A Risk-Adapted Approach Using US Features and FNA Results in the Management of Thyroid Incidentalomas Identified by F-18-FDG PET

Abstract

Purpose: To assess the risk of malignancy of thyroid incidentalomas found on F-18-FDG PET/CT by US features and cytologic results, and to evaluate the clinical usage of a combination of US features and cytology for post-FNA management of thyroid incidentalomas on F-18-FDG PET/CT. Materials and Methods: From September 2006 to December 2008, 132 patients with 134 thyroid incidentalomas detected on F-18-FDG PET/CT who had undergone US and US-FNA were included in this study. We evaluated the malignancy rate of thyroid incidentalomas in different subgroups subdivided by US features and US-FNA cytology results. Several variables were compared between the benign and malignant group. Results: The risk of malignancy was 58.2% (78/132) in thyroid incidentalomas on F-18-FDG PET/CT. Age, gender, and tumor size were not significantly different between the malignant and benign group. Malignancy rate of thyroid incidentalomas was significantly higher in the suspicious malignant (88.9 %) than in the probably benign group (11.3 %) on US (p < 0.001). Malignancy rates were high in thyroid nodules with "malignancy", "suspicious for malignancy", or "follicular neoplasm" on cytologic results, regardless of US features. However, malignancy rates of thyroid incidentalomas with "unsatisfactory" or "benign" results on cytology were higher in the suspicious malignant (75 %, 12.5 %, respectively) than in the probably benign (0 %) group on US. Conclusions: This study demonstrated that the risk of malignancy was high in thyroid incidentalomas on F-18-FDG PET/CT even without suspicious US features. However, there was no malignancy in nodules with no suspicious US features and benign cytology. Based on these results, we concluded that US may not replace FNA in the diagnosis of PET incidentalomas, and that a follow-up may be considered of thyroid incidentalomas with benign cytology and no suspicious US features.