Prognostic Prediction of Femoral Head Collapse Using Necrotic Size and Lateral Extent on Computed Tomography Images in Osteonecrosis of the Femoral Head.

BACKGROUND: The accurate prediction of femoral head collapse is essential for managing osteonecrosis of the femoral head (ONFH). In this study, we aimed to evaluate the utility of combining the coronal vertical diameter ratio (CVDR), as a simple alternative for necrotic size, and the lateral necrotic extent ratio (LNER), as an indicator of lateral lesion extension for prognostic prediction. METHODS: We retrospectively analyzed 121 hips (87 patients) with ONFH followed up for at least one year. The CVDR and LNER were measured using mid-coronal computed tomography images. Receiver operating characteristic (ROC) curve analyses were used to determine optimal cutoff values for predicting femoral head collapse of ≥ 3 mm. The independent predictors of collapse were evaluated using multivariable Cox proportional hazards models. Survival rates with an endpoint of ≥ 3 mm collapse were assessed using Kaplan-Meier analysis. RESULTS: The optimal cutoff values for predicting collapse were 49% for CVDR and 90% for LNER. Multivariable analysis revealed that CVDR ≥ 49% (hazard ratio [HR], 5.45; P < 0.001) and LNER ≥ 90% (HR, 4.43; P < 0.001) were independent and significant predictors of femoral head collapse. Risk stratification showed that 100% (36 of 36 hips) of the patients satisfying both criteria progressed to collapse, compared to only 9.5% (four of 42 hips) of the patients satisfying neither criterion. The 5-year collapse-free survival rate was significantly lower in the high-risk group than in other groups (P < 0.01). CONCLUSION: The combination of 49% CVDR and 90% LNER cutoff values provides a more accurate prognostic prediction in patients who have ONFH and can serve as a valuable indicator for determining treatment strategies.