Most of the 240,000 men in the US diagnosed with prostate cancer annually [1] have low-risk, organ-confined disease eligible for targeted therapy, which spares patients from undesired side effects [2]. Focal Laser Ablation (FLA) rapidly heats targeted cancerous tissue using a laser catheter. Despite its minimally invasive nature, FLA has several technical limitations, mostly related to the difficulty of (1) locating the tumor accurately during surgery; (2) placing the catheter safely to fully ablate the tumor, which can require several placements for large tumors; and (3) monitoring the ablation zone.

In this paper, a robotic system was designed to improve MRI-guided prostate FLA. We hypothesized that accurate placements of laser catheters could be achieved under MRI guidance and robotic positioning, avoiding unnecessary gland punctures and ensuring the ablation covers the entire tumor, maximizing the utility of a minimally invasive system. The plan could be updated iteratively after each ablation using preoperative parametric MRI information combined with intraoperative MR thermometry [3].

Robotic Positioner

In an MRI setting, any robotic system must be made of plastics and small amounts of non-ferromagnetic metals for patient safety and image quality. Space in the cylindrical bore (55-70cm) of an MRI scanner is limited, and the small size of the prostate requires that the robot must occupy a square targeting region of approximately 50mm in the transverse plane; additional space between the needle guide and the bore opening is necessary to permit operation.

The design solution chosen is a slim robot design built around a Core-XY belt system (Fig. 1a). Targeting is controlled by two pneumatic motors attached to a single belt. A pair of fiber optic lines is integrated with each motor, generating quadrature pulses for positional encoding with a resolution of 0.005o.

The robot registers on the MRI coordinate system using five gadolinium fiducial markers (PinPoint, Beekley Inc., USA). This robot uses Visualase® laser ablation catheter kits.

Remote Insertion

A remote insertion guide (Fig. 1b) allows the user standing at the bore opening to insert the catheter under real-time MRI guidance. The remote insertion guide is aligned manually with the robot, and the catheter moves through the guide to the robot. The alignment cone (Fig. 1c) funnels the inserted catheter into the end-effector needle guide.

Software & Control

A LabVIEW™ (National Instruments, Austin, TX) interface displays the targeting workspace, end effector position, and controlled target position. Operators drive the end effector either manually or automatically. A position-seeking algorithm targets the tumor. The required motor rotations are calculated based on the desired change in position. Optical encoding tracks the current position. PID modulated PWM compressed air supply controls the motors.

Accuracy & Efficacy Testing

The system was tested for MRI compatibility, insertion accuracy, and workflow streamlining in a clinical 3.0T MRI system at the National Institutes of Health Clinical Center.

MRI Compatibility

The signal-to-noise (SNR) ratios of MR images in a homogenous phantom filled with CuO4 dilated saline were measured with the robot (1) absent, (2) present, and (3) in motion. The maximum SNR variation was within the acceptable range at <7.5%.

Accuracy Validated in MRI

Insertion accuracy was tested by performing brachytherapy seed placement on a prostate phantom (Fig. 2a-d). The positioning errors for the seed placements (n=10) were μ=0.9mm, σ=0.4mm perpendicular to the insertion axis, and μ=1.9mm, σ=2.7mm parallel to the insertion axis. The seed positions were calculated from a volumetric scan with perpendicular pixels of 1.18mm per side and image slices 3.5mm thick.

System Efficacy

Catheter ablation accuracy was also measured in trials with anatomical prostate phantoms with physical tumors (n=5) and temperature-sensitive phantoms with virtual tumors (n=8). Ablation position error was under 2mm (μ=1.7mm, σ=0.2mm). In every test (Fig. 2e), 100% of the tumor volume was ablated.

The robot was tested as an assistive technology for MRI-guided laser ablations. This FLA system increases accuracy in needle placement while enabling manually-driven remote insertions. The target tumor volumes were ablated according to the treatment plan, preserving the surrounding tissue regions. Future work includes animal and human studies to validate the effectiveness of the robot and software for MRI-guided FLA.