Precision insertion of sacroiliac screws using an augmented reality-based navigation system is feasible and accurate

Wang et al Fig 3
Intra-operative drilling under the AR-based navigation with virtual images superimposed on the HMD


An experienced orthopaedic surgeon was able to successfully and accurately place percutaneous sacroiliac screws into a number of cadavers using a new augmented reality (AR) navigation system during a pilot study published in International Orthopaedics. The study authors hope that AR navigation will provide “a valuable tool for assisting percutaneous sacroiliac screw insertion in live surgery.”

Conventional modern surgical navigation systems—using fluoroscopy and CT imaging—hold a number of drawbacks for surgeons. As well as exposing both surgeon and patient to radiation, the visualisation of navigational data on monitors can impede surgical outcomes. Visualisation takes place at a different viewing angle to the surgical site, requiring unfamiliar hand-eye coordination techniques. This steepens the learning curve for new surgeons, and can hinder concentration and cause distractions. The aim of this pilot study from the Shanghai Jiao Tong University School of Medicine, Shanghai, China, was to assess the feasibility and accuracy of a head-mounted AR-based navigational system, which could address some of these problems. Principal investigator, Qiugen Wang, told Spinal News International, “AR can provide an intuitive visualisation for the surgeon of three-dimensional images of virtual planning and of internal critical anatomical structures.”

The AR system consisted of an optical see-through head-mounted display (HMD), an optical tracking system a graphical workstation with an LCD monitor. The HMD was connected to the workstation, and, using infrared-reflecting (IR) markers, the optical tracking system was able to capture the position of objects such as the drill, the HMD and the pelvis. This system allows the surgeon to view virtual images of the surgical site, which correspond to head movements in real-time. The virtual images are projected onto the transparent screen of the HMD, which should allow a surgeon to see the real world unhindered. This allows for native hand-eye coordination, which could theoretically mediate the steep learning curve, and potential distraction, associated with visualisation through a conventional monitor.

3D models of each of the pelvis and vessels for the six cadavers in the study were created preoperatively with CT data and the Mimics (Materialise, Belgium) planning software. This allowed for the virtual planning of ideal sacroiliac screw trajectories, which were modelled by cylinders.  Twelve percutaneous screws were placed in the cadavers—which had intact pelvises—by an experience orthopaedic surgeon, who was allowed time before the surgeries to familiarise himself with the AR navigation system. Post-operative CT images were compared with the ideal screw trajectory cylinder models by two independent raters, to assess the outcomes of each surgery.

Pre-operative 3D reconstruction of the pelvis, planned trajectory, and adjacent vessels
Pre-operative 3D reconstruction of the pelvis, planned trajectory, and adjacent vessels

All 12 screws were placed successfully, with no perforations. The mean time required for the experiment was 13.6±2.2 mins, and for the procedure was 11.1±2 mins. The mean deviation between the planning cylinders and the actual placement of screws was 2.7±1.2mm (1.3-5.5mm) at the bony entry point of the screw, and 3.7±1.1mm (1.1-5.2mm) at the tip. The mean angular deviation was 2.9°±1.1° (1.6-4.8°), and the mean distance between the centres of the planning cylinders and the inserted screw at the level of the nerve root tunnel region on the sagittal plain was 3.6±1mm (1.4±4.7). Whilst the result showed lower accuracy than other study results for CT-3D-fluoroscopy navigation, the authors presume that this level of accuracy is acceptable assuming that the planning cylinder “is always positioned more than 5mm away from the outer cortical margin on the sagittal plane near the nerve root tunnel area. The authors write that a comparison between the accuracy of the AR system and current navigation systems would be an important area for future research.

The accuracy could be improved, they write, by further refinement of the software and instrumentations used for the procedure, and by training users to be more familiar with the system. “I believe that with the constant development of AR technology, it will eventually become a part of the surgical routine,” Wang told Spinal News International.


The study was limited by the small number of cases, and by the normal physiology of the cadavers. Further research into the use of the system in cases with sacroiliac fractures or dislocations would be necessary to test the application of AR to cases more serious than those with minimal sacroiliac joint displacement.