Comparison of three commonly used cytologic preparations in effusion immunocytochemistry.

Discrepant results in effusion immunocytochemistry are often the result of specimen processing. Smears, cytospins, cell blocks, and monolayer preparations have all been used in various published studies; thus, there is no consistency in the immunostaining process for cytology to compare with the surgical pathology "gold standard" results. We sought to evaluate optimal specimen preparation for the immunostaining of effusion samples. Fourteen reactive and 15 malignant effusion samples (various epithelial/mesothelial neoplasms) were each prepared in three forms: air-dried cytospins (postfixed in ethanol), formalin-fixed, paraffin-embedded cell blocks, and liquid-based thin-layer (ThinPrep, CYTYC, Boxborough, MA) processing. All slides were immunostained with antibodies commonly used in effusion cytology: HBME-1, calretinin, E-cadherin, BerEP4, B72.3, LeuM1, and CA19-9. Cytospin and ThinPrep samples performed in a similar manner: high background staining was encountered in 66% of cases, most evident in three-dimensional clusters of cells. In addition, membrane staining patterns were difficult to interpret. Cell blocks provided the best milieu for morphologic interpretation, with less background staining (only 17% of cases) and results that most closely approximated those reported in the surgical pathology literature. The cost per test for cell block immunocytochemistry was also the most economical for our laboratory.