Semi-Quantitative Assessment of Surgical Navigation Accuracy During Endoscopic Sinus Surgery in a Real-World Environment.

Introduction: Although surgical navigation is commonly used in rhinologic surgery, data on real world performance are sparse because of difficulties in collecting measurements for target registration error (TRE). Despite publications showing submillimeter TRE, surgeons do report TRE of >3 mm. We describe a novel method for assessing TRE during surgery and report findings with this technique.

Methods: The TruDi navigation system (Acclarent, Irving, CA) was registered using a contour-based protocol. The surgeon estimated target registration error (e-TRE) at up to 8 points (anatomic regions of interest [ROI]) during endoscopic sinus surgery (ESS). System logs were used to simulate the localization for quantitative assessment of TRE (q-TRE).

Results: We performed 98 localizations in 20 patients. The ROI in the sinuses were ethmoid (33 sites), maxillary (28 sites), frontal (17 sites), and sphenoid (22 sites). For localizations, mean qTRE and eTRE were 0.93 and 0.84 mm (P = .56). Notably, 80% of qTRE and 81% of eTRE were 1 mm or less. Mean qTRE and eTRE were less for attending-performed registrations at the maxillary, frontal and sphenoid.

Conclusion: Surgical navigation accuracy, as measured by qTRE and eTRE, approaches 1 mm or better at all sinus sites in a real-world setting for 80% of localizations. The qTRE method provides a unique approach for assessing TRE. Surgeons underestimate TRE (overstate navigation accuracy), but this difference does not seem to be statistically significant. Registration performed by trainees yields higher TRE than registration performed by attendings. These data may be used to guide navigation optimization.