ABSTRACT
Objective To investigate the diagnostic value of 18F-FDG PET/CT for thyroid nodules.Methods From January 2008 to May 2012,34 patients (13 males,21 females; age range:21-73 years,mean (53.00± 12.57) years) with thyroid nodules on 18 F-FDG PET/CT and with histopathological results were retrospectively analyzed.From January 2011 to December 2011,20 cases (9 males,11 females; age range:40-55 years,mean (45.00±4.72) years) were selected as control group.Wilcoxon rank sum test and ROC analysis (AUC ≥0.7 was considered the standard of medium-high accuracy) were used.PET/CT features taken to suggest malignant thyroid nodules were:focally high uptake on PET,indistinct boundary or heterogeneous density on CT with punctuate,round or curved calcifications,or with hypermetabolic cervical lymph nodes as ancillary supportive findings of metastasis.The sensitivity,specificity,positive predictive value,negative predictive value and accuracy of PET/CT for diagnosing thyroid nodules were calculated.Results (1) There were 18 patients with malignant and 16 with benign thyroid nodules.The SUVmax of benign,malignant nodules and normal controls were 7.59±8.69,5.75±4.48 and 1.38±0.57,respectively.The differences between malignant thyroid nodules and controls,between benign nodules and controls were significant (u=3.553,3.408,both P＜0.01).There was no significant difference between benign and malignant thyroid nodules (u =0.207,P＞0.05).(2) The AUC for the differentiation of benign and malignant thyroid nodules by ROC analysis was 0.557 (＜0.70).(3) The sensitivity,specificity,positive predictive value,negative predictive value and accuracy of 18F-FDG PET/CT for the differentiation of benign and malignant thyroid nodules were 72.2％ (13/18),75.0％ (12/16),76.5％ (13/17),70.6％ (12/17) and 73.5％ (25/34),respectively.Conclusions 18F-FDG PET/CT has limited value for the differentiation between benign and malignant thyroid nodules based alone on the degree of metabolic intensity.It may have improved diagnostic certainty if combined with the morphological features on CT.