ARRS 2022 Abstracts


1962. Validated Radiographic Measurements for Hip Dysplasia Can Be Used During 3-D Ct Evaluation for Pre-surgical Planning
Authors * Denotes Presenting Author
  1. Bayan Mogharrabi *; University of Texas at Southwestern Medical Center
  2. Jason Lin; University of Texas at Southwestern Medical Center
  3. Yin Xi; University of Texas at Southwestern Medical Center
  4. Joel Wells; University of Texas at Southwestern Medical Center
  5. Avneesh Chhabra; University of Texas at Southwestern Medical Center
This study aimes to evaluate and compare the commonly used angular measurements for hip dysplasia(HD) on 3DCT and plain radiographs. We hypothesize that the mean differences of angular measurements from the two modalities are within the standard errors of the means.

Materials and Methods:
A retrospective cross-sectional analysis of 171 patients with a confirmed diagnosis of hip dysplasia were prospectively recruited and informed consent was obtained. All patients received full series x-rays consisting of anteroposterior, false profile, bilateral 45-degree Dunn and frog lateral views as well as 3D CT imaging. 3DCT scans were uniformly acquired on two multidetector CT scanners in 0.65mm and 3mm slice reconstruction in all three planes were available. Two trained readers recorded the anterior center edge angle (ACEA), lateral center edge angle(LCEA), and Tönnis angle on x-ray. LCEA and Tönnis angles were recorded on CT at three different depths (anterior, central, and posterior) with respect to the acetabulum on axial view. Correlation analysis using Spearmen partial correlation coefficient was used to evaluate the relationship of the angles between radiographs and CT, and a multi-variable rater agreement calculator assessed the inter-reader accuracy of the measurements.

Bland-Altman plots between radiography and CT for the LCEA and Tönnis angles yielded a mean standard difference ranging from -2.19 - 2.56. LCEA had the smallest difference at the median depth on the left hip and the anterior depth on the right hip. ACEA had a mean standard difference of -26.9 on the left hip and -23.25 on the right. Tönnis angle mean standard difference was least on the posterior depth for both hips ( -1.83 on left and -1.74 on the right). All Spearman partial correlation coefficients revealed statistically significant relationships between radiograph and CT angles. Correlation coefficients between LCEA and Tönnis angles had p-values < 0.0001, while right ACEA was p < 0.05, and left ACEA was p < .0001. Agreement between the readers was moderate to good for 20/24 measurements (ICC of 0.4 - 0.8).

LCEA and Tönnis angle measurements on radiographs demonstrate statistically significant correlation with the same angle on 3DCT at posterior, anterior, and central depths. ACEA demonstrates a statistically significant correlation between 3DCT and x-ray. Posterior depth for the left hip and anterior depth for the right hip are the optimal sections on CT to evaluate LCEA, and the posterior depth is more suited for Tönnis angle assessment. Using the diagnostic advantage of each depth on a 3D-CT to generate simulated radiographs, guides us towards a future with reduced pre-surgical radiation dosing while maintaining a high level of diagnostic accuracy.