Aims: Hip dysplasia (HD) leads to premature osteoarthritis. Timely detection and correction of HD has been shown to improve pain, functional status, and hip longevity. Several time-consuming radiological measurements are currently used to confirm HD. An artificial intelligence (AI) software named HIPPO automatically locates anatomical landmarks on anteroposterior pelvis radiographs and performs the needed measurements. The primary aim of this study was to assess the reliability of this tool as compared to multi-reader evaluation in clinically proven cases of adult HD. The secondary aims were to assess the time savings achieved and evaluate inter-reader assessment. Methods: A consecutive preoperative sample of 130 HD patients (256 hips) was used. This cohort included 82.3% females (n = 107) and 17.7% males (n = 23) with median patient age of 28.6 years (interquartile range (IQR) 22.5 to 37.2). Three trained readers’ measurements were compared to AI outputs of lateral centre-edge angle (LCEA), caput-collum-diaphyseal (CCD) angle, pelvic obliquity, Tönnis angle, Sharp’s angle, and femoral head coverage. Intraclass correlation coefficients (ICC) and Bland-Altman analyses were obtained. Results: Among 256 hips with AI outputs, all six hip AI measurements were successfully obtained. The AI-reader correlations were generally good (ICC 0.60 to 0.74) to excellent (ICC > 0.75). There was lower agreement for CCD angle measurement. Most widely used measurements for HD diagnosis (LCEA and Tönnis angle) demonstrated good to excellent inter-method reliability (ICC 0.71 to 0.86 and 0.82 to 0.90, respectively). The median reading time for the three readers and AI was 212 (IQR 197 to 230), 131 (IQR 126 to 147), 734 (IQR 690 to 786), and 41 (IQR 38 to 44) seconds, respectively. Conclusion: This study showed that AI-based software demonstrated reliable radiological assessment of patients with HD with significant interpretation-related time savings. Cite this article: Bone Jt Open 2022;3(11):877–884.

BACKGROUND Hip dysplasia (HD) is characterized by insufficient acetabular coverage of the femoral head, leading to a predisposition for osteoarthritis. While radiographic measurements such as the lateral center edge angle (LCEA) and Tönnis angle are essential in evaluating HD severity, patient-reported outcome measures (PROMs) offer insights into the subjective health impact on patients. AIM To investigate the correlations between machine-learning automated and manual radiographic measurements of HD and PROMs with the hypothesis that artificial intelligence (AI)-generated HD measurements indicating less severe dysplasia correlate with better PROMs. METHODS Retrospective study evaluating 256 hips from 130 HD patients from a hip preservation clinic database. Manual and AI-derived radiographic measurements were collected and PROMs such as the Harris hip score (HHS), international hip outcome tool (iHOT-12), short form (SF) 12 (SF-12), and Visual Analogue Scale of the European Quality of Life Group survey were correlated using Spearman's rank-order correlation. RESULTS The median patient age was 28.6 years (range 15.7-62.3 years) with 82.3% of patients being women and 17.7% being men. The median interpretation time for manual readers and AI ranged between 4-12 minutes per patient and 31 seconds, respectively. Manual measurements exhibited weak correlations with HHS, including LCEA (r = 0.18) and Tönnis angle (r = -0.24). AI-derived metrics showed similar weak correlations, with the most significant being Caput-Collum-Diaphyseal (CCD) with iHOT-12 at r = -0.25 (P = 0.042) and CCD with SF-12 at r = 0.25 (P = 0.048). Other measured correlations were not significant (P > 0.05). CONCLUSION This study suggests AI can aid in HD assessment, but weak PROM correlations highlight their continued importance in predicting subjective health and outcomes, complementing AI-derived measurements in HD management.