ABSTRACT
INTRODUCTION: Several techniques for pedicle screw placement have been described including freehand techniques, fluoroscopy assisted, computed tomography (CT) guidance, and robotics. Image-guided surgery offers the potential to combine the benefits of CT guidance without the added radiation. This study investigated the ability of a neural network to place lumbar pedicle screws with the correct length, diameter, and angulation autonomously within radiographs without the need for human involvement. MATERIALS AND METHODS: The neural network was trained using a machine learning process. The method combines the previously reported autonomous spine segmentation solution with a landmark localization solution. The pedicle screw placement was evaluated using the Zdichavsky, Ravi, and Gertzbein grading systems. RESULTS: In total, the program placed 208 pedicle screws between the L1 and S1 spinal levels. Of the 208 placed pedicle screws, 208 (100%) had a Zdichavsky Score 1A, 206 (99.0%) of all screws were Ravi Grade 1, and Gertzbein Grade A indicating no breech. The final two screws (1.0%) had a Ravi score of 2 (<2 mm breech) and a Gertzbein grade of B (<2 mm breech). CONCLUSION: The results of this experiment can be combined with an image-guided platform to provide an efficient and highly effective method of placing pedicle screws during spinal stabilization surgery.
ABSTRACT
PURPOSE: Augmented reality-based image overlay of virtual bony spine anatomy can be projected onto real spinal anatomy using computer tomography-generated DICOM images acquired intraoperatively. The aim of the study was to develop a technique and assess the accuracy and feasibility of lumbar vertebrae pedicle instrumentation using augmented reality-assisted surgical navigation. SUBJECTS AND METHODS: An augmented reality and artificial intelligence (ARAI)-assisted surgical navigation system was developed. The system consists of a display system which hovers over the surgical field and projects three-dimensional (3D) medical images corresponding with the patient's anatomy. The system was registered to the cadaveric spine using an optical tracker and arrays with reflective markers. The virtual image overlay from the ARAI system was compared to 3D generated images from intraoperative scans and used to percutaneously navigate a probe to the cortex at the corresponding pedicle starting point. Intraoperative scan was used to confirm the probe position. Virtual probe placement was compared to the actual probe position in the bone to determine the accuracy of the navigation system. RESULTS: Four cadaveric thoracolumbar spines were used. The navigated probes were correctly placed in all attempted levels (n = 24 levels), defined as Zdichavsky type 1a, Ravi type I, and Gertzbein type 0. The virtual overlay image corresponded to the 3D generated image in all the tested levels. CONCLUSIONS: The ARAI surgical navigation system correctly and accurately identified the starting points at all the attempted levels. The virtual anatomy image overlay precisely corresponded to the actual anatomy in all the tested scenarios. This technology may lead more uniform outcomes between surgeons and decrease minimally invasive spine surgery learning curves.
