Learning curve of computer-assisted navigation system in spine surgery.

BACKGROUND: Spine surgery using computer-assisted navigation (CAN) has been proven to result in low screw misplacement rates, low incidence of radiation exposure and excellent operative field viewing versus the conventional intraoperative image intensifier (CIII). However, as we know, few previous studies have described the learning curve of CAN in spine surgery. METHODS: We performed two consecutive case cohort studies on pedicel screw accuracy and operative time of two spine surgeons with different experience backgrounds, A and B, in one institution during the same period. Lumbar pedicel screw cortical perforation rate and operative time of the same kind of operation using CAN were analyzed and compared using CIII for the two surgeons at initial, 6 months and 12 months of CAN usage. RESULTS: CAN spine surgery had an overall lower cortical perforation rate and less mean operative time compared with CIII for both surgeon A and B cohorts when total cases of four years were included. It missed being statistically significant, with 3.3% versus 4.7% (P = 0.191) and 125.7 versus 132.3 minutes (P = 0.428) for surgeon A and 3.6% versus 6.4% (P = 0.058), and 183.2 versus 213.2 minutes (P = 0.070) for surgeon B. In an attempt to demonstrate the learning curve, the cases after 6 months of the CAN system in each surgeon's cohort were compared. The perforation rate decreased by 2.4% (P = 0.039) and 4.3% (P = 0.003) and the operative time was reduced by 31.8 minutes (P = 0.002) and 14.4 minutes (P = 0.026) for the CAN groups of surgeons A and B, respectively. When only the cases performed after 12 months using the CAN system were considered, the perforation rate decreased by 3.9% (P = 0.006) and 5.6% (P < 0.001) and the operative time was reduced by 20.9 minutes (P < 0.001) and 40.3 minutes (P < 0.001) for the CAN groups of surgeon A and B, respectively. CONCLUSIONS: In the long run, CAN spine surgery decreased the lumbar screw cortical perforation rate and operative time. The learning curve showed a sharp drop after 6 months of using CAN that plateaued after 12 months; which was demonstrated by both perforation rate and operative time data. Careful analysis of the data showed CAN is especially useful for less experienced surgeon to reduce perforation rate and intraoperative time, although further comparative studies are anticipated.