This chapter presented an architecture for closed-loop , robot assisted, MRI guided
percutaneous procedure motivated from prostate biopsy procedure. It described
various modules of such a system and their role in the complete procedure. Also,
it described various features expected from each of the module to integrate such a
system for any percutaneous intervention under MRI guidance. Presented architec-
ture is validated by two procedures, Chapter 4 describes fully autonomous needle
steering under real-time MRI guidance, while Chapter 5 presents an integrated sys-
tem for closed-loop monitoring of conformal brain tumor ablation under real-time
Chapter 4
Closed-loop Needle Steering under
MRI Guidance
4.1
Contributions
Aim of this research was to demonstrate and validate closed-loop needle steering
under real-time MRI guidance. An integrated system consisting of MRI-compatible
robot designed by my lab colleague Hao Su [65], real-time scanner control, au-
tonomous needle tracking algorithm and needle steering algorithm was used. My
primary contribution in this project was development of real-time, autonomous nee-
dle tracking coupled with MRI scanner control and interfacing of it with needle
steering robotic platform comprising of 2-DOF (needle rotation and insertion), Ap-
plication for real-time scanner control, robot control, real-time needle tracking in
MR images and needle steering algorithm were developed. Communication inter-
faces for performing closed-loop operation were developed allowing transfer of MR
images to real-time tracking application, needle steering parameters to robot control
4.2
Introduction
Needle-based percutaneous interventions such as biopsies, brachytherapy, and ab-
lation are some of the most common types of minimally invasive procedures. The
success rate of these procedures is closely related to the accuracy of the needle tip
placement. To perform needle-based procedures using MRI guidance, the surgeon
must overcome challenges including a limited workspace, MRI compatibility of in-
struments, and difficulties of acquiring real-time MR images of the region of interest
during the intervention. The use of robotic systems inside the MRI environment can
help the surgeon to perform needle-based procedures with MRI guidance [57,78–81].
Still, most of the procedures proposed so far are still open-loop due to the lack of an
effective MRI-based needle tracking method. The development of an autonomous
needle tracking system is crucial for the implementation of closed-loop needle in-
sertions. However, MR image acquisition time, difficulties to keep the needle tip
in the field of view, and quality of fast and continuous MR images for tracking the
needle are issues that make MRI-based needle steering a challenging task. Active
needle and catheter tracking techniques were proposed in [82–84], but an RF coil
has to be placed on the needle tip to be tracked. Just a few works deal with passive
tracking. In [85] needle tracking for two-dimensional (2D) MR images is presented
while in [86] a system to track the plane of an active loopless-antenna needle is
described. Although these systems were used for needle tracking, feasibility of using
these methods for closed-loop needle steering has not been evaluated. Using needle
tracking for closed-loop flexible needle steering can improve accuracy of needle-based
interventions.
Needle could be steered in a soft tissue using concentric continuum tubes, pre-
of conventional rigid needles increases the steerability. The enhanced steerability
improves accessibility to a lesion which might be obstructed by structures such as
nerves, blood vessels or bones on the insertion path. These needles deflect when
they are inserted into soft tissue due to the asymmetric interaction forces between
the tissue and the bevel tip, which could be used in controlled manner to compen-
sate for undesired needle deflection and tissue movement. Robotic systems used to
insert flexible bevel-tipped needles have been presented using video cameras [87],
fluoroscopic images [88], ultrasound images [89] and electromagnetic trackers [90]
to provide needle tip position feedback. These systems use needle insertion and
rotation around the insertion axis to steer the needle towards a target. In [47] duty-
cycled rotations were used to provide different needle curvatures, while in [91] only
the natural curvature of the needle was used for steering. A feasibility study of man-
ual steering of a flexible needle in the MR environment was presented in [92], where
the manual insertions and rotations induced significant errors. Therefore, using
an MRI-compatible robotic system to automate insertion and rotation with MRI-
based needle tracking system can result in an accurate MRI-guided bevel-tipped
needle steering.
In this chapter, a robot-assisted closed-loop flexible needle steering using real-
time MR images (in MR imaging context real-time means continuously acquired
images at about 1-2 Hz) as feedback is presented. The system consists of an MRI-
compatible robot, fast and continuous MR image acquisition, autonomous needle
tracking incorporating control of the MRI scan plane geometry, and a needle steering
algorithm to insert a flexible bevel-tipped needle towards a pre-defined target as
shown in Fig. 4.1. To the best of our knowledge this is the first time that robot-
I
Figure 4.1: Overall system setup showing: user interfaces for robot controller application and needle tracking application (inlay, in console room), MRI compatible robot controller (beside scanner), and the MRI compatible robot which resides on the scanner bed beside the phantom and holds the bevel-tipped needle.