Quantitative Evaluation of Angular Measurements on Plain Radiographs in Patients With Slipped Capital Femoral Epiphysis: A 3-Dimensional Analysis of Computed Tomography-Based Computer Models of 46 Femora

Background: In cases of slipped capital femoral epiphyses (SCFE) findings on plain radiographs help to determine the further necessary course of action. In severe cases possible surgical procedures are commonly indicated and planned using angular measurements on plain radiographs to describe the extent and direction of the slip. The aim of this study was to quantify the amount of angular errors deriving from this method. Methods: Data and imaging of 23 consecutive patients with SCFE (31 affected and 15 unaffected femora) were included in this study. We determined shaft-neck/shaft-physis angles on antero-posterior and torsional angles on lateral radiographs in a clinical setting. As a reference we enabled similar angular measurements on CT-based three-dimensional computer models of the same femora bearing no projectional errors and malpositioning problems. Results: In average, shaft-neck- and shaft-physis-angles were overestimated (6.5° and 10.1°) on plain radiographs and neck torsion underestimated (−15.7°). In general the variability was high, especially for neck and physeal torsional measurements with standard deviations of +/−11.8° and +/−16.7°. Three out of four torsional measurements on affected femora were outside a +/−10° window of error, about every third outside a +/−20° window. Conclusion: Our results suggest to be careful when using plain radiographs as a source to determine the slippage extent in SCFE. Before using a plain radiograph to reject or indicate and plan a correction osteotomy in an individual case of SCFE the surgeon should reassure that radiographic method and patient positioning provide a reproducible and accurate depiction of the femoral geometry. Level of Evidence: Level II; 23 consecutive patients with SCFE in the senior authors practice; evaluation of the reliability of angular measurements on plain radiographs; CT based 3D computer models of the same femora as a reference.