Purpose or Case Report: Cervical spine trauma, particularly at the craniocervical junction, can produce significant morbidity and mortality in the pediatric population. Although normative pediatric cervical spine CT measurements have been reported, the accuracy of these measurements for predicting ligamentous injury remains unknown. The purpose of this study is to evaluate the sensitivity and specificity of normative pediatric CT measurements for prediction of traumatic ligamentous injury at the craniocervical junction, using MRI as gold standard.
Methods & Materials: Following IRB approval, a retrospective review of 64 pediatric (age <=21 years) trauma patients who underwent cervical spine CT and gold standard cervical spine MRI was performed. Craniocervical junction measurements were obtained from the CT cervical spine including: basion-dens interval (BDI), anterior atlanto-dental interval (AADI), lateral atlanto-dental interval (LADI), prevertebral thickness at C2, atlanto-occipital interval (AOI), lateral atlantoaxial interval (LAAI), and Powers ratio. Cervical spine MRIs were retrospectively reviewed to identify ligamentous injury to the craniocervical junction. Thresholds for normal and abnormal CT measurements were established from normative pediatric cervical spine CT measurements and compared to the MRI cervical spine to determine the sensitivity and specificity of individual measurements, any single abnormal measurement, and combinations of abnormal measurements for detection of ligamentous injury.
Results: Sensitivity and specificity of individual measurements were as follows: BDI 0.28/1.0, AADI 0.44/0.58, LADI 0.17/0.77, prevertebral thickness 0.28/0.7, AOI 0.67/0.81, LAAI 0.86/0.31, Powers ratio 0.22/0.98. Sensitivity and specificity of one or more of the described measurements being abnormal is 1/0.12, two or more 0.94/0.19, three or more 0.89/0.42, four or more 0.72/07, five or more 0.67/0.84, six or more 0.33/0.93 and seven or more 0.28/0.95.
Conclusions: Sensitivity and specificity of CT criteria for prediction of traumatic ligamentous injury at the craniocervical junction are generally nonspecific in isolation but improve in validity when five or more criteria are met as abnormal.