Pediatric cervical spine injuries on CT: difference in accuracy of interpretations by pediatric versus non-pediatric radiologists.

Purpose: To retrospectively compare the accuracy of interpretation of initial cervical computerized tomography (CCT) by a non-pediatric radiologist (NPR) versus a pediatric radiologist (PR).

Methods: IRB approval and consent waiver were granted to review all injured children from 2010 to 2014 in the trauma registry with CT and magnetic resonance imaging (MRI) of the cervical spine. Patients with negative CCT who subsequently had positive MRI from a single institution comprised the study group. Patients with negative CCT and MRI, matched by age, gender, and severity scores, comprised the control group. The CCTs from both groups were initially interpreted at the time of service by a NPR. Subsequently, a single PR with 20 years of experience blinded to clinical/imaging data reinterpreted these CCT examinations. CT interpretations were then compared with MRI results and evaluated for statistical significance using SSPS software. The data analysis utilized summary statistics, two-tailed binomial test, and univariate χ test. Significance for all comparisons was assessed at P < 0.05.

Results: The study group was comprised of the 21 patients with negative CCT and positive MRI. Of the cohort included, 76% (16) were male and 24% (5) were female. The age range was 1 month-17 years, with a mean age of 9.7 years. CCT interpretation by NPR had a specificity of 91.7% (sensitivity 71.2%, positive predictive value 81.3%, and negative predictive value 86.3%) compared with results of MRI. Six of the 21 negative CCTs were interpreted by the PR as positive, mainly craniocervical junction injuries, and confirmed by MRI (28.6%, P < .001 compared with the NPR); no control CCT was interpreted by the PR as positive (sensitivity 100%, positive predictive value 100%, and negative predictive value 58.3%).

Conclusion: In our retrospective study, a pediatric radiologist has improved recognition of pediatric cervical spine injuries on CT compared with non-pediatric radiologist.