Comparison of Acetabular Measurements Between 2 Validated Software Programs Used in Hip Preservation Surgery

Abstract

Background: Validated software tools (Clinical Graphics [CG] and Hip2Norm) permit measurement of the percentage of femoral head coverage (%FHC), which aids in morphological classification and prediction of outcome after hip preservation surgery. Purpose: (1) To assess whether acetabular parameter measurements determined from 2 commonly used software systems are comparable. (2) To determine which parameters influence the correlation or differences between software outputs and measurements. Study Design: Cohort study (diagnosis); Level of evidence, 2. Methods: The study included 69 patients (90 hips) who underwent periacetabular osteotomy and had comprehensive preoperative imaging available. Lateral center-edge angle (LCEA), acetabular index (AI), and %FHC were determined using 3-dimensional computed tomography (CT) measurements by CG and Hip2Norm software. Images of 18 pelvises were segmented to determine spinopelvic parameters and subtended acetabular angles. Between-group measurements were compared using correlation coefficients and Bland-Altman analyses. The difference in the outputs of the 2 programs was defined as delta (Δ). Radiographic parameters were tested to assess whether they were responsible for differences in %FHC between software programs. Results: Strong correlations between LCEA (ρ = 0.862) and AI (ρ = 0.825) measurements were seen between the Hip2Norm and CG programs. However, weak correlation was seen in the estimate of %FHC (ρ = 0.358), with the presence of a systematic error. Hip2Norm consistently produced lower anterior, posterior, and total %FHC values than CG. The %FHC determined by CG, but not Hip2Norm, correlated with acetabular subtended angles (P < .05). Pelvic tilt measured on CT did not correlate with pelvic tilt estimated by Hip2Norm (P = .56), and ΔPelvicTilt strongly correlated with the difference in %FHC by the 2 software programs (ρ = 0.63; P = .005), pelvic incidence (ρ = 0.73; P < .001), and pelvic tilt (ρ = −0.91; P < .001) as per CT. Conclusion: The correlation of %FHC between Hip2Norm and CG was weak (ρ = 0.358). The difference in measurements of %FHC correlated with ΔPelvicTilt. The %FHC determined by CG strongly correlated with the segmented acetabular subtended angles and thus more likely reflected true values. Hip preservation surgeons should be aware of these measurement differences because %FHC is important in the diagnosis and prognosis of acetabular dysplasia.