Soft fluid-driven robotic positioners for intra-operative MRI-guided interventions

Abstract

Magnetic resonance imaging (MRI) plays a critical role in managing debilitating diseases, including heart rhythm disorder, neurological abnormalities, and many forms of cancer. Providing high-contrast two-dimensional (2D) and three-dimensional (3D) images of soft tissue without harmful ionizing radiation, it is a compelling and powerful tool for clinicians. Needle-based procedures, e.g., ablation treatments, have significantly benefited from MRI guidance, with which thermal distributions can be evaluated in real-time to protect surrounding critical structures while ensuring adequate ablation margins. The rapidly developing robotic assistance has widened the horizons of MRI-guided surgery, bringing many advantages, including improved intervention precision, reduced operating time and costs, as well as smaller incisions. However, not only does the confined space of the MRI bore and patient coils restrict the access of the surgeons and external devices, but also the high-strength magnetic and radiofrequency (RF) fields generated by the MRI scanner prevents the use of materials that are commonly used to construct robotic systems.

The core focus of this thesis is to investigate and tackle several key technical challenges faced in intra-operative (intra-op) MRI-guided interventions, with the development of small form factor robotic assistants which can provide precise and dexterous instrument positioning. To this end, fluid-driven soft manipulators are proposed with reinforcements on the soft chambers, allowing a high elongation ratio while maintaining bending flexibility. FEA was conducted to simulate the robot characteristics while being applied with fluidic pressure, thus facilitating optimization towards balanced soft robot fatigue life, structural rigidity and actuation linearity. Serving as essential components, the soft manipulator is incorporated into a patient-mounted robotic device for MRI-guided percutaneous needle procedures (e.g., liver tumor ablation). The compact (Ø108 mm × 115 mm height) and lightweight (189 g) robot architecture allows direct installation of the robot on the patient body. Multiple robots could be deployed for simultaneous targeting. Further building on these concepts, a multi-stage robotic positioner is proposed for MRI-guided bilateral stereotactic neurosurgery. The system incorporates intuitive, hands-on initial adjustment, followed by soft-robotic final positioning. While maintaining a sufficient workspace (± 35°), the positioner remains lightweight (203 g) and compact (Ø 81 mm × 97 mm height), enabling skull-mounted usage within most standard imaging head coils and in a bilateral configuration. An MRI-guided robotic positioner is also developed for focused ultrasound systems with diaphragm-based hydraulic actuation. Its foci can be steered across wide spatial coverage for abdominopelvic organ disease treatment. The high positional frequency response of the robot allows for respiratory motion (<0.2 Hz) compensation. Lab-based testing is conducted to evaluate the soft actuation and the integrated platforms’ performances, e.g., positioning accuracy and feedback control responsiveness. MRI-based experimental validation of overall system workflows is also conducted, including robot registration and tracking with custom-made omni-directional MRI markers. For the robotic positioner aiming at MRI-guided bilateral stereotactic neurosurgery, an MRI-based targeting accuracy test in a phantom is also conducted. No notable reduction in imaging quality was observed during all robot prototypes’ operation in the MRI compatibility tests.