An MRI-compatible endonasal surgical robotic system: Kinematic analysis and performance evaluation

Neurosurgical interventions are often complicated and require procedure-specific solutions for a better outcome. A pituitary tumor is one of the common brain tumors resected by trans-sphenoidal surgery. Magnetic resonance imaging (MRI) provides better vision when compared to other imaging modalities. Simultaneous resection and imaging of pituitary tumors under MRI can improve the surgical outcome. Herein, we propose the first MRIcompatible robotic system for pituitary tumor removal via trans-sphenoidal/endonasal access. The presented system follows the current gold standard procedure of pituitary tumor resection and integrates it with a commercial MR scanner to provide a widely applicable system. The robotic system is the procedure-, anatomy-, and geometry-specific with 6-degree-of-freedom actuated by a bed-type actuation platform in the MRI room. A K means clustering algorithm detects the tumor and develops and updates the brain model periodically during MR guided interventions. The robot has procedure-specific stiffness changing capability with a unique stiffness dependent kinematic model. The workspace subtended by the robotic system in all stiffness cases satisfies the workspace required for surgical procedures. The accuracy and repeatability of the robot are also in a desirable range. The procedure and patient-specific robot design are evaluated by in-vivo experiments in the live porcine under MRI. This work is a step toward a dual-arm surgical system for pituitary tumor removal under MRI guidance and experimental results validate the proposed solution and supplement further development to achieve a clinically applicable system.

Automated summary

This study proposes the first MRI-compatible robotic system for pituitary tumor removal via trans-sphenoidal/endonasal access. The presented robot follows the gold standard procedure of pituitary tumor resection and integrates with a commercial MR scanner, providing wide applicability. It has procedure-, anatomy-, and geometry-specific design, with 6-degree-of-freedom actuated by a bed-type actuation platform in the MRI room. The robot's accuracy and repeatability are within a desirable range, and in vivo experiments on live porcine under MRI validate the proposed solution.