Investigating Type B Basilar Invagination Through Cephalic Indices

OBJECTIVE: We sought to compare the diagnostic accuracy of cephalic indices for type B basilar invagination (BI). METHODS: This retrospective study examined head and neck magnetic resonance imaging sequences of 31 Group B BI cases and 96 controls. Two examiners blinded to diagnostic data evaluated the cephalic indices of each magnetic resonance imaging sequence, described as width/length (WLI) and height/width (HWI). The distance of the odontoid process apex to Chamberlain line and clivus canal angle were measured. The interexaminer and intraexaminer reproducibility of the cephalic indices was calculated using intraclass correlation coefficient. The diagnostic accuracy was discerned by the receiver operating characteristic (ROC) curve. All analyses were scrutinized with a 95% confidence interval. RESULTS: Cephalic indices showed interexaminer and intraexaminer reproducibility >= 94%. The areas under the ROC curve were 0.639 (WLI) and 0.874 (HWI) (95% confidence interval: P < 0.05). The HWI showed a sensitivity of 74.7% and a specificity of 85.5% for the cutoff criterion .58. The WLI presented a sensitivity of 53.3% and a specificity of 66.7% for the cutoff criterion >= 86. CONCLUSIONS: The HWI showed the largest area under the ROC curve in comparison with the WLI, with robust sensitivity and specificity values, indicating that the proportions between cranial height and width can help clinicians in investigating type B BI.