Artificial Intelligence-Assisted Microcatheter Shaping for Intracranial Aneurysm Coiling: A Preliminary Study

Abstract

To evaluate the efficacy of artificial intelligence (AI) technology-assisted microcatheter shaping for coil embolization of intracranial aneurysms. From June 2019 to May 2021, 30 aneurysms in 24 patients were treated with coiling embolization using computer software-assisted microcatheter shaping at our institution. All patients underwent digital subtraction angiography (DSA) before coiling embolization. After three-dimensional (3D) rotational angiography, digital imaging and communications in medicine (DICOM) data were extracted and imported into computer software based on an AI algorithm. 3D images of the parent artery and aneurysm were constructed with the software and data including the central axis of the parent artery, aneurysm location, aneurysm size, and 3D structure were automatically obtained. The optimal microcatheter path was calculated and the shape of the mandrel was automatically generated. Surgeons shaped the mandrel and microcatheter following the AI-generated template and completed the endovascular procedure. All patients successfully completed the endovascular procedure without perioperative complications. The microcatheters shaped as per the AI template accurately entered the aneurysm sacs in 1 attempt; 15 aneurysms required no microguidewire assistance in catheterizing the aneurysm sac and 15 did. The stability of the microcatheters during the procedures was satisfactory. No rebound incidence was observed and no reshaping was necessary. The AI-assisted microcatheter shaping technology provides a new method to generate the optimal shape for the mandrel and microcatheter during endovascular procedures. The technology facilitates microcatheter accuracy and stability during coiling embolization and provides technical support for surgeons.