Use of Intraoperative Navigation Update to Improve Accuracy during Skull Base Surgery

The use of augmented reality in the form of a heads-up display (HUD) during intracranial surgery has shown promise for improving approach planning. HUD highlights target pathology but also alerts operators to critical structures that are often difficult to appreciate using conventional preoperative imaging modalities. HUD allows for increased spatial awareness of these structures while providing the opportunity for surgical approach adjustment in the pre- and perioperative periods. Navigation Update (NU) is a BrainLab software that permits enhanced registration using relevant intraoperative structures under magnification.

Quantitative and qualitative data on the utilization and usability of NU with HUD in 11 neurosurgical procedures was obtained from standardized surveys completed at the Mount Sinai Hospital between March 2018 and August 2018. Variables included patient demographics, pathology type, location, as well as intraoperative manipulation of navigation information and registration. The team implemented a robust registration method using BrainLab’s Softtouch. Registration points were taken at eyes, ears, and nose but also using a distribution over the head, which was found to aid in deep registration of the z-axis.

Eleven cases (8 males with mean age of 57 years) were included in the analysis. The most common pathologies included meningioma (n = 3) (27%) and AVM (n = 2) (8.2%). The most common locations for pathologies included the anterior skull base (27.3%) and cerebellar pontine angle (27.3%). Most cases included identified the pathology as a deep (63.6%) compared with a superficial (36.4%) lesion.

In addition to visualizing the target pathology, HUD was used to maintain spatial awareness of critical structures, primarily the vasculature (90.9%), cranial and optic nerves (63.6%). Initial HUD platform accuracy as measured prior to skin incision was rated as either “excellent” (72.7%, defined as <1 mm deviation from the true location) or “good” (27.3%, 1–2 mm). 3D visualization and HUD utility were measured as change in surgical plan (73% of cases) where 55% of cases had a change in approach due to HUD prior to skin incision.

After skin incision if registration was deemed inaccurate, objects were observed to shift linearly by 1 to 10 mm. The most common cause of intraoperative inaccuracies was attributed to brain shift during exposure. The NU tool used planned objects (77.7%) and maximum intensity projection (MIP) (22.3%). Objects were used for adjusting registration through the NU tool in a drag and drop fashion to the orientation of HUD to patient’s anatomy. Choice of re-registration technique was made according to surgeon preference.

HUD is most helpful when accuracy remains constant intraoperatively. NU is able to help recover HUD registration accuracy even when the shift is relatively high (10 mm). Users experience greater satisfaction when NU corrects errors of greater than 3 mm, the size of nerves and vessels of interest in the field. HUD can still be useful even when inaccurate if the painted structures have at least 50% overlap with actual anatomy. Users reported that successful application of NU restores HUD benefits to the surgeon including: awareness of critical structures out of field, improved confidence and accuracy in dissection, and improved safety.

No conflict of interest has been declared by the author(s).