Objective: The purpose of this study was to evaluate the risk of malignancy of thyroid incidentalomas detected on (18)F-FDG PET and the diagnostic accuracy of sonography for differentiating benign from malignant focal thyroid incidentalomas that were detected on FDG PET.
Materials and methods: Retrospective review was performed of a database of 87 focal thyroid lesions seen on FDG PET and sonography. Forty-two focal lesions were malignant. We compared the accuracy of the maximum standard uptake value (SUV) to differentiate benign from malignant thyroid lesions. We classified the thyroid nodules as probably benign or suspicious for malignancy by the sonographic features. Statistical analyses compared two subgroups by sonographic classifications between benign and malignant thyroid lesions.
Results: The maximum SUV of the malignant nodules was not significantly higher than that of benign lesions. Thirty-seven (75.5%) of 49 lesions with suspicious sonographic findings revealed malignancy on cytopathology, compared with five (13.2%) of 38 lesions that showed probably benign sonographic findings. These differences were statistically significant using a kappa test (kappa = 0.675, p = 0.001) and logistic regression (odds ratio = 26.2, p = 0.001).
Conclusion: The probability (48.3%) of malignancy of focal thyroid incidentalomas seen on FDG PET is high. The maximum SUV of thyroid cancer is not significantly higher than that of benign lesions. The probability (13.2%) of malignancy is much lower when the sonographic findings appear benign, as compared with a significantly higher probability (75.5%) of malignancy when the sonographic findings are suspicious for malignancy.