Abstract
To assess the diagnostic performance of chest CT in the detection of rib fractures in children investigated for suspected physical abuse (SPA).
Medline, Web of Science and Cochrane databases were searched from January 1980 to April 2020. The QUADAS-2 tool was used to assess the quality of the eligible English-only studies following which a formal narrative synthesis was constructed. Studies reporting true-positive, false-positive, true-negative, and false-negative results were included in the meta-analysis. Overall sensitivity and specificity of chest CT for rib fracture detection were calculated, irrespective of fracture location, and were pooled using a univariate random-effects meta-analysis. The diagnostic accuracy of specific locations along the rib arc (anterior, lateral or posterior) was assessed separately.
Of 242 identified studies, 4 met the inclusion criteria. Of these, 2 were included in the meta-analysis. Chest CT identified 142 rib fractures compared to 79 detected by initial skeletal survey chest radiographs in live children with SPA. Post-mortem CT (PMCT) has low sensitivity (34%) but high specificity (99%) in the detection of rib fractures when compared to the autopsy reference standard. PMCT has low sensitivity (45%, 21% and 42%) but high specificity (99%, 97% and 99%) at anterior, lateral and posterior rib locations, respectively.
Chest CT detects more rib fractures than initial skeletal survey chest radiographs in live children with SPA. PMCT has low sensitivity but high specificity for detecting rib fractures in children investigated for SPA.
• PMCT has low sensitivity (34%) but high specificity (99%) in the detection of rib fractures; extrapolation to CT in live children is difficult. • No studies have compared chest CT with the current accepted practice of initial and follow-up skeletal survey chest radiographs in the detection of rib fractures in live children investigated for SPA.