Inter and intra-observer agreement of the 3-dimensional CT based anterior inferior iliac spine classification system shows fair-to-moderate agreement among high volume hip surgeons

Abstract

Purpose: The purpose of this multicenter, multinational study is to evaluate the agreement level of the Hetsroni's classification system across high-volume hip surgeons who specialize in hip preservation surgeries.

Methods: Four surgeons from three countries reviewed a digital survey that included 93 3D CT images of the hip from 53 patients. The population was composed of individuals who had undergone a pelvis CT scan in a tertiary hospital between 2000 and 2016. Each rater reviewed the images and classified each image according to AIIS subtype I, II, or III. After a minimum of two months, the raters repeated the survey. The inter-rater and intra-rater agreement was then assessed. The kappa values were calculated to determine variability.

Results: Inter-rater agreement levels yielded fair agreement for both sessions (Kappa = 0.4, p value < 0.001 in the first and Kappa = 0.27, p value < 0.001 in the second). Inter-rater agreement levels separating non-pathological Type I from pathological Types II and III yielded moderate to fair inter-rater agreement levels (K = 0.47, p value < 0.001 in the first session and k = 0.32, p value < 0.001 in the second). Intra-rater reliability displayed moderate agreement (average K = 0.53).

Conclusion: The current 3D CT-based AIIS classification system shows fair-to-moderate inter- and intra-rater agreement among high-volume hip surgeons. According to this study, the agreement of the Hetsroni classification system is not able to be sufficiently reproduced. Since accurate classification of the AIIS morphology is imperative in establishing proper treatment for SSI, this classification system there is therefore limited in its clinical value.

Level of evidence: III.

Keywords: 3D Reconstruction; AIIS morphology; Hetsroni classification; Subspine impingement.