• ABSTRACT
    • We evaluated the reproducibility of the Brooker classification for heterotopic ossification (HO) and, based on the results and weaknesses observed, proposed a simplified system with addition of objective criteria. Six observers classified radiographs of 169 total hip arthroplasties, using the Brooker classification and a modified system consisting of i) absence of HO or islands measuring <1 cm in length, ii) islands >1 cm or spurs leaving at least 1 cm between femur and pelvis, and iii) spurs leaving <1 cm between opposing surfaces or bony ankylosis. Reproducibility was calculated using kappa statistics. For the Brooker classification, interobserver kappa averaged 0.43 (range, 0.74-0.18) (poor). Intraobserver kappa averaged 0.74 (fair). For the modified classification, interobserver kappa averaged 0.59 (range, 0.51-0.76) (fair). Intraobserver kappa averaged 0.78 (good). Interobserver differences were significant (P=.0085). Interobserver consistency to detect severe HO (Brooker 3 and 4, or grade C) improved from 52% to 76% with the modified system. The new classification showed adequate interobserver reproducibility, less variability, and improved consistency for classification of significant HO.