X-Ray-free neuroendovascular training using optically-tracked 3D-printed phantoms

Transactions on Additive Manufacturing Meets Medicine

Abstract

Endovascular interventions for cerebrovascular diseases require extensive training due to their technical complexity and high-risk nature. Virtual reality simulators offer repeatable, risk-free practice but often either lack anatomical realism and accurate haptic feedback, or are expensive, limiting skill transferability to clinical settings [1,2]. While physical phantoms using 3D-printed patient-specific vasculature provide superior anatomical fidelity and realistic device-vessel interactions, they typically require fluoroscopic imaging for visualization, exposing trainees and instructors to ionizing radiation.. This radiation dependency limits training frequency and accessibility. Hybrid approaches combining physical phantoms with alternative imaging modalities remain underexplored [3]. This study developed a radiation-free hybrid system combining modular patient-specific vascular phantoms with optical tracking for endovascular training. 3D-printed transparent modular phantoms were created from contrast-enhanced computer tomography data (Formlabs Form 4, Clear resin) with flattened walls at camera positions to reduce distortion. Three cameras tracked colored instruments via color segmentation and background subtraction; 3D positions were estimated through ray-surface intersection. A C-arm simulator generated fluoroscopy-like projections using CT data from the Visible Human Project. The system tracked multiple instruments simultaneously and delivered radiation-free, realistic fluoroscopy projections. Additionally, force sensors tracked the forces applied to the aneurysm wall to warm about its puncture. Limitations included optical dead zones at vessel bifurcations, limited range of vessel diameters, and phantom transparency. This hybrid system enables radiation-free endovascular training with patient-specific anatomy, supporting multiple training scenarios and device validation.

Authors 5

1. Roman Leonov Aachen Fraunhofer IMTE \/

   RWTH Aachen University

   Affiliation as printed

   Fraunhofer IMTE \/ RWTH Aachen

2. Sonja Wichelmann

   Fraunhofer Institute for Digital Medicine

   Affiliation as printed

   Fraunhofer Institute for Digital Medicine MEVIS, Lübeck, Germany

3. André Behrends

   Fraunhofer-Einrichtung für Individualisierte Medizintechnik

   Affiliation as printed

   Fraunhofer Research Institution for Individualized Medical Technology and Engineering IMTE, Lübeck, Germany

4. Thomas Friedrich

   Fraunhofer-Einrichtung für Individualisierte Medizintechnik

   Affiliation as printed

   Fraunhofer Research Institution for Individualized Medical Technology and Engineering IMTE, Lübeck, Germany

5. T Buzug

   Fraunhofer-Einrichtung für Individualisierte Medizintechnik

   Affiliation as printed

   Fraunhofer Research Institution for Individualized Medical Technology and Engineering IMTE, Lübeck, Germany