A new breast carcinoma antigen defined by a monoclonal antibody

A new monoclonal antibody (MoAb), 3E1-2, to human breast carcinoma cells was made. With the use of the immunoperoxidase technique, 3E1-2 was tested on Formalin-fixed and fresh sections of 27 normal and 81 neoplastic tissues, including 37 carcinomas of the breast, 15 lung tumors, 5 colon tumors, and other tumors. Strong uniform staining of the cytoplasm and membrane occurred with the breast carcinoma, whereas with normal breast tissue less intense staining of the luminal membrane was seen; not all cells were reactive with the MoAb. Most other human tumors (with the exception of some lung, kidney, and uterine carcinomas) were nonreactive, and few normal tissues were reactive. The unique features of this new MoAb are: a) reaction with Formalin-fixed as well as fresh tissue; b) lack of a reaction with the cell surface of 43 established cell lines, including 10 lines derived from breast carcinoma cultures; c) variable staining patterns in different breast carcinomas, varying from all cells staining to dense cytoplasmic staining to minimal membrane staining of a few cells; d) a great differential in staining patterns between normal and neoplastic tissue (nonetheless, some normal tissues were 3E-1.2+). The antibody does not detect a tumor-specific antigen, but has a high carcinoma-to-normal breast ratio of staining. In addition, preliminary studies on the sera of 20 patients with carcinoma of the breast have shown that the antigen detected by 3E1-2 is elevated in their serum; 3E1-2 thus has the potential to be used for diagnosis of this disease.     

Tissue distribution of an epithelial and tumor-associated antigen recognized by monoclonal antibody F36/22

Murine monoclonal antibody F36/22 was derived by immunizing BALB/c mice with human breast cancer cells. This antibody reacts with an antigen located both on differentiated mammary ductal epithelia and on breast carcinomas, as examined by indirect immunoperoxidase techniques. Although the expression of this antigen correlated with estrogen receptor levels of breast tumors, antibody F36/22 did not directly react with estrophilin. In contrast to the expression of classical differentiation antigens, this antigen was found in a high percentage of poorly differentiated carcinomas of the breast. Staining intensities were similar for well- and poorly differentiated tumors; thus, antigen expression was not related to tumor grade. Intratumoral heterogeneity of antigen expression was observed in the majority of tumors. Since a subset (64 of 80) of the breast carcinomas examined have expressed the antigen, McAb 36/22 was of use for the immunological subclassification of tumors which were indistinguishable by conventional histopathological staining techniques. The antigen was also present on other adenocarcinomas (ovary, colon, stomach, pancreas, and prostate); however, these tumors usually exhibited reduced staining intensity compared with that observed in breast cancer. The normal counterpart tissues at these histotypes contained no detectable levels of the antigen, and increased expression of the antigen was associated with tumorigenesis at these sites. Tumors of mesenchymal origin and carcinomas other than adenocarcinomas exhibited undetectable levels of the antigen. Therefore, depending on the organ site, McAb F36/22 recognizes an epithelial and/or tumor-associated antigen.     

Reactivity of monoclonal antibody F36/22 with human ovarian adenocarcinomas

Monoclonal antibody F36/22 recognizes high-molecular-weight glycoprotein components associated with neoplastic development of the ovary. Indirect immunoperoxidase staining techniques were performed on a panel of nonmalignant ovarian tissues, primary ovarian tumors, exfoliated ascitic tumor cells, and metastatic lesions. Normal ovarian tissue components (n = 20) failed to exhibit detectable levels of antigen, whereas benign ovarian tissues show a low incidence of immunostaining (three of 26) restricted to some ductal elements. One hundred % (19 of 19) of the immunopositive primary malignant tumors were histologically classified as adenocarcinomas. Each of the predominant adenocarcinoma histotypes consistently showed expression of the antigen with 30 to 100% of the tumor cells scored as immunopositive. Ascitic tumor cells obtained from all of the ovarian adenocarcinoma patients examined (47 of 47) displayed immunopositive reactions, whereas normal mesothelial cells in these specimens exhibited undetectable staining. In addition, ovarian adenocarcinoma metastases (12 of 12) exhibited very intense immunoreaction products. No detectable antigen was expressed by nonadenocarcinoma ovarian tumor cells.     

Measurement of a breast cancer associated antigen detected by monoclonal antibody SP-2 in sera of cancer patients

An enzyme-linked immuno-absorbent assay has been developed for the detection of a circulating tumor-associated antigen, 90K, recognized by murine monoclonal antibody SP-2 (Iacobelli et al., Cancer Res 46: 3005-3010, 1986). This assay was found to be simple and reproducible. Using this method, 6% of 165 apparently healthy individuals and 15% of 91 patients with benign breast disease showed 90K levels above 1.7 units/ml. Approximately 50% of 129 patients with advanced breast cancer demonstrated serum antigen levels above 1.7 units/ml. All histological types of breast cancer were positive, and no association between the incidence of elevated 90K levels and other prognostic variables could be detected. The titers of 90K were significantly higher in sera from advanced-stage (3 and 4) patients than in those from patients with limited-stage (1 and 2) disease. Elevated 90K levels were also observed in patients with carcinomas of other sites, including gastrointestinal, gynecological, and lung tumors. By means of the immune blotting technique, the antigenic components carrying the determinants in serum and extracts of breast cancer cells have been identified. The levels of 90K did not correlate with those of CA 15-3 or CEA. The measurement of 90K in sera appears to be a useful adjunct to other available assays for the detection and monitoring of breast cancer and other malignant tumors.     

An evaluation of ovarian carcinoma-associated antigen defined by murine monoclonal antibody CF511 in sera from patients with ovarian carcinoma

Murine monoclonal antibody CF511, raised against human ovarian clear cell carcinoma, detects a glycoprotein (Mr 600 kDa) called CF511 antigen which is elevated in the serum of many patients with ovarian carcinoma. A competitive enzyme-linked immunosorbent assay was developed to detect CF511 antigen in human serum and used to detected CF511 antigen in subjects with ovarian carcinoma and other diseases. No raised levels (less than 18 unit (U) ml-1) of the antigen were found in the serum of 220 normal individuals or of patients with germ cell tumours (n = 6), granulosa theca cell tumour (n = 1), gastric carcinomas (n = 10) and colo-rectal carcinomas (n = 8). Raised serum levels of CF511 antigen were found in 6/46 patients (13.0%) with benign gynaecological tumours (including endometriosis or ovarian cyst), in 5/7 patients (71.4%) with breast carcinoma and 16/21 (76.2%) lung carcinoma patients. In patients with ovarian carcinoma, 42.3% (11/26) of stage I and II, and 96.0% (24/25) of stage III and IV had levels of greater than or equal to 18 U ml-1. In all patients with serial determination of CF511 antigen levels before and after the surgery, the levels of antigen correlated with the clinical course of disease. Determination of CF511 antigen levels may be useful for detection of ovarian carcinoma as well as lung and breast carcinomas and for monitoring progress of disease and response to therapy.     

Identification of an H antigen-like blood group antigen in sera of cancer patients using a novel monoclonal antibody raised against endometrial carcinoma

An IgM class monoclonal antibody (MAb) was derived by immunizing BALB/c mice with a human endometrial carcinoma cell line. This MAb, termed C12, exhibited strong reactivity against endometrial carcinoma, but lesser reactivity against normal endometria. The antigen recognized by MAb C12 (C12 antigen) was detected by radiometric assay in sera from patients with various carcinomas, but not in sera from patients without carcinomas or in sera from normal individuals. MAb C12 was found to agglutinate blood type O erythrocytes, but not A, B, or AB erythrocytes. To clarify the specificity of MAb C12, tissue staining experiments were performed in parallel using MAb C12, Ulex europaeus lectin I (anti-H), and a monoclonal anti-H antibody. In endometrial carcinoma tissues, both H and C12 antigens increased, but the C12 antigen showed a prominent increase, in contrast to the H antigen. Further, the C12 antigen was not found in endothelial cells of blood type O patients. In sera, the level of the H antigen varied according to the host's blood type. The sera from blood type O individuals possessed higher levels of the H antigen than those with blood type A or B. Thus, the H antigen showed no value as a tumor-associated serum marker. In contrast, the presence of the C12 antigen in sera was not determined by ABO blood group status. Thus, MAb C12 was demonstrated to be a unique MAb that reacts with an H-like antigen occurring in the sera of patients with carcinomas irrespective of ABO blood group status. MAb C12 may prove to be a useful marker for cancer patient serum.     

Tumor-associated antigen defined by a monoclonal antibody against neuraminidase-treated human cancer cells

A monoclonal antibody, MH-A6, was produced by immunization with a human gastric cancer cell line, MKN 74, treated with neuraminidase. The antigen defined by the monoclonal antibody was detected on various tumor tissues and a limited number of normal tissues in immunoperoxidase assay, and the expression of MH-A6 antigen was not influenced by neuraminidase treatment except for some cases of tumor tissues. Interestingly, neuraminidase treatment enhanced binding of the antibody on some adenocarcinomas, but diminished binding of the antibody on squamous cell carcinomas. Both treatment of the immunizing tissues with trypsin and periodic acid diminished binding of the antibody. In isolation of MH-A6 antigen from MKN 74 cells by the monoclonal antibody coupled-affinity column, the epitope exists on molecules with molecular weights of 30,000 and 72,000, and with an acidic pH range in two-dimensional electrophoresis. CEA and CA 19-9 activities were not detected in purified MH-A6 antigen by solid-phase radioimmunoassay, and the reactivity of the MH-A6 antibody with CEA and CA 19-9 was not detected in enzyme-linked immunosorbent assay. Hemagglutination observed between erythrocytes (Lewisa, Lewisb, or NE-treated) and anti-Lewisa, anti-Lewisb sera, or anti-T-agglutinin (peanut lectin), respectively, was not inhibited by MH-A6 antigen. The results suggest that MH-A6 antigen is a tumor-associated antigen, probably glycoprotein, and different from CEA, CA 19-9, Lewisa, Lewisb, and Thomsen-Friedreich (T) antigen.     

Detection of a soluble antigen defined by monoclonal antibody 83D4 in serous effusions associated with breast carcinoma.

Monoclonal antibody (MoAb) 83D4 was generated by immunization with cell suspensions obtained from sections of formol-fixed paraffin embedded human breast cancer. It recognized an antigen expressed in breast carcinomas but not in normal breast tissue. Pleural and ascitic fluids from 66 patients were studied by an 83D4 heterologous sandwich radioimmunoassay (SRIA) using solid-phase immobilized wheat germ agglutinin to detect the 83D4 soluble antigen. Using a cutoff level of 5 units/ml of 83D4 antigen, higher values were found in 22 of 27 breast cancer-associated effusions (mean = 10.72 +/- 6.80 units/ml). The 20 nonmalignant effusion fluids tested showed lower values (mean = 1.16 +/- 1.49 units/ml, P less than 0.001). The antigen was undetectable or present in low levels in effusions from patients with hematologic malignancies. When SRIA results were compared with conventional cytologic diagnosis in breast-cancer effusions, elevated levels of 83D4 soluble antigen were found in all patients (8 of 8) in whom malignant cells had been detected, in 4 of 8 patients with the diagnosis of "suspected malignancy," and in 10 of 11 patients with negative cytologic findings. Using an immunoglucosidase method on cell smears of various origins, MoAb 83D4 stained metastatic cells of breast and ovary carcinomas but did not reactive with mesothelial cells and other normal or malignant cell types. These results suggest that quantitation of the 83D4 soluble antigen may be used to improve the diagnosis of cancer in serous effusions.     

Membrane antigen in small cell carcinoma of the lung defined by monoclonal antibody SM1

Mouse myeloma cells were fused with spleen cells from BALB/c mice immunized with a cell line derived from human small cell carcinoma (SCC) of the lung. The cloned hybridoma SM1 produced antibody that was reactive with the surface membrane of SCC cell lines and SCC tumors but not with the membrane of several non-SCC cell lines and tumors. SM1 ascites fluid was used to screen for reactivity of the antibody with other human cancer cell lines, tumor tissues, and normal tissues. SM1 antibody was found to be unreactive with neuroblastoma, adrenal carcinoma, melanoma, and bronchial carcinoid. Reactivity was detected with some breast carcinoma cell lines but not with breast cancer tissue specimens. In the same individual, the antibody was reactive with SCC lung tumor and SCC metastatic to the liver but not with normal tissues, including bronchus, lung parenchyma, liver, kidney, and brain. Human erythrocytes and marrow cells were also unreactive. Since SM1 detects an antigen that is present in greatest amounts on the surface membrane of SCC of the lung, this antibody may be useful in tracing the lineage patterns of human lung cancers.     

Histochemical studies of human breast cancer using a monoclonal antibody against an oestrogen receptor-related antigen

The presence or absence of an oestrogen receptor-related antigen in breast tumours has been examined histochemically using a monoclonal antibody ('Ds' - Coffer & King, 1981). In frozen sections, fixed either by the method of Tamura et al. (1980) or in methanol, staining was apparent in 14/24 (58%) and 22/26 (85%) of the breast cancers respectively. In paraffin sections fixed in ethanol, staining was present in 25/33 breast cancers (76%). In either type of section, staining was predominantly in the cytoplasm of the epithelial cells. When staining was scored by independent observers (2 or 3) and related to the tumour oestrogen receptor activity, determined by a standard biochemical technique, antigen was present in both receptor-positive and receptor-negative tumours. No significant association was found between the presence of antigen and receptors in the frozen sections, but for the series of paraffin sections, there was a weak association (r = +0.48) between the presence of the two proteins. Histochemical processing of paraffin sections from 9 tumours under conditions of higher sensitivity increased the staining significantly in 2/9 tumours, but did not alter the relationship between staining and receptor status. Six tissues were stained after exposure to 'receptor-translocating' conditions (25 degrees C/2 nM oestradiol/both for 1 h): this did not consistently change the subcellular staining pattern, though all tissues tended to stain more after exposure to 25 degrees C. Staining was not blocked by absorption of the D5 antiserum with a variety of pure proteins or human serum but at higher concentrations (approx. 2-15 mg protein ml-1), extracts from human uterus, an oestrogen-receptor-positive breast cancer and an oestrogen-receptor-negative breast cancer all effectively abolished staining in sections from another breast cancer. These results are consistent with other reports suggesting that the D5 antibody detects an antigen which is not the oestrogen receptor, but which may be associated with the receptor in its tissue distribution.     

Carbohydrate antigen defined by a monoclonal antibody raised against a gastric cancer xenograft

A monoclonal antibody, ST-4-39, was obtained by using a human gastric cancer xenograft, St-4, as an immunogen. Immunization was achieved by transferring immunocompetent mouse spleen cells into a nude mouse bearing St-4. Hybridomas were produced with the spleen cells of the mouse after rejection of the tumor and screened for immunohistochemical reactivity with cancers and normal tissues on formalin-fixed paraffin sections. ST-4-39 immunohistochemically reacted with various cancers including gastric, colorectal and pancreatic cancers as well as some normal tissues. ST-4-39 and NS 19-9 differed in immunohistochemical reactivity, although they reacted with some cancers and a few normal tissues in common. PBS extracts of normal and cancer tissues were examined for antigen reactive with ST-4-39 by sandwich enzyme immunoassay. Extractable antigen was detected in adenocarcinomas of colon, stomach and lung, while it was detected only in salivary gland and trachea among normal tissues examined. Gel filtration analysis of the antigen indicated a molecular weight of greater than or equal to 1 X 10(6), and the antigenic determinant was suggested to be a carbohydrate chain with terminal sialic acid by studies using periodic acid, neuraminidase and pronase treatments. Furthermore, the ST-4-39 antigen affinity-purified from two gastric cancer strains was shown to contain multiple carbohydrate determinants including sialyl-Lewisa and sialyl-Lewisx, suggesting the antigen to be a mucin.     

Monoclonal antibodies (F36/22 and M7/105) to human breast carcinoma

Cloned hybridoma cell lines were obtained from fusions of murine myeloma cells with lymphocytes of mice immunized against human breast cancer cells. Hybridomas F36/22 and M7/105 produced antibodies whose binding to breast cancer cells could not be inhibited by prior absorptions with fibroblasts, lymphoblastoid cells, or erythrocytes. Results from cell surface binding assays using a panel of tumor cell lines indicated that antibodies F36/22 and M7/105 recognized determinants expressed maximally on breast cancer cells. Antibody F36/22 reacted with normal mammary epithelial membranes and milk fat globule membranes, whereas antibody M7/105 produced no detectable binding to these specimens. Antigens carrying these epitopes each showed reactivity with concanavalin A lectin. The determinant corresponding to antibody F36/22 was detectable in histological sections of a subset of breast tumors obtained at surgery.     

Detection of circulating adenocarcinoma-associated antigen in the sera of cancer patients with a monoclonal antibody

The antigen (YH206 antigen) corresponding to the monoclonal antibody (MoAb) YH206 (IgM), which was produced against a lung adenocarcinoma cell line A549, was found to be an extremely high-molecular-weight protein of over 330,000 daltons by means of SDS-PAGE and Western blot analysis. The immunohistological distribution of the antigenic determinant recognized by MoAb YH206 was found to be limited to adenocarcinoma tissues. It was shown by reversed passive hemagglutination assay (RPHA) that the antigen could be detected not only in the spent medium from human lung cancer cell line A549 but also in the sera from cancer patients. Only 3 out of 30 (10%) healthy donors and 4 of 31 (12.9%) patients with benign diseases had serum antigen levels of more than 1/64 dilution. In contrast, 42 of 87 (48.3%) patients with lung cancer and 29 of 50 (58.0%) patients with cancers of the digestive organs had elevated levels of the antigen. As regards the relation between antigen levels and clinical stages of lung cancer, values of more than 1/128 dilution were detected only in patients with stage III or IV disease.     

A new tumor-associated antigen expressed on breast carcinomas, defined by monoclonal antibody BCA 227

After immunization of mice with the human breast carcinoma cell line MCF-7, we produced monoclonal antibody (mAb) BCA 227, which allowed us to characterize a new tumor-associated antigen. This molecule is strongly expressed by well differentiated mammary carcinoma cell lines and by some other tumor cell lines of epithelial origin. Immunohistological study of frozen sections of different tissues and tumors confirmed its expression by tumor cells of epithelial origin, particularly infiltrating duct carcinomas of the breast. The antigen is also expressed, to a lesser extent, by some normal epithelial cells. Its biochemical characterization revealed a Mr 71,000 protein without an N-linked sugar moiety. Six to 40 x 10(3) binding sites are present on breast tumor cell surfaces. Although mAb BCA 227, which was found to be of the IgG2a isotype, did not mediate antibody-dependent cell-mediated cytotoxicity with either human or mouse effector cells, a 50% inhibition of SK-BR5 tumor growth was obtained in nude mice, suggesting that another mechanism is responsible for this inhibition. Biodistribution studies of radiolabeled F(ab')2 fragments of mAb BCA 227 in tumor-bearing nude mice showed a preferential localization in the tumor. All these data are in favor of the use of mAb BCA 227 as an immunodiagnostic tool for breast cancer.     

Epitope analysis of monoclonal antibody NCRC-11 defined antigen isolated from human ovarian and breast carcinomas

NCRC-11 is an IgM monoclonal antibody which defines an antigen found in most epithelial malignancies. The antigen has previously been shown to be a high mol. wt. glycoprotein (greater than 400,000) and in this study, antigen preparations were isolated by immunoadsorbent chromatography from ovarian mucinous and ovarian serous cyst adenocarcinoma and from breast carcinoma. Other monoclonal antibodies, against products in normal human milk, and antibodies of the Ca series (Bramwell et al., 1985) reacted with all three antigen preparations. Tests involving epitope mapping were performed to probe the relationships of the various epitopes to that defined by the NCRC-11 antibody, and, of note, the three antigen preparations from different tumour sources were remarkably similar with respect to their relative levels of epitope expression and to their topographical distribution of epitopes. The major differences in epitope expression could be attributed to the degree of sialylation in the three antigens. The antigens from ovarian tumours expressed I(Ma) blood group determinants (defined by the antibody LICR-LON-M18) which were partially masked by sialic acid. With NCRC-11 defined antigen from breast carcinoma, this determinant was totally masked by sialic acid although neuraminidase treatment clearly exposed epitopes reactive with M18 antibodies.     

Expression of a rat ovary-independent mammary tumor-associated antigen defined by a monoclonal antibody, TAK-B1

A monoclonal antibody, TAK-B1, was produced by immunization of BALB/c mice with mammary carcinoma induced in inbred Sprague-Dawley rats by treatment with 7,12-dimethylbenz[a]anthracene. TAK-B1 reacted with ovary-independent mammary carcinoma cells which had been transformed from ovary-dependent mammary carcinoma cells, but did not react with original mammary carcinoma cells or with cells from mammary glands exhibiting fibrocystic changes or normal mammary glands. However, TAK-B1 reacted not only with basal cells of the epidermis and epithelial cells of the bottom portion of crypts of the small intestine in adult rats, but also with basal cells of epidermis in skin and mesenchymal cells around developing hair follicles in fetuses. We therefore classify TAK-B1 as an ovary-independent rat mammary tumor-associated antigen. Immunoelectron microscopic examinations revealed that the antigen recognized by TAK-B1 was localized in the cell surface membrane of ovary-independent mammary carcinoma cells. Immunoprecipitation assay revealed that the antigen recognized by TAK-B1 was composed of M 220,000 protein and four other minor proteins.     

A renal cell carcinoma neoplastic antigen detectable by immunohistochemistry is defined by a murine monoclonal antibody

BALB/c mice were hyperimmunized with ACHN (ATCC CRL 1611, American Type Culture Collection, Rockville, Maryland), a stable in vitro cell line derived from a malignant pleural effusion in a 22-year-old man with renal cell carcinoma. The hyperimmune spleen cells were fused with NS-1 murine myeloma cells using polyethylene glycol. Hybridoma supernatants were screened for the presence of IgG reactive with detergent extracts of ACHN and nonreactive with detergent extracts of normal kidney tissue. A stable, rapidly growing clone named 5F4 was isolated. Supernatant from 5F4 was used as a primary antibody preparation for avidin-biotin complex immunoperoxidase staining of multiple cases of renal cell carcinoma, normal tissues, and other tumors. 5F4 produced IgG which reacted with a cytoplasmic structure in paraffin-embedded sections of all renal cell carcinomas tested. There was occasional, weak, granular, cytoplasmic staining of isolated tubular lining cells in adjacent normal kidney.     

Conjugates of monoclonal antibody 791T/36 with methotrexate in cancer therapy

The design and evaluation of drug-antibody conjugates for tumor therapy is considered with respect to the use of the anti-human osteogenic sarcoma monoclonal antibody designated 791T/36. Conjugates have been constructed by linking antibody to methotrexate; the agent has been linked to antibody, either directly or via a human serum albumin bridging agent. In each case, conjugates have been assessed for retention of antibody reactivity with osteogenic sarcoma target cells by comparing their capacity with that of unmodified antibody in competing with fluorescein isothiocyanate-labeled 791T/36 binding, as measured by flow cytometry. Retention of drug cytotoxicity was evaluated by in vitro cytotoxicity and colony inhibition assays. Conjugates with acceptable retention of drug and antibody reactivities were then examined for in vivo suppression of human tumor xenografts.     

Detection efficiency of colorectal carcinoma recurrence using technetium pertechnetate-anti-carcinoembryonic antigen monoclonal antibody BW 431/26

Background. A new anti-carcinoembryonic antigen (CEA) antibody, BW 431/26 (Scintimun, Behring-Werke, Marburg, Germany), labeled with technetium pertech-netate (Tc-99m), is an intact immunoglobulin G₁ monoclonal antibody that has been used to image colorectal cancer (CRC). This report is part of a prospective multicenter clinical trial initiated by the International Atomic Energy Agency to evaluate the role of this antibody in radioimmunoimaging of patients with suspected disease recurrence. Methods. A group of 31 consecutive patients underwent radioimmunoimaging with Tc-99m-BW 431/26 after resection of their primary CRC. Patient referral was based on either a persistent rise in serum CEA levels of unknown origin and/or questionable findings by other imaging studies. Whole-body planar scans and single photon emission computed tomography scans of selected body regions (e.g., chest, abdomen) were performed up to 24 hours after the intravenous antibody injection. Pathologic antibody concentration localizations by radioimmunoimaging were correlated with surgical, clinical, and other imaging modality findings to validate the accuracy of radioimmunoimaging in detecting CRC recurrence. Results. A total of 75 detected tumoral lesions was evaluated: 26 of 75 were of known origin (36%), and 49 of 75 were of unknown origin (65%). There were four true-negative lesions, one false-negative lesion, and no false-positive lesions; all others were true-positive lesions. Sensitivity was 96.8%, specificity 100%, and accuracy 98.6%. The study was easy to perform, without untoward side effects on patients after antibody administration. Conclusions. Anti-CEA antibody radioimmunoimaging is a highly reliable diagnostic procedure in detecting CRC recurrence and is useful especially for the diagnosis of patients with rising CEA blood levels of unknown origin, thereby significantly affecting patient management. Radioimmunoimaging should become part of the diagnostic workup of patients suspected of having CRC recurrence. Cancer 1995; 76:215-22.