel tipped flexible needle from its initial pose to a desired 3D point within a tissue medium. We also apply robust control techniques to compensate for uncertainty in needle curvature arising from tissue variability (the mechanical properties of the same organ in two different people can be different), tissue inhomogeneity, other unmodeled effects. The end result is the first closed-loop feedback controller of which we are aware that is able to deliver a bevel- steered needle to a desired point within tissue. Experimental contributions include validation of this controller in tissue phantoms and biological tissue (bovine liver).
The remainder of this chapter is organized as follows: Section 6.2 outlines the issues and prior research of robot guided needle insertion into soft deformable tissue. In Section 6.3, a brief review of the steerable needle and its kinematic model are presented for completeness. The controllability issues of the steerable needle are discussed in Section 6.4. Problem formulation for the steerable needle is described in Section 6.5. Section 6.6
outlines the coordinate transformations and feedback control law that positions the needle tip to a desired target. To estimate the state variables from noisy measurements a continuous-discrete extended Kalman filter is introduced in Section 6.7. Section 6.8 discusses robustness of the control system in the presence of modeling uncertainty. We describe our experimental testbed in Section 6.9. The efficacy of the proposed method is demonstrated by extensive simulations and experimental results in Section 6.10, and finally the summary of this chapter is presented in Section 6.11.