Assists clinicians with analysing Hip Dysplasia and impingement on X-rays
Analysing Hip Dysplasia and Impingement on X-rays
X-ray
Measures and Analysis:Acetabular IndexLCE AngleAlpha AngleMiminium demoracetabular joint sace in mmObturator foramen index
The study analysed 78 pelvic radiographs of adults (average age 50.1 years, 46% female), focusing on hip measurements for conditions like hip dysplasia.Agreement between RbHip and human readers showed a strong correlation. Bland-Altman analyses provided a mean measured difference of 0.37° to 9.56° for the lateral center edge angle (LCEA) and 0.35° to 2.06° for the acetabular index angle (AIA), depending on the reader's experience. High consistency was observed, particularly in LCEA measurements for dysplasia assessment.
Bland-Altman plots were used to assess bias, showing that RbHip results closely aligned with human measurements, especially with the senior orthopedic surgeon. Bias estimates for LCEA ranged from 3.56° (95% CI: 2.41 to 4.74) to 10.01° (95% CI: 8.37 to 11.82), depending on the reader. Confidence intervals for measurement variances were generally narrow, indicating reliable algorithm performance.
RbHip algorithm completed measurements within seconds, offering significant time savings compared to manual evaluations by radiologists, which can take several minutes per radiograph.
The automated nature of RbHip allows clinicians to quickly access accurate measurements, streamlining workflows in radiology departments, especially for routine assessments in high-throughput settings.
By automating hip measurements, RbHip reduces the need for additional follow-up assessments, which can minimize costs associated with misdiagnosis or delayed diagnoses. Automated measurements provide a consistent approach to identifying hip dysplasia early, potentially reducing healthcare costs associated with later-stage interventions.The algorithm aids non-specialist readers in making reliable assessments, freeing radiologists to focus on more complex cases. This efficiency is valuable in resource-limited settings or high-volume centers.
