Portal de Periódicos da CAPES

Classification of isolated tumor cells and micrometastasis

2000; Wiley; Volume: 89; Issue: 3 Linguagem: Inglês

10.1002/1097-0142(20000801)89

1097-0142

Gábor Méhes, Armin Witt, E. Kubista, Peter F. Ambros,

Cancer Genomics and Diagnostics

The growing number of publications clearly necessitate a classification system of isolated or micrometastatic tumor cells of distant sites1 that will enable us to translate different kinds of laboratory findings into a common language. However, although micrometastases in regional lymph nodes are diagnosed according to well established immunohistochemical and morphologic criteria, a methodologic uncertainty is palpable in the detection of isolated tumor cells (ITC) in hematologic samples. The basic discrepancies in the appraisal of clinical findings concerning the significance of isolated tumor cells can, at least in part, be explained by methodologic problems.2 In the accompanying editorial by Page et al.,3 two major categories of methodologic problems, inherent in the detection of ITC by immunocytochemistry (ICC), are raised: false-positive reactions and the lack of tumor cell quantification. In addition to artificial reactions and ectopic epithelial cells, which both are mentioned by the authors, endogenous alkaline phosphatase (AP) activity also is a frequent cause of false-positive immunocytologic reactions in cytologic preparations. Double ICC staining identifies the majority of AP directly-reactive cells in the bone marrow of epithelial carcinoma patients as CD45 positive (+)/Ig κ or λ+ and thus, these cells can be identified as mature plasma cells.4 Conversely, only a rough estimation of the tumor cell content of a particular sample can be given because no adequate measures for quantification exist. Despite standardized slide processing, differences as great as 3.6 times in the total cell number could be observed between identically handled cytocentrifuge slides, resulting in an inaccurate estimation of the tumor cell infiltrate (unpublished data). In recent years special efforts were made in our laboratory to eliminate the limiting technical factors inherent in conventional ICC analysis of hematopoietic samples from adult and pediatric solid tumor patients. Our data indicate that the immunofluorescent detection of tumor cell specific antigens significantly lowers the number of false-positive events. A computerized automatic microscope was designed to scan slides for immunofluorescent labeled cells, with the possibility of further morphologic, immunologic, and molecular cytogenetic analyses. (MetaCyte; MetaSystems, Altlussheim, Germany). The approach completes the conventional microscopic evaluations with four main elements: 1) cells with a specific immunofluorescence pattern can be stored on disk and relocated; 2) the absolute number of tumor cells can be determined and the quantification of the total analyzed cell number is performed simultaneously (DAPI counterstaining); 3) false-positive results due to endogenous enzyme activity are eliminated by the use of immunofluorescence; and 4) in doubtful cases, the neoplastic origin of a given cell can be verified by fluorescence in situ hybridization analysis after automatic repositioning. The automatic detection system was calibrated for the immunofluorescence detection of cells expressing cytokeratin (CK), (5D3; Novocastra, Newcastle, UK; and MNF116; Dakopatts, Glostrup, Denmark). In addition, the expression of epithelial type-1 mucin (BM12; Medac Diagnostica, Vienna, Austria) could be analyzed as a second parameter in the same cells. In spiking experiments, the exact number of tumor cells (e.g., MCF-7, ZR-75.1) could be recovered after an automatic search for CK positivity. Three tumor cells per 1 million mononuclear cells (MNCs) could be recovered reliably. False-positive results could be excluded by analyzing the mucin expression as well as by sequential molecular cytogenetic studies. False-negative results were not observed. Bone marrow samples from 2 puncture sites in 45 women who presented with Stage I (40 patients) or Stage II (5 patients) breast carcinoma (FIGO Staging System) were analyzed for the presence of ITCs. After CK/mucin type 1 immunofluorescence staining on cytospins, an automatic search for tumor cells was performed and the total cell number was determined. A mean number of 2 million MNCs per sample could be analyzed (range, 0.8–7.3 × 106). Surprisingly, none of the 45 cases showed tumor cell positivity. This is in contrast with previous results reporting on bone marrow ITCs in between 7–38% of lymph node negative breast carcinoma patients5-7 using antibodies against epithelial membrane antigen, CKs, or type 1 mucin. In a recent study by Braun et al.,8 the detection of special epitopes of CK 8, 18, and 19 by an antibody cocktail was found to be prognostically relevant in women with Stage I, II, and III breast carcinoma. However, none of these results explains why antibodies widely used in pathology departments for the detection of epithelial tumor cells should not react with ITCs in the bone marrow. According to our data, despite the relatively low number of cases analyzed by our method, we must conclude that the frequency of positive bone marrow samples in patients with localized breast carcinoma is significantly lower than described by the conventional light microscopic studies. These data clearly indicate that the application of different methods for ITC detection lead to different results. Therefore, as the next step toward understanding the issue of ITCs in the hematopoietic system and micrometastases at distant sites, new consensus laboratory protocols should be established. Such protocols should be appropriate as clinical tests but also should be constructed according to the strictest quality assurance guidelines. In addition to simplicity and cost-effectiveness, other important aspects, such as accuracy and reproducibility, should be borne in mind. The detection of tumor cells by automatic fluorescence microscopic analysis to our knowledge allows for the first time the verification of the neoplastic nature of the ITC, thus minimizing the risk of false-negative and false-positive cells. Gabor Mehes M.D.*, Armin Witt M.D. , Ernst Kubista M.D. , Peter F. Ambros Ph.D. , * CCRI, Children's Cancer Research Institute, St. Anna Kinderspital, Vienna, Austria, Clinic of Special Gynecology, General Hospital, Vienna, Austria, CCRI, Children's Cancer Research Institute, St. Anna Kinderspital, Vienna, Austria