METHODS IN PATHOLOGY - MULTIPARAMETER FLOW CYTOMETRIC DNA ANALYSIS OFEFFUSIONS - A PROSPECTIVE-STUDY OF 36 CASES COMPARED WITH ROUTINE CYTOLOGY AND IMMUNOHISTOCHEMISTRY

Single-parameter flow cytometry (SFCM) is limited in its ability to de tect aneuploid and diploid malignant cells or accurately estimate S-ph ase fractions (SPF) in effusions because of the high degree of contami nation by benign mesothelial cells and inflammatory cells. We examined 36 pleural and peritoneal fluids by conventional cytology and multipa rameter FCM (MFCM) to analyze the DNA content of cells expressing epit helial markers cytokeratin, epithelial membrane antigen, carcinoembryo nic antigen, BRST-1, or BRST-3 (B72.3) and compared the results to tho se found with SFCM. The cases were also studied by immunohistochemistr y using the same antibody panel. By routine cytology, 14 of the 36 cas es were classified as carcinomas, 11 as reactive, 1 as mesothelioma, a nd 10 as suspicious. MFCM allowed reclassification of 5 of the 10 susp icious cases as carcinomas and the remaining 5 as reactive cases based on ploidy and marker expression. Whereas SFCM detected only 13 nondip loid carcinomas, MFCM detected 4 diploid and 15 nondiploid carcinomas. All reactive cases were diploid by SFCM or MFCM. The mesothelioma cas e showed two distinct peaks by MFCM, a diploid peak with SPF of 13.4% and a tetraploid peak with SPF of 36.1%. The SPF of the nondiploid car cinomas ranged from 5.9 to 50.4% and diploid carcinomas, from 3.5 to 1 4.5% when gated on epithelial cells. The reactive cases had SPF rangin g hom 0.4 to 4.4% In this MFCM study, we observed that epithelial memb rane antigen was the most sensitive marker (100%) for malignant effusi ons followed by BRST-1 (89%) and BRST-3 (84%) to characterize the cell s in the effusion. The results of our study indicate that, in the work up of equivocal effusions, compared with SFCM, MFCM is a superior anci llary test characterized by high sensitivity, specificity, and accurac y. However, in this study, because the immunohistochemical study resul ts were comparable in sensitivity and specificity to MFCM results with regard to marker expression, immunohistochemistry on cell block mater ial would be a more cost-effective method for routine evaluation of di fficult effusion cases.