Non-Invasive Prediction of Fracture Risk Due to Benign and Metastatic Skeletal Defects

Abstract

ABSTRACTThe skeleton is the third most common site of metastatic cancer and a third to half of all cancer cases metastasize to bone. While much has been learned about the mechanisms of metastatic spread of cancer to bone, little headway has been made in establishing reliable guidelines for estimating fracture risk associated with skeletal metastases.Our hypothesis is that a change in bone structural properties as a result of tumor-induced osteolysis determines the fracture risk in patients with skeletal metastases. Our goal was to develop an image based clinical tool to monitor the fracture risk associated with individual lesions in patients with skeletal metastases and to use this tool to optimize treatment and to monitor a patient's response to treatment. If bone is considered a rigid porous foam undergoing remodeling by osteoblasts and osteoclasts in response to local and/or systemic modulators of their activity, it follows that changes in bone material properties reflect the net effect of this remodeling activity. Therefore, image based methods that measure both bone mineral density and whole bone geometry can be used to monitor the response of skeletal metastases to anticancer treatment and to predict whether a specific lesion has weakened the bone sufficiently such that pathological fracture is imminent. In a series of laboratory experiments we demonstrated that the reduction in the load carrying capacity of a bone with simulated skeletal defects could be predicted accurately and non-invasively using computed tomography (CT). We also demonstrated that bone material properties from tissue excised from normal, osteoporotic and metastatic cancer bone specimens could be modeled analytically using a bivariate function of bone tissue density and bone volume fraction (Vvb). Since these bone specimens were inhomogeneous with respect to density distribution (as is the case for pathologic bonein-situ), the sub-region with the minimum-Vvbaccounted for more of the variability in the measured mechanical properties compared to the average Vvbfor the entire specimen. Therefore the “weakest” subregion governed most of the mechanical behavior of the pathologic bone specimens. We applied our methods for predicting fracture risk to analyze bones from children with benign bone defects and showed that our relatively simple methods were much better at predicting fracture (100% sensitive, 94% specific) than current radiographic based guidelines (66% accurate). Using CT based data to calculate the load bearing capacity of vertebrae infiltrated with metastatic breast carcinoma, we also predicted with 100% sensitivity and 69% specificity the occurrence of a new vertebral fracture in women with metastatic breast cancer. These results are in contrast to the best available fracture risk criteria based on the size and location of the lesion on CT images of the spine, which were only 22% specific. Our non-invasive, image based method that measures both bone mineral density and whole bone geometry may also be used to monitor the response of skeletal metastases to anticancer treatment and to predict whether a specific lesion has weakened the bone sufficiently such that pathological fracture is imminent.