Diagnostic accuracy of screening tools for pediatric blunt cerebrovascular injury: An ATOMAC multicenter study

Background: Blunt cerebrovascular injury (BCVI) is rare but significant among children. There are three sets of BCVI screening criteria validated for adults (Denver, Memphis, and Eastern Association for the Surgery of Trauma criteria) and two that have been validated for use in pediatrics (Utah score and McGovern score), all of which were developed using retrospective, single-center data sets. The purpose of this study was to determine the diagnostic accuracy of each set of screening criteria in children using a prospective, multicenter pediatric data set.

Methods: A prospective, multi-institutional observational study of children younger than 15 years who sustained blunt trauma to the head, face, or neck and presented at one of six level I pediatric trauma centers from 2017 to 2020 was conducted. All patients were screened for BCVI using the Memphis criteria, but criteria for all five were collected for analysis. Patients underwent computed tomography angiography of the head or neck if the Memphis criteria were met at presentation or neurological abnormalities were detected at 2-week follow-up.

Results: A total of 2,284 patients at the 6 trauma centers met the inclusion criteria. After excluding cases with incomplete data, 1,461 cases had computed tomography angiography and/or 2-week clinical follow-up and were analyzed, including 24 cases (1.6%) with BCVI. Sensitivity, specificity, positive predictive value, and negative predictive value for each set of criteria were respectively 75.0, 87.5, 9.1, and 99.5 for Denver; 91.7, 71.1, 5.0, and 99.8 for Memphis; 79.2, 82.7, 7.1, and 99.6 for Eastern Association for the Surgery of Trauma; 45.8, 95.8, 15.5, and 99.1 for Utah; and 75.0, 89.5, 10.7, and 99.5 for McGovern.

Conclusion: In this large multicenter pediatric cohort, the Memphis criteria demonstrated the highest sensitivity at 91.7% and would have missed the fewest BCVI, while the Utah score had the highest specificity at 95.8% but would have missed more than half of the injuries. Development of a tool, which narrows the Memphis criteria while maintaining its sensitivity, is needed for application in pediatric patients.

Level of evidence: Diagnostic Test/Criteria; Level II.