| Literature DB >> 11992379 |
Güliz Akdas Barkan1, Mark A Rubin, Claire W Michael.
Abstract
The purpose of this study was to compare the cytologic features of melanoma fine-needle aspirates (FNAs) prepared by ThinPrep (TP) with those in conventional smears (CS) and to identify any diagnostic pitfalls. Fifty-one aspirates diagnosed as melanoma were obtained, 36 of which were prepared by both TP and CS. The preparations were evaluated for cellularity, cell aggregates, cellular appearance, melanin pigment, cytoplasmic, and nuclear features. Categorical data were analyzed by the chi-square test and continuous data by the Wilcoxin-signed rank test. Correlation was determined by Spearman's test for bivariate correlations (rho). Good correlation between the two methods was identified for the following features: cellularity, cell type, bi/multinucleated cells, cytoplasmic features, NC ratio, and presence of macronucleoli. TP exhibits coarser chromatin compared to CS (P = 0.005). Six of 36 CS contained large cellular groups; none of the TP contained them (P = 0.018). Twenty-five of 36 CS contained intranuclear inclusions as opposed to 12/36 TP (P < 0.001). The number of inclusions was significantly reduced on TP. The amount of intracellular melanin was the same with both techniques. Background melanin was markedly reduced on TP except when either trapped by fibrin or attached to cellular clusters (P = 0.006). Background blood was also markedly reduced on TP (P < 0.005). In summary, the cytological features of TP and CS for FNA evaluation of melanoma correlate well; however, one needs to be aware of the cytologic alterations introduced by TP. TP is a sufficient preparation method in the diagnosis of melanoma FNA aspirates when performed by clinicians. It is also a useful adjunct in bloody or low-cellular aspirates, where it tends to reduce the background blood and concentrate the cells. Copyright 2002 Wiley-Liss, Inc.Entities:
Mesh:
Year: 2002 PMID: 11992379 DOI: 10.1002/dc.10099
Source DB: PubMed Journal: Diagn Cytopathol ISSN: 1097-0339 Impact factor: 1.582