Robot-assisted versus pedicle projection localization in unilateral biportal endoscopic spine surgery: a trade-off analysis of fluoroscopy use, procedural time, and learning curve.

Precise localization of the entry point is essential in unilateral biportal endoscopic (UBE) spine surgery. While the standard pedicle projection method (P-UBE) is simple, it depends substantially on surgeon experience and real-time fluoroscopy. Robot-assisted localization (r-UBE) has been developed to facilitate more reliable guidance in UBE. However, its clinical value in this context remains underexplored. This study assesses the trade-offs between the two techniques, focusing on fluoroscopy requirements, operative efficiency, and an exploratory learning-curve characterization for robotic localization. This retrospective, non-randomized, single-center cohort study included 86 consecutive patients who underwent single-level UBE for lumbar spinal stenosis or disc herniation. Group allocation was based on surgeon preference and availability of the robotic system, resulting in 32 patients in the r-UBE group (using the TiRobot system) and 54 in the P-UBE group (using conventional fluoroscopic landmarks). Primary outcomes were fluoroscopy frequency, number of localization adjustments, and localization confirmation time. Secondary outcomes included total operative time, clinical scores (VAS, ODI, MacNab), and an exploratory nonlinear learning-curve analysis for the r-UBE group. The r-UBE group was associated with a lower fluoroscopy frequency (2.1 ± 1.3 vs. 4.4 ± 1.8; mean difference - 2.3, 95% CI - 3.0 to - 1.6; p < 0.01) and fewer localization adjustments (1.2 ± 0.4 vs. 1.8 ± 0.9; mean difference - 0.6, 95% CI - 0.9 to - 0.3; p < 0.01). However, localization confirmation time was significantly longer in the r-UBE group (12.5 ± 3.5 min vs. 8.5 ± 2.5 min; mean difference 4.0 min, 95% CI 2.7 to 5.3; p < 0.01), primarily due to the registration process (8.2 ± 2.8 min). An exploratory nonlinear reconstruction suggested rapid early improvement followed by stabilization; these estimates are based on a small case series and require confirmation in larger cohorts. Total operative time did not differ significantly (92 ± 21 min vs. 88 ± 18 min; mean difference 4.0 min, 95% CI - 5.1 to 13.1; p = 0.32), and both groups achieved similarly favorable clinical outcomes (96.9% vs. 94.4% excellent/good, p > 0.99). No robot-related issues were found. In this cohort, robot-assisted UBE localization was associated with reduced intraoperative fluoroscopy frequency and fewer repositioning attempts, but required a longer initial time investment, which may decrease with accumulated experience. While P-UBE is efficient and less resource-intensive for routine cases, r-UBE may provide a useful alternative, particularly in scenarios where P-UBE's limitations are most pronounced. The choice should be based on patient anatomy, surgeon experience, and resource availability. Importantly, the proposed advantages for complex anatomies or early-career surgeons remain hypothesis-generating and require prospective validation; furthermore, fluoroscopy frequency is a surrogate for intraoperative radiation and does not directly measure radiation dose.