Interobserver agreement for the Chest Wall Injury Society taxonomy of rib fractures using computed tomography images

Abstract

BACKGROUND In 2020, a universal nomenclature for rib fractures was proposed by the international Chest Wall Injury Society taxonomy collaboration. The purpose of this study is to validate this taxonomy. We hypothesized that there would be at least moderate agreement, regardless of the observers' background. METHODS An international group of independent observers evaluated axial, coronal, and sagittal computed tomography images on an online platform from 11 rib fractures for location (anterior, lateral, or posterior), type (simple, wedge, or complex), and displacement (undisplaced, offset, or displaced) of rib fractures. The multirater κ and Gwet's first agreement coefficient (AC1) were calculated to estimate agreement among the observers. RESULTS A total of 90 observers participated, with 76 complete responses (84%). Strong agreement was found for the classification of fracture location (κ = 0.83 [95% confidence interval (CI) 0.69-0.97]; AC1, 0.84 [95% CI, 0.81-0.88]), moderate for fracture type (κ = 0.46 [95% CI, 0.32-0.59]; AC1, 0.50 [95% CI, 0.45-0.55]), and fair for rib fracture displacement (κ = 0.38 [95% CI, 0.21-0.54], AC1, 0.38 [95% CI, 0.34-0.42]). CONCLUSION Agreement on rib fracture location was strong and moderate for fracture type. Agreement on displacement was lower than expected. Evaluating strategies such as comprehensive education, additional imaging techniques, or further specification of the definitions will be needed to increase agreement on the classification of rib fracture type and displacement as defined by the Chest Wall Injury Society taxonomy. LEVEL OF EVIDENCE Diagnostic Test or Criteria; Level IV.