Open‐Source and Commercial 3‐Dimensional Computed Tomography Methods Show Limited Agreement for Classifying Acetabular Morphology in Patients Undergoing Hip Arthroscopy
Ashley A. Ellingwood, Rajvarun S. Grewal, Henry M. Gass, Alisha Chen, Daniel Woo, Byron V. L. Beasley, James D. Bomar, Kevin C. Parvaresh, Vidyadhar V. Upasani, Matthew R. SchmitzPurpose
To compare acetabular morphology classifications derived from a validated open‐source (freely available) method with those derived from commercially available software in patients undergoing hip arthroscopy for femoroacetabular impingement syndrome.
Methods
We analyzed preoperative 3‐dimensional computed tomography imaging in patients with femoroacetabular impingement syndrome using open‐source and commercial software. The open‐source method calculated acetabular coverage across 5 regions and generated z ‐scores relative to an internal normative database. Hips with z ‐scores ≥2 or ≤−2 were classified as abnormal. The commercial method categorized hips as normal, undercovered, or overcovered based on published thresholds (lateral center‐edge angle, Tönnis angle, coverage percentage, and crossover sign) and focal depth mapping. The primary outcome was categorical agreement between methods.
Results
Seventy‐six hips (55 women, 21 men; mean age 35.5 ± 12.3 years) were included. The open‐source method classified 58/76 hips as normal, 10/76 as dysplastic, and 8/76 as overcovered. The commercial method classified 14/76 hips as normal, 29/76 as undercovered, and 33/76 as overcovered. A total of 23/76 hips were classified the same in both systems and called agreed. Disagreements (53/76 hips) fell into 3 categories: (1) essentially similar despite different labels because of borderline z ‐scores or small focal areas (n = 24), (2) disagreements driven by Tönnis angle alone (n = 8), and (3) true disagreements (n = 21). Notably, 5 hips (6.6%) showed opposite classifications (e.g., dysplastic vs overcovered). Subanalysis of z ‐scores showed that 69.8% of disagreements reflected borderline morphology ( z ‐scores of ±1‐2).
Conclusions
Although both open‐source and commercial methods provide 3‐dimensional classification of acetabular morphology, categorical agreement was limited. Most disagreements reflected borderline morphology or small focal areas of uncertain clinical relevance. A minority showed true classification opposites.
Clinical Relevance
Awareness of the methodological differences between open‐source and commercial 3‐dimensional computed tomography platforms may help surgeons appropriately interpret acetabular morphology classifications during preoperative planning for patients undergoing hip arthroscopy for femoroacetabular impingement syndrome.