Objective: The aim of the present study was to evaluate the prevalence of thyroid lesions detected by 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) incidentally, determine malignancy risk and its relationship with maximal standardized uptake value (SUVmax) and FDG uptake pattern. Methods: Between February 2009 and February 2014, a total of 12713 patients underwent 18F-FDG PET/CT. Incidental thyroid uptake was seen in 710 patients and further diagnostic evaluation was performed on 147 patients with focal or diffuse FDG uptake. The 18F-FDG PET/CT findings of these patients and their association with malignancy were retrospectively reviewed. Results: The prevalence of thyroid incidentalomas detected by 18F-FDG PET/CT was 5.6% (710/12713). Of the 147 patients who underwent biopsy or thyroid surgery, histology was benign in 99 and malign in 48 patients. The malignancy risk of incidental thyroid lesions was calculated as 32.7% (48/147). The median SUVmax was 2.9 (0.6–27.4) in benign group, whereas 11.8 (2.4–72.9) in malign group and the difference between these groups was statistically significant (p < 0.001). According to the ROC analysis, a SUVmax above 6 was more likely to be malign with statistical significance (p < 0.001). The sensitivity, specificity, positive predictive value and negative predictive value were 87.4%, 81.7%, 70.1% and 93% respectively. Conclusion: The malignancy risk of incidental thyroid lesions on 18F-FDG PET/CT is high. Although it is obvious that higher SUVmax values are tended to be malign, an overlap between benign and malign groups is still remaining. In case of absence of clinical contraindications, further examination should be recommended.