2107. Correlation Between Machine Learning Generated Radiographic Hip Measurements and 3D MRI Findings Patients With for Hip Dysplasia
Authors * Denotes Presenting Author
  1. Shuda Xia *; UT Southwestern Medical School
  2. Avneesh Chhabra; UT Southwestern Medical School
  3. Gaurav Sharan; UT Southwestern Medical School
  4. Uma Thakur; UT Southwestern Medical School
  5. Holden Archer; UT Southwestern Medical School
  6. Yin Xi; UT Southwestern Medical School
  7. Joel Wells; UT Southwestern Medical School
Hip dysplasia (HD) is characterized by structural abnormalities of the acetabulum causing under coverage of the femoral head. HD usually presents with symptoms of hip pain and/or instability in both adult and pediatric populations. If left untreated, it can lead to premature osteoarthritis and internal soft tissue damage due to improper mechanics and stress overload at the joint. A wide range of radiographic measurements are used to assist in the initial diagnosis and assessment of HD, including lateral center edge angle (LCEA), Tonnis, extrusion index, and others. Recent advancements in artificial intelligence (AI) software have enabled reproducible and automated generation of radiographic measurements. MRI is obtained for assessing internal and external hip soft tissues, especially the labrum, hyaline cartilage, synovium, and regional muscles. Additionally, 3D MRI used in concordance with 2D MRI has been shown to enhance the evaluation of labrum and cartilage lesions of HD. 3D MRI opens opportunities to measure the extent of labral tears with labrum-specific reconstructions with a higher degree of accuracy, not possible with traditional 2D MRI. The aim of the study was to assess the correlations between MRI findings and AI-derived radiographic hip measurements with the hypothesis that longer labral tears and worsening cartilage damage seen on 3D MRI would correlate with worsening AI-generated radiographic HD measurements.

Materials and Methods:
A total of 156 hips from 139 consecutive patients diagnosed with HD (ages 16 - 68 years, both genders) were included. All patients had complete data sets including 2D and 3D MRI, 4 view x-rays, AI-generated radiographic measurements, and qualitative and quantitative analysis of labral and cartilage injuries. A 3D reconstruction of the labrum was generated for each hip on an independent software, and a multireader study was performed. Interreader analysis (i.e., ICC) and spearman correlations were calculated.

Among the 139 patients, 70% (97/139) were women and 30% (42/139) were men. The largest and most common labral tear measured was in the anterosuperior location (90%, 133/156) with most cases (59%, 93/156) showing only one labral tear, and 34% (53/156) of cases showing para-labral cysts. Although there were no statistically significant correlations (p > 0.05) between the length of labral tears and hip measurements, there were statistically significant, though weak, correlations, between the presence of para-labral cysts and femoral head coverage, extrusion index, LCEA, and Tonnis measurements.

This study presents a novel application of AI and 3D MRI in patients with HD. There were statistically significant, though weak correlations seen between various AI-generated radiographic measurements and the presence of para-labral cysts and the number of labral tears. This suggests a role for AI to predict findings seen on 3D MRI based on radiographic findings for patients with HD.