Tumor marker studies of cervical smears. Potential for automation

The distribution of three tumor markers, epithelial membrane antigen (EMA), carcinoembryonic antigen (CEA) and prekeratin (PK) was studied in exfoliated epithelial cells in cervical smears using an immunoalkaline phosphatase staining technique to demonstrate the antigens. EMA was expressed by abnormal cells in a consistent and reproducible fashion whereas the other two markers were variably expressed by both normal and abnormal cells. Our results suggest that immunocytochemical staining for EMA could be of value not only for the diagnosis of cervical intraepithelial neoplasia but also for the automated screening of cervical smears.     

Immunophenotyping of Mesothelial Cells and Carcinoma Cells With Monoclonal Antibodies to Cytokeratins, Vimentin, Cea and Ema Improves the Cytodiagnosis of Serous Effusions

This paper presents an immunocytochemical study performed on cytocentrifuged deposits from 109 peritoneal and pleural effusions including 20 transudates, 43 malignant metastatic effusions and 46 effusions containing atypical cells, unidentifiable as reactive mesothelial or malignant epithelial cells on the classical morphological criteria. A panel of four monoclonal antibodies (MAb) was used, including KL1 directed to cytokeratins (KER), V9 to vimentin (VIM), NEO 723 to carcinoembryonic antigen (CEA) and E29 to epithelial membrane antigen (EMA). In most transudates the reactive mesothelial cells coexpressed VIM and KER with a ring-like pattern for the latter proteins. In contrast, they were unreactive to anti-CEA and weakly and inconsistently reactive to anti-EMA. In malignant effusions, most carcinoma cells coexpressed EMA, CEA and KER with a predominant diffuse cytoplasmic pattern for the latter. Only a few malignant epithelial cells from five metastatic adenocarcinomas weakly expressed VIM. When used on the 46 effusions with unidentifiable cells, the panel of MAb allowed reactive mesothelial cells and malignant epithelial cells to be distinguished from each other in 39 of 46 cases (85%).     

Monoclonal antibodies in the cytodiagnosis of serous effusions

In order to assess the value of immunocytochemical staining as a method of discriminating between benign reactive mesothelial cells and malignant epithelial cells in serous effusions, we have studied the reactions of a panel of commercially available antibodies on cells harvested from 83 pleural and peritoneal fluids and compared the results with the clinical and cytological diagnoses. The antibodies used were raised against cytokeratin (PKK1), epithelial membrane antigen (EMA), carcino-embryonic antigen (CEA), pregnancy specific B1-glycoprotein (SP1) and leucocyte common antigen (LCA). Anti-CEA was positive in 16 of 39 effusions (41%) containing carcinoma cells. Pregnancy specific B1-glycoprotein (SP1) was positive in 33% of the same samples. Mesothelial cells did not stain with these antibodies. Thus anti-CEA and SP1 can be used to discriminate between benign mesothelial and malignant epithelial cells in effusions. Anti-PKK1 stained both benign reactive mesothelial cells and malignant epithelial cells and cannot be used to discriminate between these two cell types. Strong positive staining of malignant cells was noted with anti-EMA. However, as occasional weak staining of mesothelial cells was also noted, strong staining with this antibody may be regarded as suspicious but not conclusive of malignancy.     

Combined staining of TAG-72, MUC1, and CA125 improves labeling sensitivity in ovarian cancer: antigens for multi-targeted antibody-guided therapy.

Single antigen-targeted intraperitoneal radioimmunotherapy for ovarian cancer has shown limited success. Due to the heterogeneous expression of tumor antigens on cancer cells, a multi-antigen targeting approach appears logical to augment the therapeutic efficacy of antibody-guided therapy. In the interest of developing this novel approach, ovarian cancer tissue microarray slides containing cancer and benign/non-neoplastic tissue samples (n=92) were processed for single-, double-, and triple-antigen labeling using antibodies for the tumor-associated antigens TAG-72, MUC1, and CA125. Among all ovarian cancer types, 72%, 61%, and 50% of the samples showed immunolabeling for TAG-72, MUC1, and CA125, respectively. Expression level of these antigens was significantly (p<0.005) higher in advanced stage carcinomas compared with early stage. Of the 48 epithelial ovarian cancer samples, individual anti-TAG-72, MUC1, and CA125 antibody probing showed labeling in 89.5%, 87.5%, and 73.0% of the cases, respectively. In the majority of the cancer samples (>70%), a heterogeneous labeling pattern was observed (only 30-40% of the cancer cells within the sample were labeled). However, upon combining the three antigens (triple-antigen labeling), 98% of the epithelial ovarian cancer samples were labeled and >95% of the cancer cells within each sample were labeled. Our data indicate that the heterogeneous expression of cancer antigens appears to be a major obstacle in antibody-guided therapy, and this can be overcome by multiple antigen targeting. Therapeutic efficacy of antibody-guided therapy for ovarian cancer treatment will be enhanced by the combined targeting of TAG-72, MUC1, and CA125.     

The value of epithelial membrane antigen in the diagnosis of ovarian tumors.

The value of epithelial membrane antigen (EMA) in the diagnosis of ovarian tumors was investigated using an indirect immunoperoxidase staining technique on 91 histologic sections (88 tumors and 3 normal tissues) and 39 ascitic fluid smears (28 from patients with epithelial ovarian tumors and 11 from cases of myoma uteri). The rate of positive EMA staining was highest in malignant tumors (89.2%), second highest in tumors of low malignant potential (33.3%) and lowest in benign tumors (25.0%); normal ovarian tissues were negative for EMA. Of the malignant tumors, all 48 serous cystadenocarcinomas (100%) and 18 of 26 mucinous cystadenocarcinomas (69.2%) stained positively for EMA. In serous cystadenocarcinomas, the EMA staining was mainly localized on the luminal membrane of cells in well-differentiated tumors, but appeared on the entire cell surface and cytoplasm of cells in poorly differentiated tumors. The results of EMA staining on ascitic fluid smears were almost the same as the results for the histologic sections. The intensity and the localization of EMA staining were related to the grade of malignancy in these ovarian tumors. In comparison with staining for other antigens (carcinoembryonic antigen, CA-125 and human keratin protein), EMA was found to be one of the most sensitive markers for the diagnosis of ovarian cancer.     

Distinction between carcinoma cells and mesothelial cells in serous effusions. Usefulness of immunohistochemistry

The usefulness of an immunoperoxidase battery to distinguish carcinomatous from benign effusions was examined. Cell block sections from 90 previously diagnosed effusions were stained with antibodies to Leu-M1, B72.3, epithelial membrane antigen (EMA), carcinoembryonic antigen (CEA) and vimentin. The 90 cases comprised 69 carcinomas (23 mammary, 16 ovarian, 10 pulmonary, 7 gastrointestinal [GI] and 13 others), 2 malignant mesotheliomas and 19 cases with reactive mesothelial cells only. EMA and vimentin were the most useful markers for distinguishing carcinoma cells from reactive mesothelial cells. EMA reacted with 86% of the carcinomas while vimentin reacted with 90% of the reactive cases. Leu-M1, B72.3 and CEA, although generally less sensitive than EMA, were also helpful in this regard. Additionally, the use of Leu-M1 and CEA together may help to distinguish pulmonary from GI carcinomas.     

CEA, ZGM and EMA localization in cells of pleural and peritoneal effusion: a preliminary study

Carcinoembryonic antigen (CEA), the zinc glycinate marker (ZGM) and epithelial membrane antigen (EMA) have been described as epithelial or tumour markers of varying specificity. These antigens were studied by immunoperoxidase localization in selected cell blocks of 62 pleural or peritoneal effusions and compared to cytological findings and review of the clinical records. By cytological criteria, 25 of the cell blocks were positive for malignancy, 30 negative, and 7 inconclusive. CEA, ZGM, and EMA by immunoperoxidase staining were localized on the cell surface and often in the cytoplasm of malignant cells, in 11/25 (44 per cent), 17/25 (68 per cent) and 22/25 (88 per cent) of the positive cell blocks respectively. Ten (40 per cent) of these cases were positive for all three antigens, 7 (28 per cent) for two, and 6 (24 per cent) for one. Of the 7 cases which were inconclusive on routine cytological reporting, 5 were positive for at least one marker. In 3 of the 5 a diagnosis of malignancy was confirmed, and in the other two was strongly suspected as malignant on clinical grounds. Macrophages were sometimes positive for one or more markers (but showed cytoplasmic staining only) and mesothelial cells in some cases stained positively for EMA but were always negative for CEA and ZGM. Localization of the 3 antigens in cells of malignant effusions was compared with their localization in the primary tumours in 9 cases. Localization corresponded for CEA in 7 of 9 cases, for EMA in 8 of 8 an for ZGM in only 2 of 9. Effusion fluid levels for CEA were compared with the cytological and immunocytochemical findings in 30 cases. Mucin stains performed on the cell blocks were also compared with the immunoperoxidase findings.     

The value of the immunoperoxidase slide assay in the diagnosis of malignant pleural effusions in breast cancer

Whether immunocytochemical studies of malignant pleural effusions due to breast cancer would increase the diagnostic yield as compared with conventional effusion cytology was examined in 30 cases with biopsy-proven metastatic spread to the pleura. Conventional cytology was performed on air-dried smears as well as on cytocentrifuge preparations stained with the May-Grünwald-Giemsa stain. Immunocytochemistry was performed with monoclonal antibodies against carcinoembryonic antigen (CEA), epithelial membrane antigen (EMA) and human leukocyte antigen (HLA) and the peroxidase-antiperoxidase technique on glass slides after Ficoll-Hypaque centrifugation. By conventional cytology, 13 cases (43%) were positive for malignant cells, 6 cases (20%) were suspicious, and 11 cases (37%) were negative. In marked contrast, all 30 cases were immunocytologically positive for malignancy. Tumor cells in all cases demonstrated a positive reaction for EMA. Some mesothelial cells were also positive for EMA, but their reaction pattern was clearly distinguishable from that of the tumor cells. Twenty-one cases (70%) also showed CEA-positive tumor cells; mesothelial cells never reacted with CEA. Some tumor cells showed a loss of HLA expression. In conclusion, this immunocytologic method can be recommended as a routine procedure for greatly increasing the diagnostic yield of cytology in pleural effusions due to breast cancer.     

In vitro and in vivo regulation of tumor antigen expression by human recombinant interferons

In vitro treatment with either type I or type II Interferon (IFN) can selectively enhance the expession of several tumor antigens, such as the carcinoembryonic antigen (CEA) and the tumor-associated glycoprotein-72 (TAG-72) in different human carcinoma cell lines and result in enhanced level of monoclonal antibody (MAb) binding to the cell surface. In vivo animal studies demonstrated that treatment of athymic mice with a type I interferon [i.e. interferon-α(A)]significantly increased the expression of a 90 kDa tumor antigen which improved the targeting of a MAb to the carcinoma xenograft. More recent studies reported that in vitro IFN treatment of human adenocarcinoma cells isolated from human malignant serous effusions selectively increased the expression of TAG-72 and CEA. One can envision that the ability of these cytokines to upregulate the level of expression of human tumor antigens presents an important experimental model in which to study the regulation of markers often correlated with epithelial cell differentiation. In addition, the increase of selective MAb-defined antigens may also be exploited in an adjuvant setting to localize higher amounts of MAbs to the tumor cell surface and, thereby, improve the effectiveness of a MAb for tumor diagnosis and, possibly, therapy.     

The use of a panel of monoclonal antibodies in ultrastructurally characterized small cell carcinomas of the lung

In recent electron microscopic studies of 51 small cell carcinomas of the lung, the ultrastructural features of epithelial differentiation, particularly the presence of desmosomes, were associated with a tendency toward localized disease, clinical resectability and relatively long survival. Thirty-three of these cases were studied with a panel of monoclonal antibodies (B72.3, B1.1, AE1-AE3 and anti-Leu-7) directed against tumor-associated glycoprotein (TAG-72), carcinoembryonic antigen (CEA), cytokeratin and Leu-7 (an antigen common to natural killer cells and small cell lung carcinomas) to assess the correlation between the immunocytochemical and ultrastructural evidence of differentiation in tumors lacking differentiation at the light microscopic level. Of four small cell carcinomas with ultrastructural glandular differentiation, two were positive for CEA, three were positive for cytokeratin, and none were positive for TAG-72 and Leu-7. Of six tumors with ultrastructural squamous features, cytokeratin was expressed by three, CEA by one and TAG-72 and Leu-7 by none. Of the 23 classic oat cell carcinomas, cytokeratin was expressed by 14, Leu-7 by 3, CEA by 1 and TAG-72 by none. While the pattern of antigen expression did not predictably reflect the submicroscopic features, there was a significant association between keratin staining and extent of disease. The prospective use of similar antibody panels with both cellular and histologic material may therefore help to define clinically relevant categories of this biologically heterogeneous neoplasm.     

Malignant ascites of serous papillary ovarian adenocarcinoma. An immunocytochemical study of the tumor cells

In 17 malignant peritoneal effusions due to papillary serous adenocarcinoma of the ovary, the reaction patterns of the tumor cells to monoclonal antibodies (MAbs) against surface antigens were studied and compared with the reaction patterns of mesothelial cells in the same effusions. The following surface markers were used with the adhesive slide method: epithelial membrane antigen (EMA), human epithelium-specific cell surface antigen (HEA-125), human endothelial antigen (BMA-120), carcinoembryonic antigen (CEA 3-13), an antibody against natural killer cells and cytotoxic cells (BMA-070), granulocyte antigen (Leu M1) and leukocyte antigen of class I (HLA-1). In all cases, from 30% to 95% of the tumor cells reacted with EMA and HEA-125. Tumor cells showed a positive staining with CEA 3-13 in only five cases. In all cases, from 75% to 95% of the tumor cells reacted positively with BMA-120. The reactivity of a few mesothelial cells with EMA and of all mesothelial cells with BMA-120 did not interfere with the identification of positive tumor cells since the reaction patterns were different. Interestingly, our study demonstrated that BMA-070, an MAb identifying natural killer cells and cytotoxic cells, is also a most useful tumor marker. The same was found to be true for Leu M1, an MAb originally thought to react only with granulocytes. The tumor cells showed a partial or total loss of the expression of HLA-1 reactivity. Since all cases were immunocytochemically positive for tumor cells while conventional cytology was positive in only 13 of the cases, the immunocytochemical analysis of malignant peritoneal effusions due to papillary serous adenocarcinoma of the ovary seems able to improve the cytologic diagnosis of the fluids.     

Complementation of expression of carcinoembryonic antigen and tumor associated glycoprotein-72 (TAG-72) in human colon adenocarcinomas.

Enzyme-labeled monoclonal antibodies (MAbs) were used in an immunohistochemical, dual-staining study of 10 colon adenocarcinomas. MAbs B72.3 and COL-4, reactive with the high molecular weight tumor-associated glycoprotein-72 (TAG-72) antigen and carcinoembryonic antigen (CEA), respectively, were labeled with horseradish peroxidase or alkaline phosphatase. Dual staining using the two MAbs on a single tissue section (formalin-fixed, paraffin-embedded) showed that greater numbers of carcinoma cells could be detected by using the combination of the two MAbs than could be detected by use of either MAb alone. In many tumors, some carcinoma cells reacted with MAb B72.3, some reacted with MAb COL-4, and some cells reacted with both MAbs. Only 1 of 10 carcinomas showed greater than 75% reactive cells when stained with each MAb individually. In 9 of 10 cases, however, greater than 75% of cells reacted when the combination of MAbs was used. Cell surface and cytoplasmic patterns of reactivity were observed with both MAbs while some pools of extracellular mucin were composed of both TAG-72 and CEA. This study supports the rationale for the use of a combination of anti-TAG-72 and anti-CEA MAbs for in vitro immunologic detection and potential in vivo immunodiagnostic and immunotherapeutic applications for these MAbs in colon adenocarcinoma patients.     

Induction of delayed-type hypersensitivity responses by monoclonal anti-idiotypic antibodies to tumor cells expressing carcinoembryonic antigen and tumor-associated glycoprotein-72

The use of anti-idiotypic antibodies as immunogens represents one potential approach to active specific immunotherapy of cancer. Two panels of syngeneic monoclonal anti-idiotypic antibodies were generated. One panel was directed against mAb CC49 and the other to mAb COL-1. mAb CC49 recognizes the pancarcinoma antigen (Ag), tumor-associated glycoprotein-72 (TAG-72), and mAb COL-1 recognizes carcinoembryonic antigen (CEA). Seven anti-idiotypic (AI) antibodies (Ab2) designated AI49-1–7 were generated that recognize the variable region of mAb CC49. These mAb were shown to inhibit the interaction of mAb CC49 (Ab1) with TAG-72 (Ag). Five anti-idiotypic antibodies designated CAI-1–5 were also generated to the anti-CEA mAb, COL-1 (Ab1). These Ab2 were shown to inhibit the interaction between COL-1 (Ab1) and CEA (Ag). Immunization of mice, rats, and rabbits with Ab2 directed against CC49 or COL-1 could not elicit specific Ab3 humoral immune responses, i.e., antibody selectively reactive with their respective target antigens. However, immunization of mice with the CC49 anti-idiotypic antibody (Ab2), designated AI49-3, could induce a delayed-type hypersensitivity response (DTH) specific for tumor cells that express TAG-72. Similarly, immunization of mice with an anti-idiotypic antibody directed against COL-1, designated CAI-1, could induce specific DTH cell-mediated immune responses to murine tumor cells that express human CEA on their surface. These results thus demonstrate that while some anti-idiotype mAb may not be potent immunogens in eliciting Ab3 humoral responses, they are capable of eliciting specific cellular immune responses against human carcinoma-associated antigens. This type of mAb may ultimately be useful in active immunotherapy protocols for human carcinoma.Some of the studies described in this paper were in partial fulfillment of requirements for the completion of Dr. Irvine's dissertation at the George Washington University     

Usefulness of ploidy, AgNOR and immunocytochemistry for differentiating benign and malignant cells in serous effusions

OBJECTIVE: The objective of this study was to establish the value of different markers in differentiating reactive mesothelial cells from metastatic adenocarcinomatous cells in serous effusions (SE). METHODS: Forty-five SE were processed for morphological examination (Papanicolaou stain), assessment of ploidy, AgNOR counting and immunocytochemical assay of carcinoembryonic antigen (CEA), epithelial membrane antigens (EMA), Ber-EP4 and Leu-M1. Ploidy was established in an image-analyser in smears stained by the Feulgen stain method. AgNOR dots were counted in the smears stained with the silver nitrate assay for non-histone proteins present in the nucleolar organizer region. CEA, EMA, Ber-EP4 and Leu-M1 were evaluated by immunocytochemistry using the streptavidin-biotin complex method. RESULTS: All the smears with positive cytology were aneuploid. Three false negatives by morphological studies were aneuploid, with AgNOR values in two of them corresponding to the neoplastic group. CEA and Leu-M1 showed a low specificity; EMA and Ber-EP4 showed moderate sensitivity. CONCLUSIONS: The assessment of ploidy and the study of AgNOR were better methods than immunocytochemistry for distinguishing between reactive mesothelial cells and adenocarcinomatous cells in serous fluid.     

Evaluation of Immunocytochemical Staining As A Method of Improving Diagnostic Accuracy In A Routine Cytopathology Laboratory

Immunocytochemical stains in a routine cytopathology laboratory can be used to distinguish between benign and malignant cells, and to identify tumour type. In our laboratory 30 problematic cases were selected for immunocytochemical stains and the results analysed in this paper. The following markers were used: cytokeratin (CAM5.2), carcinoembryonic antigen (CEA), kappa and lambda light chains, leucocytic common antigen (LCA), chorionic gonadotrophin (hCG), prostate specific antigen (PSA), L26, UCHL1, S100-protein and vimentin. Twelve FNA (four lymph nodes, one parotid swelling, two from lungs, two from pleura and chest wall, one from lumbar region, two from soft tissue masses), and 18 effusions (12 pleural effusions, five ascitic fluids, one pericardial effusion) were investigated. We found immunocytochemical stains of value in formulating the cytological diagnosis in 11/12 of FNA and 15/18 of effusions.     

Interferon α enhances expression of TAG-72 and carcinoembryonic antigen in patients with primary colorectal cancer

Interferons up-regulate the expression of human tumor-associated antigens in animal models and in vitro. The use of interferons may enhance the immunodetection and immunotherapy of tumors by monoclonal antibodies that detect tumor antigens. For this strategy to be effective, however, the interferon must have an effect at the site of the tumor. In this study, the induction by interferon (IFN) of two tumor surface antigens was evaluated in six patients with primary colorectal cancer. Patients were treated with IFN and 48 h later underwent resection of the tumor. The interferon treatment induced expression of a tumor-associated glycoprotein (TAG-72) in two patients without antigen expression prior to interferon but had no effect on one TAG-72-negative tumor. IFN did not induce expression of carcinoembryonic antigen (CEA) in the two patients whose tumors were CEA-negative prior to interferon. In all patients with heterogeneous expression of CEA and TAG-72 prior to IFN treatment, preoperative interferon increased the percentage of cells positive for CEA in two patients and TAG-72 in one patient. This study supports the addition of interferon induction to immunotherapy regimens directed at the CEA and TAG-72 cell-surface antigens.     

Immunofluorescent staining of metastatic carcinoma cells in serous fluid with carcinoembryonic antibody, epithelial membrane antibody, AUA-1 and BER-EP4

Using an indirect immunofluorescence technique, we assessed the accuracy and clinical usefulness of a panel of monoclonal and polyclonal antibodies. the panel consisted of carcinoembryonic antibody (CEA) and epithelial membrane antibody (EMA), AUA-1, and Ber-EP4 conjugated with fluorescein isothiocyanate. Twenty-six specimens from pleural, peritoneal or pericardial effusions known to contain carcinoma cells (adenocarcinoma or large cell anaplastic carcinoma) and 16 specimens without carcinoma were first examined. the sensitivity and specificity for each of the antibodies were as follows: CEA, 71% and 75%; EMA, 96% and 81%; AUA-1, 80% and 100%; and Ber-EP4, 85% and 100%, respectively. the panel of antibodies was then applied to a group of 14 'problematic' fluids. These had been identified as causing dilemmas in interpretation, either because the cells in the fluids were of equivocal appearance on light microscopy, or the cytological diagnosis was different from that expected in the light of the clinical condition of the patient. Insufficient cellular material was present in one specimen. In five (39%) of the cases the immunochemical staining supported the light microscopic diagnosis. In four (30%) cases, however, the results indicated that the original light microscopic report was incorrect. Two of these were examples of large cell carcinoma of the lung, in which false negative reports had been issued on pleural fluids. the other two were cases of benign ovarian tumours in which a false positive report had been issued. the immunostaining also clarified the final diagnosis in the three patients (23%) on whom 'suspicious' cytological reports had previously been issued. the remaining case, fluid from a patient with a high grade mixed Mullerian tumour of the ovary, was unresolved. We conclude that immunofluorescent staining by AUA-1, EMA and Ber-EP4 is an aid in the cytological interpretation of serous fluids. CEA is much less helpful.     

Expression of epithelial membrane antigen on malignant mesothelioma cells. An immunocytochemical and immunoelectron microscopic study

The presence of epithelial membrane antigen (EMA) on malignant mesothelial cells found in pleural and ascitic fluids was demonstrated immunocytochemically using a monoclonal anti-EMA antibody. Serous fluids of 25 patients with malignant mesotheliomas were investigated. In 23 cases, varying numbers of EMA-positive tumor cells were present; in 2 cases, no such cells were found. Immunoelectron microscopy was performed both on Lowicryl-embedded sediments of serous fluids and by application of preembedding techniques using the immunogold method. Expression of EMA by the immunogold method was found selectively on the villi of the malignant mesothelioma cells whereas the nonvillous, flat surfaces were largely EMA-negative. The results indicate that immunoelectron microscopy may offer a useful adjunct in the diagnosis of malignant mesothelioma in serous fluids.     

Density distribution and immunological reactivity of tumor cells from peritoneal effusions of patients with ovarian neoplasms.

Ascitic samples from 19 patients with primary ovarian non-mucinous carcinomas, three with Krukenberg tumors and eight with noncancerous peritoneal effusions were studied by conventional cytology and immunocytochemical staining. Density gradient centrifugation was applied to fractionate ascitic fluid cells. The enrichment of cell types by this method facilitated their cytomorphological characterization and identification of neoplastic cell subpopulations existing in peritoneal effusions. Immunophenotypic studies of cells were made using monoclonal antibodies (mAbs) against ovarian carcinoma-associated antigens (OC 125, 10B, 8C) and carcino-embryonic antigen (CEA). Non-specific cross-reacting antigen (NCA) was applied as a marker for granulocytes which often accompany peritoneal effusions. Our results indicated that immunofluorescence (IF) staining contributed to the distinction between the primary and secondary ovarian carcinomas. Density gradient centrifugation appeared to be a useful method for separation of mesothelial cells.     

Carcinoembryonic antigen in effusions. A diagnostic adjunct to cytology

A total of 189 effusion specimens (100 benign and 89 malignant) submitted for cytologic examination were assayed for carcinoembryonic antigen (CEA) by an enzyme immunoassay to determine whether the addition of CEA evaluation to cytologic study would improve the diagnostic accuracy for the detection of malignancy. The sensitivity and specificity were 78% and 90%, respectively, for a cytologic diagnosis of malignancy and 68% and 99%, respectively, for a positive CEA (greater than 5 ng/mL). CEA assay was negative in the most common epithelial malignancies of the female genital tract (15 of 17 cases), mesotheliomas (5), lymphomas (7) and alveolar-cell carcinoma of lung (1). CEA assay was positive in 55 of 89 cases of malignancy, including 14 cases with cytologically negative malignant effusions. The CEA assay sensitivity for lung carcinoma (95% for adenocarcinoma, 100% for oat-cell carcinoma and 100% for carcinosarcoma), breast carcinoma (95%), and gastrointestinal carcinoma (100%) were all over 90%. No significant difference in the levels of CEA was noted between gastrointestinal and lung adenocarcinomas. Oat-cell carcinomas and squamous-cell carcinomas had lower values. In cases of an effusion with an unknown primary, an elevated CEA in the fluid is diagnostic of metastatic carcinoma arising from the breast, lung or gastrointestinal tract.