Abstract
OBJECTIVE: To assess the stability of odontoid parameters on flexion-extension motion and to validate the accuracy of the physiological cervical lordosis (CL) predictive formula across different cervical positions. METHODS: Standard cervical spine lateral radiographs in neutral, flexion, and extension positions were collected to measure odontoid incidence (OI), odontoid tilt (OT), C2 slope (C2S), CL, T1 slope (T1S), and T1S minus CL (T1S-CL). Friedman's test was used to assess the differences in parameters among the three cervical spine positions. The predictive performance of the formula CL = 0.36× OI - 0.67 × OT - 0.69 × T1S was assessed and validated using the Pearson correlation coefficient (r), coefficient of determination (R²), mean squared error (MSE), and mean absolute error (MAE). RESULTS: No significant differences were found for OI and T1S among the three different cervical spine positions (p = 0.162 and p = 0.186, respectively). There was a strong, significant positive correlation between the predicted value and actual value of physiological CL at three cervical positions (neutral, flexion, and extension). The predictions for the neutral position were the most accurate, with statistical measures of r = 0.85 (p < 0.01), R²=0.82, MAE = 4.28, and MSE = 27.77. CONCLUSION: OI is a stable and reliable anatomic parameter, not affected by cervical spine flexion-extension motion. OI can serve as a supplementary parameter for evaluating cervical sagittal balance and compensatory ability. The formula CL = 0.36×OI - 0.67×OT - 0.69×T1S provides the best predictions for physiological CL in the neutral position.