Freehand ultrasound guided robotic needle steering

As postdoc fellow at Stanford University I am working within a National Health Institute (NIH) funded project for the development of a clinically transferable platform for percutaneous ablations of liver tumors.

1. Illustration of the needle steering potential. Multi-focal tumors could be reached using less and conveniently located needle access points . Also, shaped ablations of large and irregular tumors could be addressed just by steering the needle in multiple directions; without the need for multiple needle insertions as it happens in the current practice.

1. Illustration of the needle steering potential. Multi-focal tumors could be reached using less and conveniently located needle access points . Also, shaped ablations of large and irregular tumors could be addressed just by steering the needle in multiple directions; without the need for multiple needle insertions as it happens in the current practice.

 
2. Robot and needle tip design. The robot allows managing the three degrees of freedom of the system: needle base insertion, needle base rotation (or needle spinning) and needle tip joint (bending of the flexure joint). The inset shows pictures and scheme of the flexure joint design for the needle tip.

2. Robot and needle tip design. The robot allows managing the three degrees of freedom of the system: needle base insertion, needle base rotation (or needle spinning) and needle tip joint (bending of the flexure joint). The inset shows pictures and scheme of the flexure joint design for the needle tip.

 

Percutaneous tumor ablation is one of the least invasive methods for the treatment of liver cancer, but its application is limited by the linear path of straight ablation probes (needles) and the poor targeting accuracy of manual insertion. Robotic steerable needles can be inserted along controlled curved paths. This enables accessing challenging tumor locations and can potentially improve tumor targeting accuracy thanks to the ability to course-correct the needle during insertion. However, barriers to clinical adoption of steerable needles remain. These include low steering perfomances in biological tissue, lack of integrated clinical therapy, and the need for a practical method to localize curved needles using freehand 2D ultrasound imaging.

To improve the steering performances in biological tissue (liver tissue), we use the so called “exaggerated asymmetric tip” needle steering principle. In particular, our needle design features a flexure joint which allows having full control of the tip asymmetry and therefore of the curvature in biological tissue upon needle insertion. This allows increasing the steering performance without compromising its safety and control reliability, since the tip articulation is released during any needle base rotation from the base, or needle spinning. The needle spinning is necessary to steer the needle in 3D. The tip design also enables therapy delivery at the target site. In this work the needle shaft is insulated and used for conveying radio frequency current to the tip. In a real scenario, the needle tip will be firstly placed at the tumor location, and then the radio frequency ablation will be initiated in order to destroy (by heat) the surrounding cancer tissue.

Validation tests on live animal (swine) performed at the Stanford Hospital-Radiology Department.    (a)    Freehand ultrasound guided human-in-the-loop control of the robotic steerable needle.    (b)    Validation through cone beam CT (Artis Zeego by Siemens) for 3D imaging and ground truth evaluation of the targeting error. Fake tumors are introduced and used as physical targets during pre-clinical swine tests.

Validation tests on live animal (swine) performed at the Stanford Hospital-Radiology Department. (a) Freehand ultrasound guided human-in-the-loop control of the robotic steerable needle. (b) Validation through cone beam CT (Artis Zeego by Siemens) for 3D imaging and ground truth evaluation of the targeting error. Fake tumors are introduced and used as physical targets during pre-clinical swine tests.

The control of the robotic steerable needle in liver tissue is based on feedback from freehand 2D ultrasound imaging. Ultrasound is the most ubiquitous and safest (no harmful radiations) imaging system for percutaneus interventions. Technology challenges are related to the real time and automatic needle segmentation and the 2D nature of ultrasound imaging (the needle moves on trajectories in 3D). We implemented a human-in-the-loop control strategy and a customized estimation algorithm that allows using a coarse needle localization from expert ultrasound users (interventional radiologists).

I am currently performing in-vivo liver experimentation to assess the targeting accuracy of the needle steering platform control.

This work has not been published yet, but a brief overview is presented in [2].