Indeterminate results at fine-needle aspiration biopsy (FNAB) of thyroid nodules pose a clinical dilemma, because only 20% to 30% of patients suffer from malignancy. Previous studies suggested that the false-negative ratio of [(18)F]-2-fluoro-2-deoxy-D-glucose-positron emission tomography (FDG-PET) is very low; therefore, it may help identify patients who would benefit from (hemi)thyroidectomy. A systematic literature search was performed in 5 databases. After assessment, the identified studies were analyzed for heterogeneity, and the extracted data of test characteristics were pooled using a random-effects model. Threshold effects were examined, and publication bias was assessed. The query resulted in 239 records, of which 6 studies met predefined inclusion criteria. Data from 225 of the 241 described patients could be extracted. There was mild to moderate heterogeneity in study results (inconsistency index [I(2)] = 0.390-0.867). The pooled prevalence of malignancy was 26%. Pooled sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 95% (95% confidence interval [95% CI], 86%-99%), 48% (95% CI, 40%-56%), 39% (95% CI, 31%-47%), 96% (95% CI, 90%-99%), and 60% (95% CI, 53%-67%), respectively. Sensitivity increased to 100% for the 164 lesions that measured >15 mm in greatest dimension. There was no evidence of threshold effects or publication bias. A negative FDG-PET scan in patients who had thyroid nodules >15 mm with indeterminate FNAB results excluded thyroid cancer in a pooled population of 225 patients. Conversely, a positive FDG-PET result did not identify cancer, because approximately 50% of these patients had benign nodules. The authors concluded that the incorporation of FDG-PET into the initial workup of such patients before surgery deserves further investigation.
Copyright © 2011 American Cancer Society.