Reliability of Overcoverage Parameters With Varying Morphologic Pincer Features: Comparison of EOS® and Radiography

Abstract

Multiple radiographic parameters used for diagnosis and quantification of morphologic pincer features have emerged, but the degree to which pelvic tilt or rotation affects conventional radiography and EOS(®) is unknown. We asked: (1) What is the reliability of EOS(®) and conventional radiography at increasing sizes of morphologic pincer features with varying degrees of tilt and rotation? (2) What is the effect of tilt and rotation on acetabular overcoverage measurements? Using a dry cadaveric pelvis, AP conventional radiographs and EOS(®) images were taken at intervals of increasing modeled pincer size with 0° to 15° varying tilt and rotation. Lateral center-edge angle, Sharp angle, Tönnis angle, crossover sign, and retroversion index were measured on all images. Statistical analysis was conducted. The intermodality intraclass correlation coefficients for conventional radiography and EOS(®) radiography across all pincer sizes, rotations, and tilts were excellent (0.93-0.98). Crossover sign was in perfect agreement in conventional radiography and EOS(®). Rotation of the hip away from the beam source and/or increased anterior tilt falsely increased all overcoverage parameters except for Tönnis angle. Rotation away from the beam of 10° or greater or anterior tilt of 5° or greater produced a false-positive crossover sign. EOS(®) radiography maintained excellent reliability in comparison to conventional radiography but both were equally vulnerable to the effects of tilt and rotation for quantification of hip parameters used in acetabular overcoverage assessment. A standardized pelvic radiograph ensuring that the pelvis is not excessively tilted or rotated should be used for assessing acetabular overcoverage parameters.