Introducing and Modeling a Soft Steerable Catheter Tip Using Multiobjective Design Optimization

Abstract

Mechanical thrombectomy is a life-saving procedure used to treat ischemic stroke. It requires endovascular navigation through tortuous pathways. This navigation presents a particular challenge for some anatomy, such as type III aortic arches. This letter presents a steerable soft catheter tip that can augment existing non-steerable catheters and facilitate navigation. The elastomeric device uses two embedded tendons for actuation. A mechanics model is first presented based on a Cosserat rod model. Since design optimization for such devices requires heavy reliance on solving the mechanics, which in turn, requires repeated numerical integration of the static equilibrium differential equations, a novel approach is needed to alleviate the computational burden and to facilitate the design optimization process. To achieve this computational speed enhancement, we present a method that leverages modal kinematics to encode the equilibrium shapes of the soft catheter tip. We then utilize this encoding method to formulate the sensitivity of the device shape to design parameter uncertainty (material property and locations of the termination points of the actuation tendons within the device). Using these closed-form formulations, we present a multiobjective design optimization framework allowing improved path following while minimizing sensitivity to model parameter uncertainty. We validate our approach experimentally on a device fabricated to match our optimized design and demonstrate how the device can navigate through a type III aortic arch. The bending of the device was compared to the simulation, showing an error at the tip of less than 5 mm and 21 degrees.