Abstract
PURPOSE: To determine the malignancy rate of bone lesions identified on FDG PET/CT in patients who have undergone CT-guided biopsy because of the suspicion of malignancy. METHODS: This single-centre retrospective study spanned eight consecutive years and included all patients who underwent both FDG PET/CT and CT-guided bone biopsy because of the suspicion of malignancy. The positive predictive value (PPV) for malignancy was calculated, and different patient and imaging characteristics were compared between malignant and benign bone lesions. RESULTS: Of 102 included patients with bone lesions that all showed FDG uptake exceeding mediastinal uptake, bone biopsy showed a malignant lesion in 91 patients, yielding a PPV for malignancy of 89.2 % (95 % CI 81.7 - 93.9 %). In the 94 patients with bone lesions that showed FDG uptake exceeding liver uptake, bone biopsy showed a malignant lesion in 83 patients, yielding a PPV for malignancy of 88.3 % (95 % CI 80.1 - 93.5 %). Higher age, bone marrow replacement of the lesion seen on CT, expansion of the lesion seen on CT, and presence of multifocal lesions on FDG PET/CT were significantly more frequent in patients with malignant lesions than in those with benign bone lesions (P = 0.044, P = 0.009, P = 0.015, and P = 0.019, respectively). Furthermore, there was a trend towards a higher incidence of cortical destruction (P = 0.056) and surrounding soft tissue mass (P = 0.063) in patients with malignant bone lesions. CONCLUSION: The PPV for malignancy of suspicious bone lesions identified on FDG PET/CT is not sufficiently high to justify changes in patient management without histopathological confirmation. Nevertheless, ancillary patient and imaging characteristics may increase the likelihood of a malignant bone lesion.