To assess the predictive value of CT value and cross-sectional area (CSA) of the psoas muscle at the L3 level for imminent fracture risk after vertebral fracture and compare with quantitative CT based finite element analysis (QCT-FEA) and volumetric bone mineral density (vBMD) measurement. Forty patients with imminent fractures following prior thoracolumbar fragility fractures and fifty-one matched controls without re-fracture during at least a two-year follow-up were retrospectively identified from administrative databases. The CT value and the CSA of the psoas muscle at L3 levels were measured with baseline CT images. Additionally, QCT-FEA was performed to estimate the vertebral bone strength of L1 and measured the vBMD. The predictive values of these factors were assessed using logistic regression analysis and ROC curves. Patients with imminent fractures had significantly lower psoas muscle CSA (669.5 ± 188.1 vs. 904.2 ± 189.7 mm², P < 0.001), vBMD (65.6 ± 19.6 vs. 78.5 ± 16.0 mg/cm³, P = 0.001), and FE-vertebral strength (564.5 ± 329.9 vs. 783.9 ± 225.6 N, P < 0.001) than controls, while no difference was observed in psoas muscle mean CT value ( P = 0.210). Psoas CSA emerged as the strongest independent predictor of imminent fracture (adjusted OR = 5.0, 95%CI: 2.5–11.3, P < 0.001), outperforming vBMD (adjusted OR = 2.2, 95%CI: 1.4–3.8) and FE-vertebral strength (adjusted OR = 2.7, 95%CI: 1.6–4.9). ROC analysis confirmed psoas CSA had the highest AUC among single parameters (0.812, 95%CI: 0.724–0.901), and the combination of psoas CSA and FE-vertebral strength achieved the optimal predictive performance (AUC = 0.866). Stratified analysis in females validated these findings, with psoas CSA showing an even stronger association (adjusted OR = 8.6, 95%CI: 3.4–29.4) and higher AUC (0.850). L3 psoas muscle CSA is an independent and superior predictor of imminent fracture risk in patients with prior vertebral fractures compared to QCT-FEA-derived vertebral strength and vBMD. Combining psoas CSA with FE-vertebral strength further enhances predictive efficiency. Retrospectively registered.
Jiang et al. (Sat,) studied this question.