No abstract

Background: The purpose of this study was to investigate the sensitivity and specificity of current cervical prevertebral soft tissue swelling (PVST) values in a cohort of children with known cervical fractures or dislocations. Methods: Forty two children (average age 11.9, range 1.4 to 17.0 y) with documented cervical spine injury and 61 children (average age 11.9, range 0.5 to 17.9 y) with cervical pain but no injury were reviewed (January 2004 to December 2015). PVST was measured on lateral cervical radiographs at C2, C3, and C6. Patients were stratified by age (0 to 2 y, 3 to 6 y, 7 to 10 y, 11 to 15 y, and 16 y and above). The Wilcoxon rank sum test was used to compare PVST measurements at each spine level across injury and noninjury cohorts. Sensitivity and specificity were estimated to assess the ability of abnormal reference values to detect when a true injury was present. In addition, positive predictive value and negative predictive value were also estimated. Results: The majority of c-spine injuries (31/42; 76%) involved bony fracture and 57% (24/42) were treated with a collar or brace. Comparison of PVST measurement found no difference at C2 (P=0.07), C3 (P=0.07), or at C6 (P=0.99) across injury and non-injury cohorts. Sensitivity was poor at single-level measures for C2 (26%), C3 (31%), and C6 (24%), while specificity was relatively high (92%, 87%, and 79%, respectively). When an increased value at either C2 or C3 indicated injury, sensitivity increased to 36%, and when an increased measurement at just one of the 3 measured levels indicated injury, the sensitivity increased to 48%, while the specificity decreased to 72%. While retropharyngeal measures were more likely to detect injury than retrotracheal, C6 alone was increased in 5 of the 20 injury patients. Conclusions: PVST measurements exhibit poor sensitivity but good specificity as indicators for the diagnosis of occult cervical trauma in children. Negative values do not exclude injury; positive values suggest further evaluation. Levels of Evidence: Level III