Accuracy of craniocervical measurements on CT for identifying partial or complete craniocervical ligament injuries in pediatric patients

Purpose To assess the accuracy of craniocervical measurements for identifying craniocervical injuries and the frequency of subjective findings of craniocervical injuries on CT in pediatric patients. Methods Case-controlled retrospective review of patients <= 16 years old with craniocervical junction injuries. Receiver operator curves were created for common craniocervical measurements on CT comparing patients with complete and partial craniocervical injuries to uninjured cohort. Frequency of subjective CT findings of craniocervical injury was assessed in the injured cohort. Results For complete disruption injuries (CD) (n = 27), C1-C2 distance (AUC = 0.90, 95%CI = 0.83-0.97), atlanto-occipital distance (AUC = 0.95-0.98, 95%CI = 0.90-1.00), and basion-dens distance (AUC = 0.90, 95%CI = 0.82-0.98) had excellent accuracy diagnosing injury. Powers ratio (AUC = 0.85, 95%CI = 0.76-0.94) had good, basion-posterior axial line (AUC = 0.74, 95%CI = 0.61-0.86) fair, and atlanto-dental distance (AUC = 0.69, 95%CI = 0.57-0.82) poor accuracy. For partial disruption injuries (PD) (n = 21), basion-dens distance (AUC = 0.75, 95%CI = 0.62-0.88) had fair accuracy diagnosing injury. Powers ratio (AUC = 0.63, 95%CI = 0.47-0.79), C1-C2 distance (AUC = 0.60, 95%CI = 0.45-0.75), atlanto-dental distance (AUC = 0.55, 95%CI = 0.39 = 0.71), atlanto-occipital distance (AUC = 0.63-0.65, 95%CI = 0.47-0.81), and basion-posterior axial line (AUC = 0.60, 95%CI = 0.44-0.76) all had poor accuracy. Eighty-one percent (n = 22) of CD and 38% (n = 8) of PD patients had non-concentric atlanto-occipital joints. One hundred percent of CD patients had >= 1 soft tissue finding and eighty-one percent (n = 22) had >= 2 findings. Seventy-three percent (n = 16) of PD patients had >= 1 soft tissue finding. Eighty-six percent (n = 18) of PD patients had non-concentric atlanto-occipital joints and/or soft tissue findings. Conclusion Craniocervical measurements have poor accuracy for identifying craniocervical injuries in pediatric patients with incomplete craniocervical ligament disruption. Subjective findings of craniocervical injury are frequently present on CT in pediatric patients and can help increase sensitivity for identifying injury.