A comparison of ten serological tumor markers for the detection of gastric cancer.

Gastric cancer comprises only 2% of cancer cases in the United States but represents the most prevalent cancer in less developed countries and the fourth most prevalent cancer world wide. Early diagnosis and therapeutic intervention could radically reduce the number of deaths attributed to this disease. For this reason, minimally invasive cancer specific tests are urgently sought and recently have included the serological tumor markers. The objective of this study was to compare ten tumor antigens (carcinoembryonic antigen [CEA], CA 19-9, CA 195, CA 50, CA 72-4, CA 125, CA 15-3, CA 27.29, alpha-fetoprotein [AFP], and Cyfra 21-1) for their diagnostic efficacy in gastric cancer patients. The assays used in this study included CA 72-4, CA 19-9, CA 15-3, CA 125, CA 27.29, and Cyfra 21-1 from Fujirebio Diagnostics/Centocor Inc., CA 195 and CEA from Hybritech, Inc., CA 50 from CIS bio international, and AFP from Abbott Inc. Sera from 200 healthy adults were used to determine the normal reference intervals. Diagnos tic parameters were determined using sera from 554 patients including 184 with no disease, 11 with non-malignant disease, 12 with gastric cancer, and 347 with other types of cancer. The diagnostic sensitivities included: CA 50 (70%), CA 19-9 (64%), CA 195 (58%), CEA (50%), CA15-3 (45%), CA 125 (40%), CA 27.29 (30%), CA 72-4 (27%), AFP (22%), and Cyfra 21-1 (9%). With the exception of CA 195 and CA 15-3 (75% specificity), all the markers had diagnostic specificities equal to or greater than 80% (range 80-95%). Analytical parameters were evaluated for the assays and compared favorably We concluded that CA 50 was the best tumor antigen for use in the diagnosis of gastric cancer.

Keywords: cancer, gastric cancer, stomach cancer, carcinoembryonic antigen, alpha-fetoprotein, CEA, AFP, CA 50, CA 19-9, CA 195, CA 72-4, CA 125, CA 15-3, CA 27.29, Cyfra 21-1, tumor marker.

**********

With 558,458 estimated new cases and 405,215 estimated deaths world-wide in 2001, gastric cancer is the fourth most prevalent cancer globally. Similarly, in less developed countries gastric cancer ranks first in prevalence and second only to lung cancer for incidence of new cases (World Health Organization, 2001). Originally the number one cause of cancer deaths in the United States, today the incidence and prevalence of gastric cancer have declined drastically, possibly due to the widespread use of refrigeration and antibiotics in the processing of food. This has led to a decreased consumption of salt cured and smoke cured meat and fish which have long been associated with increased risk of gastric cancer (Hossfeld and Sherman, 1990; Key et al., 1998). Additionally, Helicobacter pylori infection is considered to be a predisposing factor for gastric cancer because it can cause chronic atrophic gastritis, resulting in increased gastric pH, bacterial colonization of the stomach, and the production of carcinoge nic N-nitroso compounds from dietary proteins. The decreased incidence of Helicobacter pylori infection in the United States, due to improved sanitation and the use of antibiotics, has paralleled the observed decline in gastric cancer. No comparable decreases of infection rates or gastric cancer incidences have been observed in less developed countries. (Key et al., 1998).

Possible therapeutic methods and strategies include total gastrectomy, radical subtotal gastrectomy, resectioning of involved portions of liver, pancreas, and transverse colon, splenectomy and removal of involved lymph nodes, chemotherapy, and radiotherapy (Hossfeld and Sherman, 1990; National Cancer Institute Symptoms, 2000). The prognosis depends on the extent of tumor spread at the time of initial treatment and is generally better for gastric lymphomas than for carcinomas. The overall five-year survival rate for all patients is approximately 10%. This increases to about 40% for patients who were diagnosed and treated early (Hossfeld and Sherman, 1990; National Cancer Institute Symptoms, 2000).

Traditional methods of gastric cancer diagnosis have included biopsy, barium X-rays, gastroscopy, upper GI series with double contrast media, computer tomography (CAT scans), exfoliative cytology, and gastric cytology following brushing and washing of the stomach (National Cancer Institute Symptoms, 2000; Hossfeld and Sherman, 1990). There is evidence to support the use of serum tumor antigens as an aid in diagnosis, to measure tumor size, and to evaluate post surgical therapeutic methods and the presence of recurrent disease in gastric and other gastrointestinal cancers. (Wu and Nakamura, 1997). CA 72-4 is the principal tumor antigen in current use for the diagnosis and prognosis of gastric cancer. Other markers which have been assessed for gastric cancer include, among others, CA 19-9, CA 50 and CEA, (Wu and Nakamura, 1997). Similarly CAl95 and CA125 have been reported to have some sensitivity for gastric cancer (Hall et al., 1999). CA 19-9, CA 50, and CA195 are markers for a variety of gastrointestinal ca ncers and CA125 is a marker of ovarian cancer. Elevated CA 15-3 has been reported in a variety of adenocarcinomas including breast, lung, ovary, colon, and pancreas. It is principally used in the assessment of breast cancer patients (Lauro et al., 1999). CA27.29 is used as a marker for therapeutic monitoring in breast cancer patients and has not been reported in gastric cancer patients (Gion and Minone, 2001; Frenette et al., 1994). It has been reported in some cases of ovarian, uterine, lung, prostate, colorectal, and pancreatic cancer (Fujirebio Diagnostics, 1998). Elevated alpha-fetoprotein has been extensively used as a marker for hepatic disease, including hepatoma, and for yolk sac derived germ cell tumors. It has also been reported in a few patients with other gastrointestinal cancers (Wu and Nakamura, 1997; Butch et al., 2000). Similarly, Cyfra 21-1 is used as a marker of lung cancer and has not been reported to be useful in diagnosis and monitoring of gastric cancer (Wu and Nakamura, 1997; Hubbard, 1 990).

CA 72-4 is a 1 million kDa mucin-like glycoprotein complex (TAG 72) which is predominantly associated with human adenocarcinoma of the gastrointestinal tract (Johnson et al., 1986; Lan et al., 1987). Two monoclonal antibodies (cc49 and B72.3) have been developed against CA 72-4 (TAG 72) which detect distinct antigenic determinants expressed on the circulating antigen found in a variety of gastrointestinal cancers, and lung cancer (Patterson et al., 1986; Klug et al., 1986). The use of CA 72-4 is recommended in cases of gastric cancer and it has been used in tumor panels (ratio of CAl9-9 to CA72.4) to exclude pancreatic disease (Wu and Nakamura, 1997).

CA 19-9 is a high molecular weight (200-1000 kDa) mucin like glycoprotein which exists as a ganglioside on tumor cells. The expression of this sialylated [Le.sup.a] blood group antigen (sialylated lactoN-fucopentoeose II ganglioside) is required for the expression of CA 19-9 and hence [Lea.sup.a-b-] patients do not express the antigen and can present as false negatives (Steinberg, 1990). A monoclonal antibody was developed against CA 19-9 derived from the SW-1116 human colon carcinoma cell line (Koprowski et al., 1979). CA 19-9 is clinically useful in the detection of pancreatic, colorectal, hepatic, and other gastrointestinal cancers. It has also been described in breast and lung cancer (Wu and Nakamura, 1997). CA 50 is related to CA 19-9 but lacks a fucose residue. Its epitope is the same as that found in [Le.sup.a-b-] (Lewis negative) patients. It has been reported in patients with gastric, colon, and hepatic cancer (Wu, 1996). CA 195 is also related to CA 19-9. It is defined by the mouse monoclonal antib ody CC3C-195 and it recognizes both [Le.sup.a] and sialyl-[Le.sup.a] epitopes. Binding with higher affinity to the sialylated [Le.sup.a] blood group antigen, the antibody can bind to both the sialylated and unsialylated [Le.sup.a] blood group. CA 195 has been reported in pancreatic, colon, and gastric cancers (Wu and Nakamura, 1997).

CA 125 is a 200 kDa glycoprotein expressed by tissue of mullerian duct origin as well as by ovarian tumors. It is defined by the mouse monoclonal antibody OC 125 derived from an ovarian cancer cell line (OVCA 433). It is currently used for detecting epithelial tumors of the ovary. However, it has also been reported in breast, lung, endometrial, and gastrointestinal tumors. It can be elevated with pregnancy and with pelvic inflammatory disease. (Jacobs and Bast, 1989)

CEA is a 150-300 kDa cell surface heterogeneous glycoprotein which is structurally similar to IgG. Abnormally elevated serum levels have been reported in patients with colorectal cancer, breast cancer, and a variety of other carcinomas (Cooper et al., 1979; Reynoso et al., 1972). Additionally, CEA levels can be elevated in heavy smokers and patients with nonmalignant pathologies (Clarke et al., 1982). Consequently, CEA is currently used in therapeutic monitoring and as a diagnostic aid, but is not useful in screening for cancer.

CA 15-3 is a 300-450 kDa glycoprotein defined by two monoclonal antibodies. The 115D8 antibody recognizes human milk fat globule membranes and the DF3 antibody reacts with a breast cancer antigen extract (Kufe et al., 1984; Hilkens et al., 1984). It has been reported in cases of breast, ovarian, pancreatic, lung, and colorectal cancer (Wu and Nakamura, 1997).

CA 27.29 is a mucin antigen defined by the monoclonal antibody B27.29. This antibody recognizes an antigen extracted from ascites fluid derived from patients with breast cancer. CA 27.29 has an epitope that is shared with the DF3 antibody of CA 15-3. (Burtis and Ashwood, 1996). It is currently being marketed as a specific test for breast cancer.

Alpha-fetoprotein (AFP) is a 70,000 kDa glycoprotein which has been isolated from patients with hepatocellular carcinomas and germ cell tumors (Chan et al., 1986). Maternal serum and amniotic fluid APP levels are routinely used for the prenatal diagnosis of open neural tube disease and gastroschisis, and together with karyotyping have been used to diagnose cases of Down's syndrome (Milunsky, 1987; Knight et al., 1988). Alpha-fetoprotein has been reported to be useful in screening for hepatocellular carcinoma in high incidence areas such as Asia, and for classifying and staging germ cell tumors (Chan et al., 1986). Alpha-fetoprotein has been reported in cases of hepatocellular carcinoma, testicular and ovarian germ cell tumors, as well as pancreatic, colorectal, and gastric carcinomas (Butch et al., 2000).

Cyfra 21-1 is a 40 kDa fragment derived from cytokeratin 19. One subgroup of intermediate filament proteins, cytokeratins are found in epithelial cells. The monoclonal antibody recognizes an epitope on the Cyfra 21-1 fragment and is useful in the detection of non-small cell lung cancer, including squamous cell carcinoma of the lung (Pujol et al., 1993). It has also been reported in cases of cervical cancer and other malignancies (Bonfrer et al., 1994; Bodenmuller et al., 1992).

In a clinical laboratory, in order to compare different assay methods one must evaluate their specific performance characteristics (precision, linearity, analytical sensitivity, and analytical specificity) and their clinical performance (normal reference interval and predictive values). Precision is evaluated by assaying replicate samples and determining the mean, standard deviation, and coefficient of variation. Linearity is determined by assaying dilutions of an elevated serum sample and plotting the results and/or performing regression analysis. The minimum detectable concentration of analyte in the test (analytical sensitivity) is determined by assaying replicate samples lacking the analyte (e.g., diluent) and calculating the mean plus two standard deviations. Values falling below this cutoff are presumed to be analyte free. The analytical specificity represents the degree of assay interference from drugs or other chemicals (e.g., bilirubin) present in the specimen. This is not always reported but can be determined by spiking samples with varying concentrations of the suspected interfering drugs/chemicals.

In order to establish a healthy (normal) adult reference interval for the analyte using a particular assay, one calculates the mean plus or minus two standard deviations (95% confidence interval) on assay results from a population set of adults known to be in good health. Subsequently, any patient result which falls within this interval is considered to be "normal" or healthy; whereas, patient results which fall outside (above or below) the limits of this interval are considered to be abnormally elevated or decreased respectively. For tumor markers a low result would have no clinical significance. Therefore, one establishes the cutoff between normal (presumed negative for disease) and abnormal (presumed positive for disease) results by using the mean plus two standard deviations. Predictive validity compares the ability of a new test method to accurately diagnose/predict the presence or absence of disease with that of an established method. Predictive value results include diagnostic sensitivity and specific ity, diagnostic efficiency, and positive and negative predictive values. For the calculation of predictive values, one compares the test results with the "true results" as defined by an external test method considered to be the reference test method. For example one could compare the results of a tumor antigen assay (test results) with those obtained by the physician with histologic analysis of biopsy material (true results). Individual patient assay results are then assigned to one of four categories (true positives [TP], true negatives [TN], false positives [FP], or false negatives [FN]) from which the predictive values are derived. Predictive values include: (a) diagnostic sensitivity (% of individuals with the disease who test positive by the assay), i.e., [100 TP/(TP + FN)], (b) diagnostic specificity (% of individuals without the disease who test negative by the assay), i.e., [100 TN/(TN + FP)], (c) diagnostic efficiency (% of all test results that are either true positives or true negatives), i.e., [10 0 (TP + TN)/(TP + TN + FP + FN)], (d) positive predictive value (% of all positive test results that are true positives), i.e., [100 TP/(TP + FP)], and (e) negative predictive value (% of all negative test results that are true negatives), i.e., [100 TN/(TN + FN)].

The purpose of this study was to evaluate the analytical and clinical performances often serologic tumor marker tests (CEA, CA 19-9, CA 195, CA 50, CA 72-4, CA 125, CA15-3, CA 27.29, AFP, and Cyfra 21-1) for the detection of gastric cancer. Particular attention was paid to the comparison of their diagnostic sensitivities as this value reflects the tumor marker test's ability to detect the disease. A working hypothesis that CA 72-4 would prove to be superior to the other tumor markers was developed based on reports in the literature of its superiority (Wu and Nakamura, 1997; Spila et al., 1996).

MATERIALS AND METHODS

Assays--All assays were performed according to the directions supplied by the manufacturers. The [Tandem.sup.R]-E CEA assay (Hybritech, Inc) is a solid phase two-site immunoenzymometric assay (ELIZA) utilizing two monoclonal IgG antibodies directed against unique sites on the CEA antigen. This assay was quantitated spectrophotometrically using the Photon Immunoassay Analyzer[TM] from Hybritech, Inc. The [Tandem.sup.R]- CA 195/Hybri C Mark[TM] assay (Hybritch Europe, Inc.) is a solid phase two-site immunoradiometric assay (CA 195) (RIA) utilizing monoclonal IgM antibodies developed against the Lewis A (blood group determinant) and sialyated Lewis A epitopes on the CA 195 antigen. This assay was measured using a Genesys[TM] 5000 gamma counter (Laboratory Technologies, Inc.). The Cento[cor.sup.R] CA l9-9[TM] assay (Fujirebio Diagnostics, Inc./Centocor, Inc.) is a solid phase radioimmunoassay (CA 19-9) (RIA) using the 1116-NS-19-9 antibody for both the capture and tracer antibodies. This antibody is directed aga inst an epitope which is biochemically related to the Lewis A determinant; the assay was quantitated using a Genesys[TM] 5000 gamma counter (Laboratory Technologies, Inc.). The RIA-[gnost.sup.R] CA-50 assay (CIS bio international) is a solid phase two-site immunoradiometric assay (CA 50) (RIA) utilizing monoclonal mouse antibodies directed at two carbohydrate chains (sialylated Lewis A and sialylated lactotetraose) of the adenocarcinoma cell line Cob 205. The assay was measured using a Genesys[TM] 5000 gamma counter (Laboratory Technologies, Inc.). The [Centocor.sup.R] CA 72-4[TM} assay (Fujirebio Diagnostics, Inc./Centocor, Inc.) is a solid phase radioimmunoassay (CA 72-4) (RIA) based on two monoclonal antibodies, cc49 and B72.3, which react with distinct antigenic determinants on a tumor associated glycoprotein TAG 72. The antigen was quantitated using the Genesys[TM] 5000 gamma counter (Laboratory Technologies, Inc.). The [Centocor.sup.R] CA l25[TM] assay (Fujirebio Diagnostics, Inc./Centocor, Inc.) is a s olid phase two-site immunoradiometric assay (CA 125) (RIA) using two mouse monoclonal antibodies, 0C125 directed against the OVCA 433 ovarian cancer cell line and a second antibody directed against another CA 125 epitope. The assay was measured using a Genesys[TM] 5000 gamma counter (Laboratory Technologies, Inc.). The [Centocor.sup.R] Cyfra[TM] 21-1 assay (Fujirebio Diagnostics, Inc./Centocor, Inc.) is a solid phase immunoradiometric assay (RIA) utilizing two mouse monoclonal antibodies, KS19.1 and BM19.21, to detect cytokeratin 19 fragments in serum. The assay was quantitated using a Genesys[TM] 5000 gamma counter (Laboratory Technologies, Inc.). The [Centoco.sup.R] CA [15-3.sup.R] assay (Fujirebio Diagnostics, Inc./Centocor, Inc.) is a solid phase radioimmunoassay (RIA) using the 11 5D8 murine monoclonal antibody as the capture antibody and the [I.sup.125] labeled DF3 murine monoclonal antibody as the tracer. This assay was quantitated using an Iso [Data.sup.R] gamma counter. The [Truquant.sup.R] BR[TM] as say (Fujirebio Diagnostics, Inc./Centocor, Inc.) is a solid phase competitive inhibition radioimmunoassay (competitive RIA) using polystyrene tubes coated with CA 27.29 antigen and [I.sup.125] labeled murine monoclonal B27.29 antibody. This assay was quantitated using an Iso [Data.sup.R] gamma counter. The [IMx.sup.R] APP assay (Abbott Laboratories, Inc.) is a microparticle enzyme immunoassay (MEIA) utilizing two monoclonal antibodies directed against unique sites on the AFP antigen. This assay was quantitated using the [IMx.sup.R] Automated Analyzer from Abbott Laboratories, Inc. All dilutions were performed using diluent supplied by the manufacturers in the assay kits. These diluents contain physiological concentrations of protein which maintains the sample protein concentration within limits which do not affect the assay. Regression analysis was used to determine the linearity of the assays and the independent t-test was used to compare male and female subjects when developing the reference intervals. Stat istical analysis was performed using SPSS software.

Patients--Procedures used in this study were in accord with ethical standards established by the University of Southern Mississippi (USM). Permission for the study was granted by the USM Human Subjects Protection Review Committee (HSPRC/IRB).

All study participants were selected from patients seen in an area hospital. Five hundred and fifty four patients were randomly chosen and the assays were run in a blind fashion. Blood samples were collected using appropriate aseptic technique. Following serum separation aliquots were coded and frozen at -20[degrees]C. Subsequently, aliquots were thawed at 37[degrees]C and assayed in duplicate (sample permitting) for the tumor antigens. The diagnoses were obtained from the attending physicians and were based on pathological examination. Patient Classifications included (a) no known disease, (b) nonmalignant disease, (c) non gastric cancer, and (d) gastric cancer. Cancer patients were classified according to the primary site of the tumor, regardless of the presence or absence of metastases. Since available information on patient therapy was incomplete, statistical analyses were performed on the total patient pool without reference to this.

The normal control subjects were healthy males (100) and females (100) ranging from 18-65 years of age. Their blood samples were collected and processed in the same manner as the patient samples.

RESULTS

Precision and Linearity--Quality control samples analyzed over a 6 month period were used to determine intra- and inter-assay precision. The within-run coefficient of variation (%CV) was less than 10% for all but the CA 15-3 assay which was somewhat higher (20%) (Table 1). Similarly the between-run coefficient of variation was less than 17% for each of the assays (Table 2). Serial dilutions of abnormal pool samples exhibited good linearity (Fig. 1) with [R.sup.2] values equal to or greater than 0.989 for all the assays.

Reference Intervals--The minimum detectable concentration was determined by analyzing approximately 20 replicates of the zero calibrator/diluent and establishing the mean + 2SD as the cut-off value (Table 3 a). The normal adult reference intervals were established by determining the 95% confidence intervals for healthy control male and female subjects. The intervals (Tables 3a, 3b) were broader than those reported by the manufacturer for all but the CA 125, CA 72-4, CA 27.29, and AFP assays which were somewhat narrower. There was no significant difference between healthy adult males and females for any of the assays except CA 19-9, where the males were significantly (p <0.05) higher.

Diagnostic Parameters--In this study there were 184 patients without disease, 11 patients with nonmalignant disease, 12 patients with gastric cancer, and 347 patients with other types of cancer including: pancreatic, small intestinal, esophageal, lung, breast, ovarian, prostatic, renal, colorectal, gallbladder, hepatic, cecal, uterine, testicular, head and neck, leukemia, lymphoma, and all other types. Patients' diagnoses were made by the attending physicians and were predicated on a variety of pathologic findings including the histologic analysis of biopsy or surgical tissue. For purposes of this study, patients with gastric cancer were considered to be positive for disease. Similarly, cutoffs between normal (negative) and abnormal (positive) test results used were those listed by the manufacturers and are cited in Table 4. In Table 4, the diagnostic sensitivity of CA 50 (70.0%) is superior to that of the other markers (CA 19-9, 63.6%; CA 195, 58.3%; CEA, 50.0%; CA 15-3, 45.5%; CA 125, 40.0%; CA 27.29, 30.0 %; CA 72-4, 27.3%; AFP, 22.2%; and Cyfra 21-1, 9.1%). The diagnostic specificities of the ten assays range from 75-95% with Cyfra 21-1 having the highest value. The negative predictive and positive predictive values range from 97-99% and 3-9% respectively. The efficiency of the Cyfra 21-1 assay was the best (92.6%), presumably due to the fact that it had the highest % specificity.

DISCUSSION

The incidence and prevalence of gastric cancer make it an important medical problem world wide. For some time the medical community has sought a minimally invasive, inexpensive, and early diagnostic test for this and other types of cancer. With the exception of PSA in prostate cancer, tumor markers have generally not proven useful as screening tests either because their incidence is too low in the general public, or because the cutoff between benign and malignant disease is not sufficiently precise. Thus increased concentrations have been reported in some cases of benign disease while not observed in cases of in situ cancer when the prognosis is best (Wu and Nakamura, 1997; Roulston and Leonard, 1993). Despite this, many tumor antigens have proven useful for diagnosis and for therapeutic monitoring and the detection of recurrent disease.

In this study we compared ten serologic assays (CEA, CA 19-9, CA 195, CA 50, CA 72-4, CA 125, CA 15-3, CA 27.29, AFP, and Cyfra 21-1) for their efficacy at detecting gastric cancer. The within-run and between-run precision was slightly higher for CA 15-3 and CA 195 than for the other assays, but all values were below 20%. The linearity was excellent for all the assays. The minimum detectable concentration of analyte (zero calibrator/diluent mean + 2SD) was slightly higher for CA 125 than for the other assays. This test was therefore repeated using a patient sample that had previously given a result of 0 U/mL (data not shown). The results did not differ from those of the zero calibrator/diluent, confirming its value. The normal reference intervals were broader than those cited by the manufacturers for all the assays except CA 125, CA 72-4, CA 27.29, and AFP. The CA 19-9 assay exhibited a significantly higher reference interval for males than for females; otherwise there were no significant differences between the sexes. The assays compared favorably for cost and availability of instrumentation. With the exception of CEA and AFP, all of the assays were radiolabeled ([I.sup.125]) and therefore had shorter shelflives. The turnaround time varied from 1 hour for AFP (automated assay) to approximately 3-24 hours for the other assays (manual assays with varying incubations periods). The CEA (ELIZA assay) required only the use of a spectrophotometer and therefore might be more attractive than the other assays for use in a small lab.

Sera from 554 patients seen in a local hospital were assayed for ten tumor antigens and the diagnostic parameters were compared. The physicians' diagnoses and the manufacturers suggested cutoff values were utilized to assign the test results to the categories of true or false positives and negatives. Predictive values were calculated for gastric cancer. The most important finding of this study was the observation that CA 50 was clearly superior to CA 72-4 for the detection of gastric cancer, exhibiting a diagnostic sensitivity of 70% as compared to 27%. Similarly, CA 19-9, CA 195, CEA, CA 15-3, and CA 125 all excelled when compared to CA 72-4. The importance of this stems from the fact that CA 72-4 has been reported to be the best tumor marker for gastric cancer and is currently being marketed as a gastric/gastrointestinal cancer marker. Since CA 50, CA 195, and CA19-9 share very similar epitopes, it should not be surprising that all three react similarly with gastric as well as with other carcinomas. Similarly, CEA shares some antigenic determinants with CA 19-9 (Wu and Nakamura, 1997).

In a similar study, Pectasides et al. (1997), found CA 50 and CA 19-9 to be superior to CEA for the diagnosis of gastric cancer. Haglund et al. (1992) investigated CA 19-9 and CA 50 for their diagnostic capabilities and found them to have the same sensitivity for gastric cancer.

In two studies the authors reported a discrepancy between the markers depending on the stage of the cancer. In a study involving 100 cancer patients (44 with early cancer and 56 with advanced cancer), Kodama et al. (1995) reported that in advanced cancer CA 72-4 was superior to CEA and CA 19-9 for the diagnosis, prognosis, and detection of recurrent disease. By contrast they found CA 19-9 and CEA to be better for the detection of early stage (I and II) disease. Likewise, in a study by Van-Dalen and Kessler (1996) in which 4266 serum samples from 23 labs were analyzed for CEA, CA 15-3, CA 19-9, CA 72-4, CA 125, Cyfra 21-1, and AFP, the authors reported that CA 72-4 was the most sensitive for stage IV disease. However, the authors found CA 72-4, CA 19-9, and CEA to be equally sensitive for stage I-III disease.

By contrast, in a study of 242 patients, Spila et al. (1996) found that CA 72-4 was superior to both CEA and CA 19-9 for the diagnosis and prognosis of both primary and recurrent gastric cancer. Likewise, Fernandez-Fernandez et al. (1996) have reported that in a study of 167 patients with gastric cancer and 92 patients with benign disease they found CA 72-4 to be superior to both CA 19-9 and CEA at all stages of disease. Discrepancies between their results and ours could be the result of genetic differences in the patient populations, the stage of the tumors, the presence of pathologic complications, the prevalence of disease (gastric cancer) in the population sample, and/or the use and type(s) of therapies. Since CA 199, CA 195, CA 50, and CA 72-4 are blood group antigen type carbohydrate markers and CEA contains incomplete blood group substances, it is not surprising that patients who do not express a particular blood group antigen will have serum which does not react in tumor marker assays that use monoclo nal antibodies directed at epitopes found on these antigens (Wu and Nakamura, 1997). Thus the genetic background of a patient could cause false negative values with these tests. The greater the tumor burden and the more metastatic it has become, the greater the likelihood of increased levels of antigen and hence of positivity with a particular antigen assay. Both positive and negative predictive values are somewhat dependent on the disease prevalence in the sample population (Cembrowski et al., 2000). For this reason many studies are designed to include increased numbers of patients with the disease being studied (high prevalence), and to exclude any patients with other diseases. While this would lead to better (higher) predictive values, it doesn't reflect the local patient population. In this study, patients were randomly selected and included therefore only 12 gastric cancer patients and numerous patients with other types of cancer and with no cancer. This better represents what is actually seen in America n hospitals but could introduce a bias if the disease cohort shares some unique feature(s). The gastric cancer patients' sera were collected prior to surgery and chemotherapy but there is limited data about any medications they may have been using which could have interfered with the assay. Regretfully that information is not available at this time.

In conclusion, ten assays (CEA, CA 19-9, CA 195, CA 50, CA 72-4, CA 125, CA 15-3, CA 27.29, AFP and Cyfra 21-1) were evaluated for their efficacy at diagnosing gastric cancer. CA 50 proved to be superior to the other assays with CA 19-9, CA 195, and CEA also proving effective. In contrast to previous studies, our results did not support the use of CA 72-4 for the diagnosis of gastric cancer and therefore our hypothesis was rejected.

[FIGURE 1 OMITTED]

Table 1

Within-run precision for CEA, CA 19-9, CA 195, CA 50, CA 72-4, CA 125,
CA 15-3, CA 27.29, AFP, and Cyfra 21-1.

Sample                   n           Mean  SD      %CV

CEA Low Control          43    4.28 ng/mL  0.29   6.78
CEA High Control         40   64.04 ng/mL  2.79   4.36
CA 19-9 Low Control      20   39.66 U/mL   2.18   5.51
CA 19-9 High Control     20   76.28 U/mL   4.79   6.28
CA 195 Low Control       30   11.60 U/mL   1.10   9.53
CA 195 Mid Control       30   52.30 U/mL   3.55   6.80
CA 195 High Control      30   79.40 U/mL   7.24   9.13
CA 50 Low Control        20   12.78 U/mL   0.58   4.54
CA 50 High Control       20  100.45 U/mL   4.18   4.16
CA 72-4 Low Control      20    9.24 U/mL   0.74   8.05
CA 72-4 High Control     20   69.66 U/mL   3.57   5.13
CA 125 Low Control       20   55.16 U/mL   3.48   6.31
CA 125 High Control      20  101.39 U/mL   6.38   6.29
CA 15-3 Control          50   46.83 U/mL   9.60  20.50
CA 27.29 Control I       42   75.36 U/mL   6.61   8.77
CA 27.29 Control II      37  106.51 U/mL   9.93   9.32
AFP Low Control          10   20.36 ng/mL  2.22  10.90
AFP Medium Control       10   77.87 ng/mL  3.16   4.06
AFP High Control         10  171.22 ng/mL  4.96   2.90
Cyfra 21-1 Low Control   20    4.41 nglmL  0.28   6.27
Cyfra 21-1 High Control  20   14.17 ng/mL  0.77   5.41
Table 2

Between-run precision for CEA, CA 19-9, CA 195, CA 50, CA 72-4, CA 125,
CA 15-3, CA 27.29, AFP, and Cyfra 21-1.

Sample                  n        Mean        SD    %CV

CEA Low Control         76    4.44 ng/mL   0.37   8.33
CEA High Control        72   62.64 ng/mL   3.40   5.43
CA 19-9 Low Control     59    44.57 u/mL   4.33   9.72
CA 19-9 High Control    59    84.85 U/mL   8.65  10.19
CA 195 Low Control      62    11.67 U/mL   1.88  16.11
CA 195 Mid Control      58    52.03 U/mL   4.81   9.25
CA 195 High Control     62    80.68 U/mL  10.39  12.88
CA 50 Low Control       57    12.87 U/mL   0.86   6.68
CA 50 High Control      57   105.46 U/mL   7.73   7.33
CA 72-4 Low Control     65     9.57 U/mL   0.71   7.37
CA 72-4 High Control    66    71.17 U/mL   3.57   5.01
CA 125 Low Control      86    54.08 U/mL   5.50  10.17
CA 125 High Control     86   107.11 U/mL   8.14   7.56
CA 15-3 Control         67    45.21 U/mL   6.61  14.62
CA 27.29 Control I      73    74.99 U/mL   6.95   9.27
CA 27.29 Control II     68   117.76 U/mL  16.38  13.91
AFP Low Control         38   19.60 ng/mL   1.44   7.35
AFP Medium Control      38   78.15 ng/mL   3.88   4.96
AFP High Control        38  167.01 ng/mL   6.28   3.76
Cyfra 21-1 Low Control  78    4.45 ng/mL   0.50  11.23
Cyfra 21-1 HighControl  76   13.97 ng/mL   0.86   6.16
Table 3a

Reference intervals for CEA, CA 19-19 CA 195, CA 50, CA 72-4, CA 125, CA
15-3, CA 27.29, AFP, and Cyfra 21-1.

Sample                 n       Mean      SD     Range

Zero/Diluent Conrols

     CEA               20  0.00 ng/mL   0.35  0.00-00.70
     CA 19-9           20   0.00 U/mL   0.70  0.00-01.40
     CA 195            20   0.00 U/mL   1.50  0.00-03.00
     CA 50             20   0.08 U/mL   0.12  0.00-00.32
     CA 72-4           20   2.93 U/mL   0.36  2.21-03.64
     CA 125            20   3.20 U/mL   1.44  0.40-06.00
     CA 15-3           21   0.02 U/mL   0.08  0.00-00.18
     CA 27.29          24   0.24 U/mL   1.16  0.00-02.56
     AFP               13  0.00 ng/mL   0.01  0.00-00.02
     Cyfra 21-1        20  0.01 ng/mL   0.03  0.00-00.07

Health Adults

     CEA              264  2.82 ng/mL   2.64  0.00-08.10
     CA 19-9          199  16.01 U/mL  15.53  0.00-47.08
     CA 195           230   4.96 U/mL   6.58  0.00-18.11
     CA 50            200  14.93 U/mL  13.81  0.00-42.55
     CA 72-4          200   1.32 U/mL   1.09  0.00-03.50
     CA 125           200  10.60 U/mL   8.58  0.00-27.76
     CA 15-3          214  24.71 U/mL  14.00  0.00-52.72
     CA 27.29         200  17.74 U/mL   7.42  2.90-32.58
     AEP              214  3.60 ng/mL   1.93  0.00-07.46
     Cyfra 21-1       200  1.00 ng/mL   1.90  0.00-04.80
Table 3b

Reference intervals for CEA, CA 19-19 CA 195, CA 50, CA 72-4, CA 125, CA
15-3, CA 27.29, AFP, and Cyfra 21-1.

Sample                  n      Mean       SD     Range

Health Adultt Males

     CEA               133  3.08 ng/mL   2.36  0.00-07.80
     CA 19-9            99  18.73 U/mL  18.67  0.00-56.07
     CA 195            121   5.07 U/mL   6.50  0.00-18.07
     CA 50             100  14.84 U/mL  15.30  0.00-45.44
     CA 72-4           100   1.41 U/mL   0.91  0.00-03.23
     CA 125            100  10.44 U/mL   8.26  0.00-26.95
     CA 15-3           106  25.36 U/mL  13.92  0.00-53.20
     CA 27.29          100  18.94 U/mL   8.28  0.00-33.50
     AFP               107  3.47 ng/mL   1.79  0.00-07.05
     Cyfra 21-1        100  1.02 ng/mL   2.06  0.00-05.13

Healthy Adult Females

     CEA               131  2.55 ng/mL   2.89  0.00-08.33
     CA 19-9           100  13.33 U/mL  11.08  0.00-35.49
     CA 195            109   4.83 U/mL   6.69  0.00-18.21
     CA 50             100  15.02 U/mL  12.22  0.00-39.46
     CA 72-4           100   1.23 U/mL   1.25  0.00-03.72
     CA 125            100  10.77 U/mL   8.93  0.00-28.62
     CA 15-3           108  24.08 U/mL  14.12  0.00-52.32
     CA 27.29          100  16.54 U/mL   6.28  3.98-29.10
     AFP               107  3.73 ng/mL   2.06  0.00-07.85
     Cyfra 21-1        100  0.99 ng/mL   1.73  0.00-04.45
Table 4

Comparison of predictive values of CEA, CA 19-9, CA 195, CA 50, CA 72-4,
CA 125, CA 15-3, CA 27.29, AFP, and Cyfra 21-1 for gastric cancer.

            Sensitivity  Specificity  Predictive  Predictive  Efficiency
                 %            %       Value (+) %  Value (-) %     %

CEA            50.0         80.1         5.3         98.6        79.4
(n = 554)
CA 19-9        63.6         87.0         9.2         99.1        86.5
(n = 541)
CA 195         58.3         75.1         4.9         98.8        74.7
(n = 554)
CA 50          70.0         84.4         8.1         99.3        84.1
(n = 515)
CA 72-4        27.3         90.4         5.5         98.4        89.1
(n = 550)
CA 125         40.0         91.1         8.0         98.7        90.1
(n = 527)
CA 15-3        45.5         75.0         3.8         98.4        74.4
(n = 515)
CA 27.29       30.0         81.2         3.2         98.2        80.2
(n = 494)
AFP            22.2         86.9         3.4         98.2        85.7
(n = 418)
Cyfra 21-1      9.1         94.5         3.4         97.9        92.6
(n = 516)

             Cutoff
              Value

CEA         5.0 ng/mL
(n = 554)
CA 19-9     37.0 U/mL
(n = 541)
CA 195      10.5 U/mL
(n = 554)
CA 50       25.0 U/mL
(n = 515)
CA 72-4      5.6 U/mL
(n = 550)
CA 125      35.0 U/mL
(n = 527)
CA 15-3     35.0 U/mL
(n = 515)
CA 27.29    37.7 U/mL
(n = 494)
AFP         8.9 ng/mL
(n = 418)
Cyfra 21-1  4.8 ng/mL
(n = 516)

ACKNOWLEDGMENTS

The authors thank Ms Jan Oglesby (Kessler AFB) and Mr: Allen Keely (Fujirebio Diagnostics Inc.) for their assistance. This study was partially supported by the Aubrey Keith and Ella Ginn Lucas Research Award. Fujirebio Diagnostics Inc/Centocor Inc., Hybritech Inc., and Abbott Laboratories donated the reagents. The gift of sera from Keesler AFB Medical Center and Forrest General Hospital is gratefully acknowledged.

LITERATURE CITED

Bodenmuller, H., P. Stieber, D. Banauch, U. Hasholzner, A. Dessauer, and B. Ofenloch-Hahnle. 1992. Enzymum test Cyfra 21-1: A new marker for NSCLC, method evaluation and first experiences. [Abstract] Clin. Chem. 38:966-970.

Bonfrer, J.M.G., K.N. Gaarenstroom, and G.G. Kenter. 1994. Prognostic significance of serum fragments of cytokeratin 19 measured by Cyfra 21-1 in cervical cancer Gynecol. Oncol. 55:371-375.

Burtis, C.A., and E.R. Ashwood. 1996. Tumor markers. Pages 336-350 in C.A. Burtis and E.R. Ashwood eds. Tietz Fundamentals of Clinical Chemistry 4th Edition WB.Saunders Co., Philadelphia, PA.

Butch, A.W., N.A. Massoll, and A.A. Pappas. 2000. Tumor markers. Pages 523-540 in M.L. Bishop, J.L. DubenEngelkirk, and E.P. Fody eds. Clinical Chemistry Principles, Procedures, Correlations 4th Edition Lippincott Williams & Wilkens, Philadelphia, PA.

Chan, D.W, M. Kelsten, R. Rock, and D. Bruzek. 1986. Evaluation of a monoclonal immunoenzymometric assay for alpha-fetoprotein. Chin. Chem. 32:1318-1322.

Cembrowski, G.S., A.M. Sullivan, and T.L. Hofer. 2000. Quality control and statistics. Pages 40-76 in M. Bishop, J.L. Duben-Engelkirk, and E.P. Fody eds. Clinical Chemistry, Principles, Procedures, and Correlations 4th Edition, Lippincott Williams & Wilkins, Philadelphia, PA.

Clarke, C.A., T.P. Whitehead, and A.G.W Whitfield. 1982. Carcinoembryonic antigen and smoking. J Royal College of Phys. London 16 (2):112-113.

Cooper, M.J., C.R. Mackie, D.B. Skinner, and A.R. Moossa. 1979. A reappraisal of the value of carcinoembryonic antigen in the management of patients with various neoplasms. British J. Surg. 66:120-123.

Fernandez-Fernandez, L., E. Tejero, A. Tieso, L. Rabadan, M. Munoz, and I. Santos. 1996. Receiver operating characteristic (ROC) curve analysis of the tumor markers CEA, CA 19-9, and CA 72-4 in gastric cancer Int. Surg. 81 (4):400-402.

Frenette, P.S., M.P. Thirlwell, M. Trudeau, D.M. Thomson, L. Joseph, and J.S. Shuster. 1994. The diagnostic value of CA 27.29, CA 15-3, mucin-like carcinoma antigen, carcinoembryonic antigen, and CA 19-9 in breast cancer and gastrointestinal malignancies. Tumour Biology 15 (5):247-254.

Fujirebio Diagnostics, Inc. Truquant[R] BR[TM] Radioimmunoassay (CA 27.29) Package Insert.#014-395, 06/98.

Gion, M., and R. Minone. 2001. CA 27.29: A valuable marker for breast cancer management; a confirmatory multicentric study on 603 cases. Eur. J. Cancer 37 (3):355-363.

Haglund, C., P.J. Roberts, H. Jalanko, and P. Kuusala. 1992. Tumour markers CA 19-9 and CA 50 in digestive tract malignancies. Scand. J. Gastroenterol. 27 (3):169-174.

Hall, M., M. Jackson, J.T. Hohnson, H. Schultze, D. Fortenberry, C. Bright, H. Hua, J. Ying, and P. Sykes. 1999. A comparison of seven tumor antigens for the serodiagnosis of gastrointestinal cancers. The Chemist 76 (5):9-22.

Hilkens, J., F. Buijs, and J. Hilgers. 1984. Monoclonal antibodies against human milk-fat globule membranes detecting differentiation antigens of the mammary gland and its tumor. International 3. of Cancer 34: 197-206.

Hossfeld, D.K., and C.D. Sherman. 1990. International Union Against Cancer: Manual of Clinical Oncology, 5th ed. Springer-Verlag, New York, NY. 389 pp.

Hubbard, E. 1990. Tumor markers. Diagnostic and Clinical Testing 28:13-26.

Jacobs, I., and R.C. Bast. 1989. The CA 125 tumor-associated antigen: A review of the literature. Hum. Reprod. 4:1-12.

Johnson, B.G., J. Schlom, A.J. Peterson, J. Bennett, J.L. Magnani, and D. Colcher. 1986. Analysis of a human tumor-associated glycoprotein (TAG 72) identified by monoclonal antibody B72.3. Cancer Res. 46 (2):850-857.

Key, T., D. Forman, and M.C. Pike. 1998. Epidemiology of cancer. Pages 34-59 in L.M. Franks and N.M. Teich, eds. Introduction to the cellular and molecular biology of cancer, 3rd ed. Oxford University Press, Oxford, England.

Klug, T.L., M.A. Sattler, D. Colcher, and J. Schlom. 1986. Monoclonal antibody immunoradiometric assay for an antigenic determinant (CA 72) on a novel pancarcinoma antigen. Int. J. Cancer 38 (5):661-669.

Knight, G.J., G.E. Palomaki, and J.E. Haddow. 1988. Use of maternal serum alpha-fetoprotein measurements to screen for Down's syndrome. Clin. Obstet. Gynecol. 31:306-327.

Kodama, I., K. Koufuji, S. Kawabata, S. Tetsu, Y. Tsuji, J. Takeda, and T. Kakegawa. 1995. The clinical efficacy of CA 72-4 as serum marker for gastric cancer in comparison with CA 19-9 and CEA. Int. Surg. 80 (1):45-48.

Koprowski, H.Z., K. Steplewski, and M. Mitchell. 1979. Colorectal carcinoma antigens detected by somatic hybridoma antibodies. Cell Genet. 5:957-972.

Kufe, D., G. Inghirami, and M. Abe. 1984. Differential reactivity of a monoclonal antibody (DF3) with human malignant versus benign breast tumors. Hybridoma 3:223-232.

Lan, M.S., R.C. Bast, M.I. Colnagi, R.C. Knapp, D. Colcher, J. Schlom, and R.S. Megzgar. 1987. Co-expression of human cancer-associated epitopes on mucin molecules. Int. J. Cancer 39 (l):68-72.

Lauro, S., L. Trasatti, and F. Bordin. 1999. Comparison of CEA, MCA, CA 15-3 and CA 27.29 in follow-up and monitoring therapeutic response in breast cancer patients. Anticancer Res. 19(4C):3511-3515.

Milunsky, A. 1987. The prenatal diagnosis of neural tube and other congenital defects. Pages 300-306 in A. Milunsky, ed. Genetic disorders in the fetus: Diagnosis, prevention, and treatment. Plenum Press, New York, NY.

National Cancer Institute Symptoms. 15 June 2000. http:\cancernet.nci.nih.gov/wynth_pubs/stomach.htm#1

Paterson, A.J., J. Schlom, H.F. Sears, J. Bennett, and D. Colcher. 1986. A radioimmunoassay for the detection of a human tumor-associated glycoprotein (TAG 72) using monoclonal antibody B72.3. Int. J. Cancer 37 (5):659-666.

Pectasides, D., A. Mylonakis, M. Kostopoulou, M. Papadopoulou, D. Triantafillis, J. Varthalitis, M. Dimitriades, and A. Athanassiou. 1997. CEA, CA 19-9, and CA 50 in monitoring gastric carcinoma. Am. J. Clin. Oncol. 20 (4):348-353.

Pujol, J.L., J. Grenier, and J.P. Daures. 1993. Serum fragment of cytokeratin subunit 19 measured by Cyfra 21-1 immunoradiometric assay as a marker of lung cancer Cancer Res. 53:61-70.

Reynoso, G., T.M. Chu, D. Holyoke, E. Cohen, T. Nemoto, and J.J. Wang. 1972. Carcinoembryonic antigen in patients with different cancers. J.A.M.A. 220 (3):361-365.

Roulston, J.E., and R.C.F. Leonard. 1993. Applications to cancer. Pages 37-51 in J.E. Roulston and R.C.F. Leonard, eds. Serological tumour markers. Churchill Livingstone, New York, NY.

Spila, A., M. Roselli, M. Cosimelli, P. Ferroni, F. Cavaliere, R. Arcuri, M. Tedesco, S. Carlini, R. D'Alessandro, P. Perri, C.U. Casciani, J.W. Greiner, J. Schlom, and F. Guadagni. 1996. Clinical utility of CA 72-4 serum marker in the staging and immediate post-surgical management of gastric cancer patients. Anticancer Research 16 (4B):2241-2247.

Steinberg, W. 1990. The clinical utility of the CA 19-9 tumor-associated antigen. Am. J. Gastroenterol. 85:350-355.

Van-Dalen, A., and A.C. Kessler. 1996. A multicentre evaluation of tumour marker determinations using the automatic enzymun test systems ES 300 and ES 600/700. Eur J. Clin. Chem. Clin. Biochem. 34 (4):377-384.

World Health Organization, 2001. Globocan 2000: Cancer incidence, mortality, and prevalence worldwide, version 1.0. IARC CancerBase No 5. Lyon, 2001. http:www-dep.iarc.fr/globocan/globocan.html.

Wu, J.T., 1996. Diagnosis and management of cancer using serologic tumor markers. Pages 1064-1080 in J.B. Henry, ed. Clinical Diagnosis and Management by Laboratory Methods 19th edition, W.B. Saunders, Philadelphia, PA.

Wu, J.T., and R.M. Nakamura. 1997. Human circulating tumor markers. American Society of Clinical Pathologists, Chicago, IL. 263 pp.

Kevin L. Beason,(1) Shawn R. Clinton, (1)Sabrina Bryant, (1) James T. Johnson, (1) Margaret Jackson, (1) Harold Schultze, (1) Deborah Fortenberry, (1) Cynthia Bright, (1) Helen Hua, (1) Jiarong Ying, (1) Paul Sykes, (1) Cynthia Wilson, (2) Kay Holifield, (3) Charlton Vincent, (3) and Margot Hall (1, 4)

(1.) University of Southern Mississippi, Hattiesburg, MS 39406

(2.) University of Mississippi Medical Center, Jackson, MS 39216

(3.) Laurel Clinic for Women, Laurel, MS 39442

(4.) Author for correspondence. Department of Medical Technology, Post Office Box 5134, [e-mail]