Accuracy of fine needle aspiration biopsy with and without the use of tumor markers in cytologically indeterminate thyroid lesions.

We investigated if the use of two tumor markers, galectin-3 and CD44v6, could improve diagnostic accuracy of thyroid fine needle aspiration biopsy (FNAB) in cytologically indeterminate lesions (CIL). 351 patients with CIL [cellular follicular lesion/suspicious follicular neoplasm/suspicious Hürthle cell neoplasm (CFL/sFN/sHCN), Hürthle cell neoplasm (HCN), and follicular neoplasm (FN)] and surgical follow-up were investigated. 251 patients had FNAB diagnoses made without help of tumor markers and the rest of 100 patients had FNAB diagnoses made with a known expression of tumor markers determined by the reverse transcription (RT)-PCR. Risk of malignancy in all 351 patients with CIL was 6.8%. In the group with FNAB made without RT-PCR, there were 140 CFL/sFN/sHCN with the risk of malignancy of 4.2%, 92 FN with the risk of malignancy of 13.0%, and 19 HCN with the risk of malignancy of 5.2%. In the group with FNAB made with RT-PCR, there were 49 CFL/sFN/sHCN with the risk of malignancy of 2.0%, 40 FN with the risk of malignancy of 7.5%, and 11 HCN with the risk of malignancy of 9.0%. In the group with at least one positive tumor marker (N = 69), the risk of malignancy was 3.1% for CFL/sFN/sHCN, 11.1% for FN, and 10.0% for HCN. In the group with negative tumor markers (N = 31) there were no malignancies. The use of tumor markers, galectin-3 and CD44v6, determined by RT-PCR improves only sensitivity of thyroid FNAB in CIL. In most patients with CIL, and negative both tumor markers, conservative approach is advisable.