Computed tomography-derived structural analysis for the likelihood of pathologic fracture in canine antebrachial osteosarcoma

Determining the risk of pathologic fracture in dogs with a primary bone tumor would aid in case selection for in-situ treatment options. Prior research found strong relationships between in vitro strength of canine antebrachii with primary bone tumors and CT-derived metrics. This study assesses the prognosis for pathologic fracture in dogs with distal radial bone tumors using CT-derived structural analysis metrics. CT images of the antebrachium in dogs with aggressive osseous lesions of the radius were used to calculate structural rigidity and failure forces, including axial rigidity (AR), craniocaudal bending rigidity (BR), torsional rigidity (TR), and failure forces for a slightly-curved/asymmetric beam (Fs) or a curved beam (Fc). Metrics were compared with the clinical outcome of radial fracture. Eight of 19 dogs with CT-derived metrics developed a radial fracture. The prognostic potential of the metrics to discriminate fractured and nonfractured bones was analyzed using receiver operating characteristic curves (area under the curve), stepwise logistic regression, and classification regression (CART) analyses. Fc was the most sensitive and specific metric for prognosing fracture occurrence (AUC = 0.864). When dog body weight (BW) was included, all five metrics had AUC > 0.705. Fc was the best predictor of fracture using stepwise logistic regression and CART analysis, followed by BR. An indication of fracture probability can be determined by normalizing Fc or BR with dog BW or by using the logistic regression equation of either metric with dog BW. Results warrant further analysis of a larger cohort to evaluate fracture likelihood in dogs with antebrachial bone neoplasia.