Objective: In the present study we aimed to assess the role of galectin- 3, cytokeratin 19, thyroid peroxidase and CD44v6 in distinguishing benign from malignant follicular lesions.

Material and Method: Fifty-four malignant and 50 benign lesions were evaluated and classified according to World Health Organization 2004 histological classification. Galectin-3, cytokeratin 19, thyroid peroxidase and CD44v6 were performed immunohistochemically and the slides were evaluated by two independent investigators. Sensitivity, specificity and diagnostic accuracy were assessed for each antibody tested.

Results: Sensitivity, specificity and diagnostic accuracy were as follows respectively: Galectin-3: 59,25%, 84% and 71,15%; Cytokeratin 19: 70%, 82% and 75,4%; Thyroid peroxidase: 61%, 70% and 65,4%; CD44v6: 20,4%, 88% and 52,9%.

Conclusion: The negativity for Galectin-3 and Cytokeratin 19 can not exclude malignancy but positivity can be thought as a sign of malignant feature or potential for lesions in which there is strong suspect of malignancy. Thyroid peroxidase immunostaining failed to differantiate benign from malignant oxyphilic tumors but decreased expression can be used as a malignancy marker together with Galectin-3 and/or Cytokeratin19 positivity in suspicious cases. CD44v6 does not seem to be reliable in distinguishing benign from malignant follicular patterned thyroid lesions.

In conclusion, our approach is to take as much new samples or serial sections as possible in cases without clear-cut evidence of malignancy but with histological and immunohistochemical suspicion. Follicular variant papillary carcinoma has different criteria for malignancy and it should be always kept in mind while evaluating a benign-looking lesion with immunohistochemical signs that favor malignancy.