Department of Pathology, St. Andrea Hospital (A.B.), University La Sapienza, Rome, Italy; and Departments of Pathology (M.P.) and Clinical and Biological Sciences (F.O.), University of Turin, Turin, Italy
Address correspondence to: Armando Bartolazzi, M.D., Ph.D., Department of Pathology, St. Andrea Hospital, University La Sapienza, Via di Grottarossa 1035, 00189, Rome Italy. E-mail: Armando.Bartolazzi{at}cck.ki.se.
To the editor:
Preoperative characterization of thyroid nodules by using galectin-3 immunodetection on cytological smears from thyroid FNA is going to become a widely accepted diagnostic procedure in a large number of European institutions and university hospitals. As coauthors of the multicenter study in which this test-method has been proposed and validated (1), we are obliged to comment on the article entitled "Galectin-3 is not an universal marker of malignancy in thyroid nodular disease in children and adolescents" by M. Niedziela and colleagues, published in the September issue of JCEM (2).
To improve the conventional cytological evaluation of thyroid nodules, Niedziela and colleagues propose a RT-PCR-based technique instead of an immunocytochemical method. There are at least two important criticisms to this approach that need to be considered and discussed.
First, it should be stressed that analysis of galectin-3 expression in thyrocytes cannot be considered as a substitute for the conventional morphological evaluation of each specific lesion. For this reason, an immunocytochemical diagnostic approach to the preoperative characterization of thyroid nodules is mandatory, because this simple and cheap procedure combines a morphological and immunophenotypical evaluation of thyroid cells.
This strategy is crucial because expression of galectin-3 is normally observed in activated endothelial cells as well as in the so-called "foamy macrophages," which are commonly observed in different benign (i.e. colloid goitres, Hashimotos thyroiditis, etc.) and malignant thyroid conditions. Moreover, it should be also considered that only the cytoplasm expression of galectin-3 is suggestive of malignancy, because nuclear immunoreactivity is physiologically observed in some benign thyroid conditions (3).
Consequently, the use of RT-PCR for evaluation of galectin-3 expression on morphologically undefined cytological material, as Niedziela and co-authors propose, doesnt represent a reliable diagnostic method and is affected by a large number of false positive results.
With this in mind, we are not surprised about the reported results regarding galectin-3 expression in Hashimotos thyroiditis and other benign thyroid lesions (2). For the reason discussed above we strongly discourage the use of RT-PCR for this specific diagnostic purpose. A reliable use of RT-PCR for characterization of thyroid nodules requires the set-up of expensive and time consuming tissue micro-dissection procedures to be applied on cytological and histological preparations.
Regarding the second point of discussion, expression of galectin-3 in some Hashimotos thyroiditis has been previously described in a very limited number of cases (1). Although the number of galectin-3 positive Hashimotos thyroiditis decreased consistently when a purified monoclonal antibody to galectin-3 and a biotin-free detection system were used in immunocytohistochemistry (4), we confirm the possibility that some of these thyroid diseases may express galectin-3.
The significance of galectin-3 expression in focal areas of oncocytic changes observed in some Hashimoto thyroiditis is currently under investigation (paper in preparation), but the conclusion of Niedziela and colleagues that an "abundant number of oxyphilic cells in this disorder is a consequence of neoplastic transformation within Hashimotos goiters" appears quite imprudent, especially if this conclusion is not supported by an adequate morphological study of these intriguing lesions.
Received October 14, 2002.
References