Introduction: Thyroglobulin (Tg) measurement in needle washouts from fine-needle aspiration biopsies (FNA-Tg) increases the sensitivity of lymph node (LN) metastasis diagnosis in differentiated thyroid cancer (DTC). However, the cutoff value for FNA-Tg has not been clearly established and there are large differences between clinical studies, which hinders its interpretation. Our study aimed to investigate the optimal cutoff value of FNA-Tg and evaluate its utility in the diagnosis of LN metastasis of DTC.
Methods/design: This was a retrospective study of 211 consecutive cases of FNA from 143 patients identified from our institutional database, who underwent FNA citology and Tg measurement in needle washout for suspicious LN, between 2012 and 2016.
Results: From the total of 211 cases, 121 (57%) had personal history of DTC. FNA citology was benign in 114 (54%) and malignant in 64 (30%). The median FNA-Tg was 1168 ng/ml (interquartile range 2711974) in malignant LNs, and 0.1 ng/ml (interquartile range 00.27) in benign LNs. LN ressection was performed in 55 patients (38.4%), based on the combined results of FNA-Tg and FNA citology. Histology reported LN metastasis of DTC in 45 of these (81%). Compared to FNA-Tg values above 0.2 ng/ml, FNA citology showed superior specificity (95.4% vs 67.1%) but slightly inferior sensitivity (88.7% vs 91.5%). FNA-Tg values above 10 ng/ml showed 80.3% sensitivity and 100% specificity. Combining both diagnostic strategies (FNA citology and FNA-Tg above 0.2 ng/ml) showed superior diagnostic power than using either strategy alone (specificity 96.4% and sensitivity 91.5%). We evaluated the optimal cutoff values of FNA-Tg in determining malignant LNs from ROC analysis and the optimal cutoff value was 0.98 ng/ml (sensitivity, 85%; specificity, 95.4%).
Conclusion: Our results show that combining FNA citology and Tg measurement is useful for the investigation of LN metastasis of DTC. A FNA-Tg cutoff value above 1 ng/ml should lead to its diagnosis.
20 - 23 May 2017
European Society of Endocrinology