An Augmented Reality Navigation System for Computer Assisted Arthroscopic Surgery of the Knee

This work presents a prototype navigation system for computer assisted arthroscopic surgery. Arthroscopic surgery is minimally invasive and ensures fast post-surgical recovery. It requires, however, considerable dexterity due to the limited 2D view of the scene and cramped operating area. Augmented reality can help the surgeon overcome these limitations. The key feature of the system presented is an interactive graphical interface. A 3D model and two 2D projections of the patient’s joint and models of the arthroscope and other sensorized surgical instruments are shown. The positions of all sensorized objects, including the patient’s anatomy, are tracked in real-time by an optoelectronic localizer and reproduced in the virtual scene. The field of view of the arthroscope is dynamically highlighted on the 3D model of the joint, which is reconstructed from a preoperative CT/MRI data set and matched to the actual anatomy by means of an ICP-based non-fiducial registration algorithm. The prototype system has been positively evaluated by a selected group of skilled orthopaedic surgeons. Preliminary tests on the overall accuracy of the system show that errors can be kept within 3°/3mm.