Diagnosis of scaphoid fracture displacement with radiography and computed tomography

Background: Displacement is an important risk factor for nonunion of scaphoid wrist fractures. We compared com-puted tomography with radiographs with regard to their ability to detect displacement. Methods: Six blinded observers rated thirty scaphoid fractures (ten displaced and twenty nondisplaced) with use of radiographs and computed tomography. The radiographs were evaluated separately from the computed tomography scans and then, in a third evaluation, the two imaging studies were reviewed simultaneously. The evaluations were re-peated four weeks later. Observers were asked to evaluate specific measures of fracture displacement and then to judge the fracture as being displaced or nondisplaced. Results: Intraobserver reliability was better for computed tomography alone and the combination of radiographs and computed tomography than it was for radiographs alone (kappa values, 0.65, 0.63, and 0.54, respectively; all p < 0.001). The interobserver reliability was also better for computed tomography alone and the combination of radio-graphs and computed tomography than it was for radiographs alone (kappa values, 0.43, 0.48, and 0.27, respec-tively; all p < 0.001). The average sensitivity was 75 % for radiographs alone, 72 % for computed tomography alone, and 80 % for both; the average specificity was 64%, 80%, and 73%, respectively; the average accuracy was 68%, 77%, and 75%, respectively. The positive predictive values (assuming a 5 % prevalence of fracture displacement) were low (0.10, 0.13, and 0.16) and the negative predictive values were high (0.97, 0.98, and 0.99) for the radiographs, com