Simulating activities of daily living with finite element analysis improves fracture prediction for patients with metastatic femoral lesions

Abstract

Predicting fracture risk for patients with metastatic femoral lesions remains an important clinical problem. Mirels' criterion remains the most formalized radiographic scoring system with good sensitivity (correctly identifying clinical fractures) but relatively poor specificity (correctly identify cases that do not fracture). A series of patients with metastatic femoral lesions had Computed Tomography (CT) scans, were followed prospectively for 4 months, and categorized into fracture (n=5), non-fracture (n=28), or stabilized (n=11) groups. CT based-Finite Element (FE) modeling was used to predict fracture for these cases using axial compression (AC), level walking (LW), and aggressive stair ascent (ASA) loading conditions. The FE predicted fracture force was greater for the non-fracture compared to the fracture group for all loading cases. The ability of the FE models to predict fracture cases (sensitivity) was similar for the groups (Mirels, AC, LW: 80%, ASA: 100%). The ability of the models to correctly predict the non-fracture cases (specificity) was improved for AC (71%) and LW (86%) loading conditions, when compared to Mirels specificity (43%), but poorer for the ASA (21%) conditions. The results suggest that FE models that assess fracture risk using LW conditions can improve fracture prediction over Mirels scoring in a clinical population.