Poster 231: Three-Dimensional Statistical Shape Modeling of Hip Morphology is More Predictive of Femoroacetabular Impingement Syndrome Compared to Traditional Two-Dimensional Radiographic Metrics

OBJECTIVES: Femoroacetabular impingement syndrome (FAIS) is classically described as cam, pincer, or mixed morphology, leading to pathologic osseous contact between the femur and acetabulum (Beck JBJS). Many studies have analyzed femur shape variability and/or acetabulum shape variability in isolation, rather than utilizing meaningful articulation of a loaded joint. At present, there is limited evidence characterizing how morphologic variations of an articulated hip joint contribute to FAIS. The primary objective of this study was to assess morphologic variations of the articulated hip joint amongst patients with FAIS compared to an asymptomatic cohort using statistical shape modeling (SSM). The secondary objective was to determine which two-dimensional (2D) measures have the strongest correlation with three-dimensional (3D) parameters associated with FAIS. METHODS: This retrospective case-control study recruited 14 patients with unilateral or bilateral symptomatic FAIS and 24 asymptomatic controls with no known history of hip surgery, chronic hip- related pathology, or severe musculoskeletal injury to the lumbar spine, hip, or lower extremity. Participants underwent computed tomography (CT) scans of their bilateral pelvis and proximal femurs (0.5 x 0.5 x 1.25 mm) that were used to create preoperative subject-specific bone models of the proximal femur and hip joint. Participants also underwent static standing biplane radiographs that were used to characterize articulation of their loaded hip joint. These 3D hip positions were determined with sub-millimeter accuracy using a previously validated volumetric model-based tracking process that matched the subject-specific bone models to the biplane radiographs. Once co-registered, left-sided hip joints were mirrored for each patient, to allow for comparison between left and right hip joints, and uploaded to Shapeworks 6.3 for SSM analysis. Using principal component analysis (PCA), complex morphologic variations within the study population were distilled into more digestible categories known as modes, which represented simplified independent morphologic variations. Scores for each mode for each hip are derived from the PCA and represent the morphology of patient-specific hip joints as compared to the overall study cohort. Multivariate regression was performed to identify subject-specific mode scores and demographics that predicted FAI using Stata 16E software. Lastly, Pearson correlations were utilized to correlate 2D measurements of alpha angle, femoral version, and acetabular version to mode scores. RESULTS: Of the 14 symptomatic cases with known cam-type FAIS, 11 met inclusion/exclusion criteria, yielding 11 cases (22 hip joints). Of the 24 controls recruited, no patients were excluded, resulting in 48 additional hip joints. Therefore, the final cohort consisted of 35 patients, roughly one-third of which showed FAIS pathology and eventually underwent femoroplasty, and two-thirds had no history of pathology prior to recruitment. Table 1 details mode descriptions from a surgeon with specialized fellowship training in hip preservation and the amount of explained morphologic variation captured by each mode. Figure 1 illustrates a montage of images that represent variations across Mode 1. Only Mode 1 (p = 0.0006), Mode 6 (p = 0.031), and patient age (p = 0.002) were able to distinguish between cases and controls. However, when further specifying the laterality of surgical intervention for FAIS correction, an interaction was observed between Mode 1 and patient age. More specifically, hip joints with lower Mode 1 values were more likely to require surgery at a younger age; whereas higher Mode 1 values demonstrated no association between age and probability of undergoing surgical intervention (Figure 2). Mode 6 was not predictive of surgical intervention. When considering 2D measurements, femoral version was the only parameter associated with Mode 1 values (r = -.41, p = 0.004), while alpha angle was the only predictor of Mode 6 values (r = -.41, p = 0.001). Acetabular version was not associated with either. CONCLUSIONS: Diagnostic criteria for FAIS are heterogeneous and imprecise. Femoral neck rotation within the acetabulum in combination with anteroposterior placement of the greater trochanter relative to the femoral neck axis is the strongest predictor of symptomatic FAIS. These results highlight the limitations of using 2D static imaging to represent highly complex 3D articulating structures, especially femoral variations. Although less precise than 3D methods, 2D measurements of femoral version, rather than alpha angle, may serve as a more accurate risk assessment for FAIS. SSM more accurately predicts symptomatic FAIS requiring surgery compared to traditional 2D radiographic metrics. This technique may help to identify patients who may benefit from early intervention and facilitate joint preservation. Furthermore, SSM may help identify patients who have a morphology that may be better treated with open rotational osteotomy rather than hip arthroscopy surgery.