Immunoradiometric assay of CA 125 in effusions. Comparison with carcinoembryonic antigen

Diagnostic value of carcinoembryonic antigen in malignant pleural effusion: a meta-analysis

BACKGROUND AND OBJECTIVE

Conventional tests are not always helpful in making a diagnosis of malignant pleural effusion (MPE). Many studies have investigated the utility of pleural carcinoembryonic antigen (CEA) in the early diagnosis of MPE. The present meta-analysis determined the accuracy of CEA measurement in the diagnosis of MPE.

METHODS

A systematic review of English language studies was conducted and data on the accuracy of pleural CEA concentrations in the diagnosis of MPE were pooled using random effects models. Receiver operating characteristic curves were used to summarize the overall test performance.

RESULTS

Forty-five studies met the inclusion criteria for the meta-analysis. The summary estimates for CEA in the diagnosis of MPE were: sensitivity 0.54 (95% CI: 0.52-0.55), specificity 0.94 (95% CI: 0.93-0.95), positive likelihood ratio 9.52 (95% CI: 6.97-13.01), negative likelihood ratio 0.49 (95% CI: 0.44-0.54) and diagnostic odds ratio 22.5 (95% CI: 15.6-32.5). Analysis of a subset of 11 studies which examined the value of pleural CEA in ruling out a diagnosis of malignant mesothelioma found that the sensitivity and specificity of a CEA level exceeding cut-off values were 0.97 (95% CI: 0.93-0.99) and 0.60 (95% CI: 0.55-0.65), respectively.

CONCLUSIONS

Measurement of pleural CEA is likely to be a useful diagnostic tool for confirming MPE, and is also helpful in the differential diagnosis between malignant pleural mesothelioma and metastatic lung cancer. The results of CEA assays should be interpreted in parallel with clinical findings and the results of conventional tests.

Serum CA-125 screening for ovarian cancer in patients with dermatomyositis

Ovarian cancer is the most overrepresented malignancy diagnosed in women with dermatomyositis. Unfortunately, screening with pelvic examination rarely detects this cancer prior to the development of metastatic disease. Our objective was to examine the use of serum CA-125 antigen levels in screening patients with dermatomyositis for ovarian cancer. A single blinded, case-control study was conducted in our institution of CA-125 levels in 14 women diagnosed with dermatomyositis between 1986 and 1993, 4 of whom subsequently developed ovarian cancer. In the 4 patients who developed ovarian cancer ("cases"), CA-125 determinations were performed on serum stored 5 to 19 months prior to the diagnosis of ovarian cancer. In the remaining 10 patients ("controls"), serum was drawn for CA-125 level determination at the time of the study, and simultaneous gynecologic and endovaginal ultrasound examinations were performed to exclude clinical evidence of ovarian cancer. All CA-125 serum measurements were performed simultaneously by a technician blinded to disease status using one diagnostic kit. CA-125 was found to be elevated in 2 patients with ovarian cancer (on serum obtained 5 and 13 months prior to the date of diagnosis of ovarian cancer) and in none of the control patients without clinical or ultrasound evidence of ovarian cancer (relative risk = 20, 95% confidence interval = [0.64, 666]). In these 14 patients, the sensitivity of CA-125 elevation for detection of ovarian cancer was 50%, and specificity was 100%. Serum CA-125 screening for ovarian cancer in patients having dermatomyositis may be useful; however, prospective studies are needed to confirm this and to determine the effect of screening on cancer stage at diagnosis and long-term survival.

CA125 in peritoneal fluid of ovarian cancer patients

The presence of CA125 was assessed in peritoneal fluid from 70 patients with ovarian cancer and 32 control patients. The follow-up period ranged from 39 to 89 months (median, 56 months). The cutoff for normal peritoneal fluid CA125 levels was determined to be 250 U/ml. A positive correlation between the serum and peritoneal fluid CA125 levels was observed (P less than 0.001). Peritoneal fluid levels were higher than serum levels in all patients. Patients with evidence of active ovarian cancer showed higher peritoneal fluid CA125 levels than the control patients (P less than 0.001). Peritoneal fluid CA125 levels correlated inversely with survival (P = 0.004). The peritoneal fluid CA125 levels were higher in patients with bulky tumor than in those with small (less than 1 cm) tumors (P less than 0.001). Eight out of twenty-six patients with active cancer and available peritoneal cytology had a negative peritoneal cytology. Three of these patients showed elevated peritoneal fluid levels. Three patients out of twenty-four showed elevated peritoneal fluid CA125 levels at second-look laparotomy. These 3 patients had negative biopsies at second-look surgery, but relapsed during the observation period. At second-look laparotomy an elevated peritoneal fluid CA125 level may imply a bad prognosis, but a normal level does not exclude the presence of disease.

Cancer antigen 125, carcinoembryonic antigen, and carbohydrate determinant 19-9 in ovarian tumors

The authors studied data of combination assays of tumor markers, because simultaneous elevation of different types of tumor markers in the serum was puzzling. They interpreted such phenomena regarding cancer antigen 125, carcinoembryonic antigen, and carbohydrate determinant 19-9 in ovarian tumors. The tissue expression of the antigens was compared with preoperative serum levels. Several different factors were found to cause the simultaneous elevation of two or three of these markers in the serum. Furthermore, even when the levels of some of the tumor markers were raised in the serum, the ovarian tumor did not always produce the marker by itself. This study indicates that immunohistochemical identification of a marker in tumor tissue is prerequisite to the use of that marker in the serum to monitor disease status.

Diagnostic significance of the tumour markers CEA, CA 15-3 and CA 125 in malignant effusions in breast cancer

Concentrations of the tumour-associated antigens CEA, CA 15-3 and CA 125 were determined in serous effusions (EF) from patients (pts) with breast cancer. As controls, serous effusions from patients with benign and other malignant diseases were used. The EF levels were also compared with those of serum. CA 15-3 was elevated (greater than 35 U/ml) in 65.7% of breast cancer EF, in 39.3% of EF in various other malignant diseases and in 0% of benign EF. CEA was elevated (greater than 9 ng/ml) in 37.5% of breast cancer EF, in 17.9% of EF of various other malignant diseases and in 27% of benign EF. CA 125 was elevated (greater than 35 U/ml) in 93.8% of breast cancer EF, in 78.6% of EF of various other malignant diseases and in 58.8% of benign EF. There was a statistically significant correlation between EF and serum values for the markers studied. Sensitivity and specificity for CA 15-3 were 65.7% and 76.6%, for CEA 37.5% and 77.7% and for CA 125 93.8% and 28.7%, respectively. CA 15-3 is a marker with definite diagnostic accuracy compared to CEA and CA 125 in breast cancer EF. CA 125 appears to derive from proliferating mesothelia rather than cancer cells alone and occurs in a broad spectrum of malignant as well as benign EF. The above markers should not be used alone for diagnosis of breast cancer in patients with serous effusions.

CA 72-4 radioimmunoassay in the diagnosis of malignant effusions. Comparison of various tumor markers

We evaluated the utility of the CA 72-4, CEA, CA 125, CA 19-9 and CA 15-3 radioimmunoassays for the detection of tumor-associated antigens (TAAs) in effusions of malignant vs. benign origin. Fluids were obtained from 51 patients with adenocarcinomas, 27 with non-epithelial malignancies, and 68 with benign disorders. The CA 72-4 radioimmunoassay (cut-off value 8.5 U/ml) detected the TAG-72 antigen in 51% of adenocarcinoma patients' effusions, while only 1 of 68 benign specimens had an elevated TAG-72 level. Similarly, CEA levels above 5 ng/ml were found in 55% of the fluids from patients with adenocarcinoma and 3.2% of effusions from patients with benign disease. CA 19-9 (cut-off value 37 U/ml) exhibited a lower degree of sensitivity, with positive values in 23.5% of the effusions due to adenocarcinomas and in 4.5% of the effusions due to benign disease. At a cut-off value of 29 U/ml, CA 15-3 was positive in 49% of fluids from patients with adenocarcinoma and in 3.0% of the benign fluids. The CA 125 RIA failed to show any specificity using the established cut-off value of 35 U/ml, with approximately 80% of all the effusions giving positive results. The specificity of the assay was increased by using a cut-off value of 3000 U/ml, but with a substantial loss in sensitivity (23.5%). Using a combination of the CA 72-4 and CEA RIAs the sensitivity for malignant effusions was increased to 73.5%. No additional improvement in the overall sensitivity was observed when using the CA 72-4 assay in combination with assays for the other markers, except in the case of 1 effusion. We conclude that the CA 72-4 RIA, possibly in combination with other assays such as CEA, may be useful in distinguishing between adenocarcinomatous and benign effusions.

CA 125 in peritoneal fluid and serum from patients with benign gynecologic conditions and ovarian cancer

CA 125 was measured in peritoneal fluid from 200 patients with primary ovarian malignancies (35) and benign gynecologic conditions (165). In 86 patients CA 125 was measured both in peritoneal fluid and in serum. Patients with ovarian cancer had markedly greater serum CA 125 levels compared to patients with benign disease. CA 125 levels in peritoneal fluid were usually higher than serum levels. Twenty-six (93%) of 28 patients with ovarian cancer had peritoneal fluid levels which exceeded serum levels in paired samples. peritoneal fluid CA 125 values greater than 200 U/ml identified ovarian cancer patients with 96% sensitivity and 99% specificity. Serum CA 125 values greater than 35 U/ml identified ovarian cancer patients bearing ascites with a sensitivity of 99% and specificity of 94%. Only 2 of 165 patients with benign gynecological conditions had peritoneal fluid values above 200 U/ml. By contrast, only two values below 200 U/ml were found in ascitic fluids from 35 patients with ovarian cancer. CA 125 levels in peritoneal fluid deserve further evaluation for follow-up of patients with ovarian cancer.

Immunoscintigraphy using CA 125 antibodies in the management of ovarian cancer

Radioimmundetection (RID) using anti-CA 125 antibodies proved to be a valuable tool in the follow-up of metastasizing ovarian cancer. Sensitivity, specificity, and accuracy were high. RID had no clinical side effects. But some patients formed antibodies which interfered a) with the evaluation of the scintigram and b) with further measurement of CA 125 levels. We found 2 cm diameter metastases that were not detected by computed tomography. However, the heterogeneity of tumor metastases limits the sensitivity of this method. CA 125 serum levels, immunohistochemistry, and immunoscintigraphy did not always correlate. Monitoring serum levels of CA 125 was most valuable in clinical management of tumor spread and in the optimal use of RID.

Immunohistochemical localization and tissue levels of tumor-associated glycoproteins CA 125 and CA 19-9 in the decidua and fetal membranes at various gestational ages

To investigate the sources and biologic significance of CA 125 and CA 19-9 in amniotic fluid, immunohistochemical and biochemical localization of these tumor-associated glycoproteins in the decidua and fetal membranes was studied. Immunohistochemically, CA 125 and CA 19-9 were localized in the cytoplasm of decidua cells and amnion epithelial cells but not in the chorion and placental tissue. Biochemically, the 12,000 X g supernatant fractions of decidua and amnion tissues contained relatively large amounts of CA 125 and CA 19-9, 73% to 96% of which was present in the cytosolic fractions of these tissues. The CA 125 levels in the amniotic fluid decreased, whereas those of CA 19-9 increased with gestation, which correlated well with the respective levels in amnion tissues. These findings suggest that amnion cells produce and secrete these glycoproteins into the amniotic cavity. However, it is possible that the decidua also secretes CA 125 into the amniotic cavity through the chorion and amnion such as in the case of prolactin.

Purification and immunochemical characterization of ascitic fluid glycoproteins containing certain tumor-associated and blood group antigen markers

Ascitic fluids from patients with various types of cancer were screened for the CA 19-9 and CA 125 tumor-associated antigenic activities. Two fluids exhibiting the highest activities were tested for their binding to various lectin-Sepharose columns resulting in both being bound best to wheat germ agglutinin (WGA) Sepharose. The WGA column eluate of one fluid was further chromatographed by HPLC and three peaks were obtained with approximate molecular weights of 3.65 MDa, 664 kDa and 330 kDa, of which only the largest fraction contained the CA 19-9 activity. The fluids were also fractionated on a Sephacryl S-400 column with most of the activity being present in or near the void volume. Monoclonal antibodies were used to demonstrate that the purified glycoproteins also contained the blood group A determinant, the four Lewis determinants Le(a), Le(b), Le(x) and Le(y), and the sialylated-Le(x) determinant, while other antibody analyses failed to detect other blood group and/or carbohydrate sequence determinants. Some of the blood group expressions could be separated from the CA 19-9 and CA 125 active glycoproteins by adsorption with various lectins other than the WGA.

Detection of malignant pleural effusions by tumor marker evaluation

Cytologic examination and determination of tumor markers (PHI, LDH, alpha-1-glycoprotein, alpha-2-HS-glycoprotein, beta 2-microglobulin, ferritin [corrected], sialic acid, IgE, fetoprotein, CEA, beta HCG and beta 1-SP-glycoprotein) were carried out in pleural fluid samples obtained from 70 patients with suspected neoplasia. Tumor markers were also determined in sera. The protein content of all pleural effusions was greater than or equal to 3 g/dl. Patients were grouped according to diagnosis as follows: (a) 42 with neoplastic diseases (7 mesotheliomas and 19 lung, 4 ovarian, 3 breast and 8 miscellaneous cancers), (b) 22 with benign inflammations and (c) 6 with congestive effusions. Of the parameters examined, only CEA and beta-HCG [corrected] gave information that the effusion was probably malignant. Using 6 ng/ml as cut-off for CEA and 10 mIU/ml for beta HCG, the sensitivity was 57.1% and 45.2%, respectively, specificity was 92.8% for both parameters and test efficiency 0.75 and 0.69, respectively. When CEA and beta HCG were considered together sensitivity increased to 73.8% and efficiency to 0.78. CEA and/or beta HCG were positive in the pleural effusions of 19 of the 20 malignant pleural effusions, all with a negative cytologic examination, which subsequently became positive in 8. Because of their high specificity, these two parameters are a useful tool and can be routinely measured to evaluate pleural effusions of dubious origin, even if CEA and beta HCG cannot, on [corrected] their own, define the primary malignancy.