Assessment of malignant skeletal disease: initial experience with 18F-fluoride PET/CT and comparison between 18F-fluoride PET and 18F-fluoride PET/CT.

UNLABELLED 18F-fluoride PET/CT was performed on 44 oncologic patients to evaluate its diagnostic accuracy in assessing malignant osseous involvement and in differentiating malignant from benign bone lesions. METHODS (18)F-fluoride PET and (18)F-fluoride PET/CT were interpreted separately. Lesions showing increased (18)F-fluoride uptake were categorized as malignant, benign, or inconclusive. The final diagnosis of lesions was based on histopathology, correlation with contemporaneous diagnostic CT or MRI, or clinical follow-up of at least 6 mo (mean, 10 +/- 3 mo). RESULTS Increased (18)F-fluoride uptake was detected at 212 sites, including 111 malignant lesions, 89 benign lesions, and 12 lesions for which the final diagnosis could not be determined. In a lesion-based analysis, the sensitivity of PET alone in differentiating benign from malignant bone lesions was 72% when inconclusive lesions were considered false negative and 90% when inconclusive lesions were considered true positive. On PET/CT, 94 of 111 (85%) metastases presented as sites of increased uptake with corresponding lytic or sclerotic changes, and 16 of the 17 remaining metastases showed normal-appearing bone on CT, for an overall sensitivity of 99% for tumor detection. For only 1 metastasis was PET/CT misleading, suggesting the false diagnosis of a benign lesion. The specificity of PET/CT was significantly higher than that of PET alone (97% vs. 72%, P < 0.001). PET/CT identified benign abnormalities at the location exactly corresponding to the scintigraphic increased uptake for 85 of 89 (96%) benign lesions. In a patient-based analysis, the sensitivity of PET and PET/CT was 88% and 100%, respectively (P < 0.05) and the specificity was 56% and 88%, respectively (not statistically significant). Among the 12 patients referred for (18)F-fluoride assessment because of bone pain despite negative findings on (99m)Tc-methylene diphosphonate bone scintigraphy, (18)F-fluoride PET/CT suggested malignant bone involvement in all 4 patients with proven skeletal metastases, a potential benign cause in 4 of 7 patients who had no evidence of metastatic disease, and a soft-tissue tumor mass invading a sacral foramen in 1 patient. CONCLUSION The results indicate that (18)F-fluoride PET/CT is both sensitive and specific for the detection of lytic and sclerotic malignant lesions. It accurately differentiated malignant from benign bone lesions and possibly assisted in identifying a potential cause for bone pain in oncologic patients. For most lesions, the anatomic data provided by the low-dose CT of the PET/CT study obviates the performance of full-dose diagnostic CT for correlation purposes.