Patient-Specific IVR Endovascular Simulator with Augmented Reality for Medical Training and Robot Evaluation

Abstract

Endovascular intervention using interventional radiology (IVR) is most commonly used in cerebralvascular treatment. Medical imaging such as digital subtraction angiography (DSA) and vascular mapping make vasculature and catheters easier to read from fluoroscopy during endovascular intervention. We propose simulating IVR using augmented reality, reproducing fluoroscopic images and a patient-specific blood vessel model without X-ray imaging. The advantages of the patient-specific vascular model reproducing the human vasculature lumen with 13 μm resolution include 1) a realistic “feel,” 2) excellent tool behavior simulation during intervention, and 3) surgical training alternative to physician training in-vitro. Simulated fluoroscopic images are created in two steps: First, the blood vessel model refraction index is matched to surrounding glycerin solution to conceal the vascular model, making the silicone vasculature appear human as seen in endovascular intervention. Second, an augmented reality (AR) environment is created using image subtraction and overlap, making model-based endovascular simulation more understandable for catheter use and fluoroscopy use and reading.