Background: Fine-needle aspiration (FNA) is currently the primary diagnostic procedure in diagnosing thyroid malignancy and guides surgeons on patient selection for thyroidectomy for thyroid nodules. Diagnostic sensitivity is reported to be approximately 80%; however, patients with negative FNA results do not necessarily undergo surgery and are often not considered in statistical analysis. This may lead to bias in previous reported sensitivity of FNA. The aim of this study was to assess the diagnostic performance attributes of FNA based on a comprehensive review and summary of previous literature.
Methods: A comprehensive review of published literature from 1966 to 2005 was performed, using structured selection and appraisal methods to include all studies that have assessed the sensitivity of FNA for detecting thyroid malignancy in palpable thyroid nodules. A statistical modeling study was designed to estimate the possible true sensitivity and specificity of FNA.
Results: Twelve studies fulfilled inclusion criteria and were included in the review. Only 1 study had greater than 25% of patients with negative FNA results who proceeded to thyroidectomy. Statistical modeling indicated that the sensitivity of FNA is highly dependent on the risk of malignancy in the patients with negative FNA results who did not undergo thyroidectomy; in the “same risk” scenario, where the risk of malignancy in the whole group with negative FNA result was assumed to be the same as that in patients with negative FNA results who underwent surgical biopsy; sensitivity could be as low as 66% (confidence interval [CI]: 65–68%).
Conclusion: Based on existing reports, the true diagnostic attributes of FNA for thyroid malignancy in palpable nodules are uncertain and FNA could miss up to a third of all thyroid malignancy. Further research is required to investigate the incidence of malignancy in FNA negative cases and to determine the additive effect of clinical judgment.