Paper 65: Novel CT-Derived 3D Modeling Assessing Dysplastic Hips Versus Controls Demonstrates Decreased 2D Lateral Center Edge Angle, but not Tönnis Angle, Predicts Decreased 3D Anterolateral Coverage

OBJECTIVES: Acetabular coverage has significant implications in both acetabular dysplasia and femoroacetabular impingement syndrome (FAIS), with both undercoverage (dysplasia) and overcoverage (pincer-type FAIS) contributing to symptoms, labral tears, and degenerative joint changes. Measurements of acetabular coverage inferred from 2D imaging have been shown to be unreliable, and they provide a limited view of complex 3D morphology. Consequently, there is a need 1) for a consistent technique to allow for accurate, repeatable measurement of acetabular coverage and 2) to determine whether and to what degree 2D radiographic measures of acetabular coverage are related to 3D measures. Thus, the objectives of this study were to utilize novel 3D modeling techniques to 1) compare 3D acetabular coverage between patients undergoing HA followed by periacetabular osteotomy (PAO) for combined dysplasia and FAIS to patients undergoing HA alone for FAIS and 2) evaluate the association between 2D radiographic measures of acetabular coverage—lateral center edge angle (LCEA) and Tönnis angle (TA, acetabular index)—and acetabular coverage as measured on 3D osseous models. METHODS: An IRB-approved retrospective review of all patients who underwent HA+PAO from a large orthopedic practice from 2017-2021 was conducted. Included patients had a preoperative anteroposterior (AP) pelvic radiograph and CT scan of the full pelvis and bilateral proximal femurs. HA+PAO patients were matched 1:1 by sex, age, and body mass index (BMI) to patients who underwent HA alone. LCEA and TA were measured on AP pelvic radiographs. Dysplasia subgroups were defined as severe (LCEA < 15°), moderate (LCEA 15°-19.9°), mild/borderline (LCEA 20°-24.9°), and normal/non- dysplastic (LCEA 25°-39.9°). 3D osseous models of the femur and acetabulum were reconstructed from CT scans (Materialise Mimics v24.0) and uploaded into 3D modeling software (Materialise 3-matic v16.0). Acetabular coverage of the femoral head was measured by projecting the acetabular surface onto the femoral head surface as delineated by the acetabular rim profile. A coronal, transverse, and sagittal plane were made. The coronal plane was established by first defining a reference plane using points on both anterior superior iliac spines and the pubic tubercle on the affected side. A parallel plane was then created that passed through the affected femoral head center (FHC). The transverse plane was created by making a true horizontal plane passing through the FHC. Finally, the sagittal plane was established as a plane perpendicular to both the coronal and transverse planes and passing through the FHC. Total coverage was evaluated as the percentage coverage of the superior half of the femoral head (above the transverse plane), which was further divided by the coronal and sagittal planes into quadrants: anteromedial (AM), posteromedial (PM), anterolateral (AL), and posterolateral (PL). Percentage coverage of these quadrants was also quantified. 3D acetabular coverage, demographic characteristics, and 2D radiographic measurements were compared between the HA+PAO and HA groups using independent samples t-tests. Correlation and linear regression analyses were performed among all 52 patients. Correlations between 2D LCEA and TA versus 3D acetabular coverage were evaluated using Pearson correlation coefficients. Linear regression was performed evaluating the relationship between 2D LCEA and TA and 3D acetabular coverage. Significance was established using an a priori α of 0.05. RESULTS: Twenty-six dysplastic patients (9 severe, 12 moderate, 5 borderline) who underwent HA+PAO were matched to 26 non-dysplastic controls who underwent HA alone. There were no significant differences in sex, age, or BMI between the groups (p ≥ 0.9). Preoperative 2D LCEA and TA were significantly lower in the HA+PAO group (p < 0.001). 3D acetabular coverage was significantly decreased in the HA+PAO group globally and in the AM, AL, and PL quadrants (p < 0.001). The largest difference in mean coverage between the HA+PAO and HA groups was seen in the AL quadrant (-12.8%), followed by the PL quadrant (-11.3%, Figure 1, Figure 2). There was no significant difference in 3D acetabular coverage in the posteromedial quadrant (p = 0.117, Table 1, Figure 2). There were significant correlations of fair to very high strength between both 2D LCEA and TA and 3D coverage globally and in all quadrants (p < 0.001). The strongest correlations by quadrant were between 2D LCEA and 3D coverage anterolaterally (r = 0.835) followed by posterolaterally (r = 0.716). Upon linear regression analysis, 2D LCEA was a significant predictor of 3D coverage globally and in all quadrants (p ≤ 0.025), whereas 2D TA was not (p ≥ 0.4). CONCLUSIONS: This study utilized novel 3D modeling techniques to demonstrate that patients with acetabular dysplasia as defined by 2D LCEA have decreased 3D acetabular coverage globally and in each quadrant except the posteromedial quadrant. The decrease in coverage is most pronounced in the AL quadrant, followed by the PL quadrant. The study also demonstrated that both 2D LCEA and TA correlate significantly with 3D acetabular coverage; however, only LCEA is a significant predictor of 3D acetabular coverage globally and in all quadrants. The correlation between 2D LCEA and 3D coverage is strongest in the AL followed by the PL quadrant. These findings suggest that only 2D LCEA, not TA, is a useful radiographic measure of 3D coverage, and it is most useful for predicting coverage laterally, particularly anterolaterally. The further use of 3D modeling to uncover significant anatomical differences in multiple dimensions that are not apparent with 2D radiographic measures alone may have important impacts on patient selection, treatment modalities, surgical planning, and postoperative assessment.