Table 1.
Generalized linear model: error vs. screw location and navigation method.
| Clinical Variables | Screw Location and Error Median (95 percentile) | Navigation Method Median (95 percentile) | ||||||
|---|---|---|---|---|---|---|---|---|
| Cervical N = 9 | Thoracic N = 225 | Lumbar N = 130 | Sacral N = 6 | P-Value | Benchmark N = 209 | Experimental N = 162 | P-Value | |
| Axial Distance Error (mm) | 1.00 (1.73) | 1.05 (3.69) | 1.53 (3.94) | 1.03 (1.93) | 0.381 | 1.14 (3.92) | 1.21 (3.42) | 0.597 |
| Axial Angle Error (deg) | 2.79 (4.33) | 2.07 (8.07) | 2.48 (9.47) | 6.05 (11.59)* | 0.009* | 2.43 (8.97) | 2.15 (8.14) | 0.839 |
| Sagittal Distance Error (mm) | 1.32 (2.20) | 0.86 (1.28) | 1.28 (4.12) | 0.84 (2.13) | 0.437 | 0.83 (3.62) | 1.13 (4.25) | 0.214 |
| Sagittal Angle Error (deg) | 1.55 (4.22) | 2.57 (9.75) | 2.47 (9.48) | 3.86 (10.77) | 0.485 | 2.60 (10.06) | 2.33 (8.59) | 0.492 |
Statistical analysis using generalized linear model of axial distance, axial angle error, sagittal distance error, and sagittal angle error as a function of screw location (cervical, thoracic, lumbar, and sacral) and navigation method (benchmark navigation, experimental navigation). No significant differences were seen using the experimental navigation system. Screws located in the sacrum demonstrated increased axial angle error vs. thoracic and lumbar spine.