Evaluation of the efficiency of FDG PET/CT in detection and characterization of skeletal metastases

Abstract

Background and purpose Combined PET/CT using 18F-FDG is widely used in evaluation of various malignancies; in their initial staging and more efficiently in their follow up; hence, the importance of evaluation of its diagnostic role in the imaging of skeletal metastases. The purpose of this study is to evaluate precisely the efficiency of FDG PET/CT in detection and characterization of osseous metastatic lesions compared to isolated PET and CT in various malignancies. Patients and methods The study included 123 patients divided into seven groups of malignancies to whom PET/CT was done. In this study population, a detailed retrograde lesion based analysis was performed for a total of 1705 detected bone lesions on PET, CT and fused PET/CT images. Sensitivity, specificity, PPV and NPV of each modality were calculated. Semi-quantitative and ROC curve analysis of the lesions were performed to study the relationship between the lesion's SUV and its corresponding morphologic pattern on CT and to set a reliable SUVmax cut-off value that can predict the presence of malignant lesion. Results The calculated fused PET/CT sensitivities and specificities in various malignancies ranged from 95.2% to 99.6% and 75% to 100%, respectively. The combined PET/CT has significantly improved the low CT sensitivity (especially in lymphoma) as well as both CT and PET specificities. Our ROC analysis suggested using SUVmax of 3 as a cut off value for malignant osseous lesions. Conclusion Fused PET/CT was highly efficient in evaluation of skeletal metastases with superior performance in: detection of early bone marrow infiltration not apparent on CT, resolution of metabolic activity before definite signs of complete healing on CT, detection of missed sclerotic metastases on PET due to their relatively low metabolic activity, detection of intra and extra osseous recurrence and differentiation of benign from malignant bone lesions. © 2014 Production and hosting by Elsevier B.V.