48 chapter 2 To find the most accurate combination of immunostains, Saggiorato et al. explored the expression of galectin-3, HBME-1, thyroperoxidase, CK-19 and keratan-sulphate in 125 cytological follicular neoplasms, 24 of which were Hürthle cell lesions. Galectin-3 was not only the most accurate marker individually, but also in combination with other stains. Sequential HBME-1 staining of galectin-3negative cases reached 98% sensitivity and 98% specificity in non-oncocytic lesions. In oncocytic lesions, sequential CK-19 staining was more preferred with 100% sensitivity and 100% specificity [212]. The common denominator between all these studies is the combined use of galectin-3 and HBME1. Unfortunately, clinical validation studies regarding this combination are limited. Its seemingly promising diagnostic accuracy warrants further assessment in future prospective studies. Performance of immunocytochemistry in Hürthle cell cytology Expression of ICC markers in Hürthle cell nodules differs from non-oncocytic indeterminate cytology. Hürthle cell carcinomas were distinguished in the cytological samples by typical overexpression of markers associated with a high degree of cell proliferation, disorganized tissue structure and intermediate differentiation, such as Ki-67, laminin, cyclin D1 and cyclin D3. Overexpression reflects the known more erratic behaviour of Hürthle cell carcinoma [223, 227]. Moreover, markers that were highly diagnostic in indeterminate nodules in general, also seem differently expressed in Hürthle cell lesions. Saggiorato et al. demonstrated that two combinations of ICC markers were extraordinarily accurate: galectin-3 and CK-19 staining was 100% sensitive and 100% specific; galectin-3 and thyroperoxidase staining was 100% sensitive and 85% specific [212]. In a previous meta-analysis, inclusion of Hürthle cell lesions was related to between-study heterogeneity [201]. Hürthle cell lesions require a biotin-free ICC method, as Hürthle cells themselves are rich in biotin. Thus, much-used biotin-based methods may consequently cause false positive and highly intensive staining in Hürthle cell neoplasms [204, 210, 219]. Availability, cost-effectiveness and limitations of immunocytochemistry Current application of immunocytochemistry is limited. Clinical validation studies for all of the described immunomarkers are scarce, and no cost-effectiveness studies are available to date. Yet, the technique is widely available, relatively inexpensive and fast in comparison to other (molecular) techniques. Costs per immunostain vary up to €20, partly depending on simultaneous local application of the technique and similar stains for immunohistochemistry. Immunocytochemistry is preferably performed on cell block FNAC specimens, but can be performed in all types of cytology, from direct smears to liquid-based cytology [217, 219]. ICC is impossible when the FNAC specimen has poor cellularity or too much obscuring blood [228]. Also, immunostaining of cytology is technically more difficult than histological staining, especially in (destained) cytology smears. Technical inconsistency and interobserver variation likely lead to false-negative results [202, 220]. Stain intensity thresholds or percentage of stained cells necessary to raise suspicion of malignancy vary in the available literature. Consistent methodology and assessment thresholds should be determined to improve reproducibility of ICC results.
RkJQdWJsaXNoZXIy MTk4NDMw