Associate Professor Department of Neurosurgery, Huashan Hospital, Fudan university
Introduction: We developed an AI-driven surgical robot to automate microcatheter shaping, a critical and challenging step in endovascular aneurysm surgery. Traditionally requiring extensive experience, manual shaping is performed by skilled neurosurgeons. Accurate shaping is essential for precise catheter placement and effective aneurysm coiling, ensuring both the safety and success of the procedure.
Methods: The AI surgical robot comprises three main parts: the vascular and aneurysm modeling module, the microcatheter simulation module, and the automatic shaping module. First, enhanced DICOM sequences of intracranial arteries are imported into the vascular and aneurysm reconstruction module, where vessels and aneurysms are extracted and smoothed to remove voxel boundary artifacts. The microcatheter simulation module detects potential collisions between the microcatheter and vascular walls, modeling the microcatheter as a Discrete Elastic Rod (DER) for path simulation. The automatic shaping module then forms the mandrel into the precise shape calculated, based on specifications of various microcatheter brands.
Results: Intelligent and automated endovascular surgery using the AI surgical robots has become daily procedures for endovascular therapy of intracranial aneurysms in the neurosurgery department in 23 institutes geographically located in 15 cities across China, in which 395 embolization surgeries of intracranial aneurysms (C5 24/6.1%, C6 124/31.4%, C7 76/19.2%, M1 87/22.0%, A1 44/11.1%, BA 5/1.3%, Others 35/8.9%.)have been conducted using the AI surgical robots. The rate of successful placement with single trial was 91.9%. No severe adverse events were observed since initial clinical use in November 2022. Application of the AI surgical robots has avoided repetitive trials for shaping and placement of microcatheter due to inaccurate shaping by manual, reduced exposure to radiation, and potentially reduces incidences of complications. More importantly, the application of AI surgical robot standardized the surgical processes, so that expected to reduce the steepness of learning curve for neurosurgeons with less experience.
Conclusion : The AI surgical robot reduces radiation exposure, minimizes complications, and standardizes procedural accuracy, easing the learning curve for less experienced neurosurgeons. It shows promise for improving the safety, efficiency, and accessibility of endovascular aneurysm treatments.