Predicting pathologic fractures of long bones caused by metastatic disease continues to be a challenging clinical problem. We assessed the ability of noninvasive imaging and computational techniques to predict the strength of bones with osteolytic lesions. A murine model of induced tumor osteolysis to the distal femur was used as a model system resulting in a wide range of lesion sizes. Microcomputed tomography scans were obtained and specimen-specific, voxel-based, finite element analyses were performed and results were compared with direct measurement of biomechanical strength via axial compressive loading of the distal femur. Additional indirect surrogates of bone strength included dual-energy xray absorptiometry to determine bone mineral density, radiographic scoring, and computed tomography volume/mineral estimates. Predicted bone strength was weakest (r2 = 0.55) for the dual-energy xray absorptiometry measure and strongest (r2 = 0.91) for the direct computed tomography voxel-based, finite element analysis. The relative success of the voxel-based, finite element modeling approach to estimate bone strength in a murine osteolytic tumor model indicates this approach, with further development and validation, could serve as a way to nondestructively estimate bone strength in a clinical setting.