Darrow E. Haagensen

Evaluation of carcinoembryonic antigen as a plasma monitor for human breast carcinoma

Plasma CEA levels have been determined in 92 normal women and 768 women with benign or malignant breast diseases. Only one of 92 normal women had a CEA level above 5 ng/ml. Of 253 women with benign breast diseases (gross cystic disease, adenofibroma, fibrosis, etc.) only one had a CEA level above 5 ng/ml. Ninety‐four percent of the above two groups of women had CEA levels below 3 ng/ml. Of 164 women operated upon for Columbia Clinical Classification Stage A or B breast carcinoma, preoperative CEA levels were above 5 ng/ ml in seven (4%). Patients with a preoperative CEA level above 3 ng/ml seemed to have an increased incidence of tumor recurrence. Elevated CEA levels (>10 ng/ml) in our postmastectomy population of 288 patients have correlated with development of metastases in 14 of 46 subjects. Of 216 patients under treatment for metastatic breast carcinoma, CEA levels above 10 ng/ml have been detected in 15 percent of patients with soft tissue metastases, 38% of patients with visceral metastases and 50% of patients with osseous metastases. Of metastatic breast carcinoma patients with CEA levels above 10 ng/ml serial measurements have correlated with the patients response to therapy, progressively increasing in treatment failures and decreasing in treatment responders.

Medical adrenalectomy with aminoglutethimide: clinical studies in postmenopausal patients with metastatic breast carcinoma.

The use of adrenalectomy and hypophysectomy in the management of postmenopausal patients with metastatic breast carcinoma is reserved for highly selected patients. As an alternate approach, a pharmacologic method of inhibiting adrenal cortical secretion was developed which consisted of the daily administration of 1000 mg of aminoglutethimide to block steroidogenesis and either dexamethasone (2.0–3.0 mg/day) or hydrocortisone (40–60 mg/day) as replacement glucocorticoid. This regimen markedly suppressed plasma levels of DHA-S, androstenedione, estrone, and estradiol, and urinary levels of aldosterone. Of 50 patients treated, 19 (38%) demonstrated either a complete (8/19) or a partial (11/19) objective disease remission which lasted for 18.05 ± 3.1 months (mean ± SEM). In 10 (20%) patients, there was stabilization of disease (7.8 ± 1.2 months), accompanied by symptomatic relief of bone pain in six (12%). There was disease progression in 20 (40%) patients. The acute side effects of aminoglutethimide therapy were significant and consisted of transient lethargy (41.5%) and a cutaneous rash (35.8%). Chronic toxicity was negligible. The medical adrenalectomy regimen of aminoglutethimide plus glucocorticoid offers a suitable alternative to surgical adrenalectomy or hypophysectomy in the management of postmenopausal patients with metastatic breast carcinoma.

Evaluation of a breast cyst fluid protein detectable in the plasma of breast carcinoma patients.

A radioimmunoassay has been developed for one of the major proteins isolated from human breast cyst fluid. Im-munologically this protein is identical to a protein present in both human milk and saliva. Ninety-two normal women had plasma levels of this protein below 100 ng/ml (range 7–81 ng/ml; mean 31 ng/ml), and 85% had plasma levels below 50 ng/ml. More than one-half of the women with active gross cystic disease of the breast had plasma levels above 50 ng/ml. None of these women, however, had plasma levels above 150 ng/ml. A significant percentage of women with advanced breast carcinoma have been found to have plasma concentrations of this protein greater than 150 ng/ml.

Localization of I-131-labeled goat and primate anti-carcinoembryonic antigen(CEA) antibodies in patients with cancer.

Thirty patients with anti-carcinoembryonic antigen (CEA)-producing cancers of the colon, breast, or thyroid were injected with 1 to 2 mCi of Iodine-131 (131I)-labeled, affinity-purified, goat or baboon anti-CEA antibodies. Images were obtained daily for four days. Computerized background subtraction using technetium 99m (99mTC)-labeled compounds was used. Images obtained with and without background subtraction were correlated with other evidence of disease. Activity levels in plasma, urine, and thyroid gland were monitored. Significant deiodination of antibody occurred within the first 24 hours. The mean plasma half-disappearance-time of baboon antibody was significantly longer than the mean half-disappearance-time of goat antibody. With exogenous blockade, total thyroid uptake was less than 0.1% of the injected dose. Without background subtraction, scintigraphic localization of known tumor was possible in one of two patients with colon carcinoma, in three of 20 patients with breast cancer, and in one of five patients with medullary carcinoma of the thyroid. With background subtraction, potential false-positive results could be generated for every patients, depending on the normalization site chosen and the degree of subtraction used. In contrast to results of previous reports, CEA-producing tumor was found to be infrequently localized using highly purified goat or primate radiolabeled anti-CEA. Furthermore, the subtraction technique described by previous investigators may lead to a high false-positive rate.

The detection of elevated plasma levels of carcinoembryonic antigen in patients with suspected or established medullary thyroid carcinoma.

Plasma levels of carcinoembryonic antigen (CEA) and calcitonin (CT) were measured in 35 normal control subjects and in 37 patients with suspected or established medullary thyroid carcinoma (MTC). None of the normal control subjects had elevated basal plasma levels of either CEA (>5 ng/ml) or CT (>0.25 ng/ml). However, of the 37 patients with suspected or established MTC, 23 (62%) had elevated basal plasma levels of CEA (range 9.8–7,000 ng/ml) and 27 (73%) had elevated basal plasma CT values (range 0.30–500 ng/ml). Generally, patients with clinically apparent MTC, either primary or metastatic, had higher plasma CEA and CT levels than those with occult disease. A positive correlation was found (r = 0.785, p < 0.01) when comparing basal plasma CEA and stimulated plasma CT levels in 20 patients. A marked increase above the basal plasma level of CT but not CEA was detected in each of six MTC patients following intravenous calcium or pentagastrin. These data demonstrate that basal plasma levels of CEA and CT were increased in a large percentage of patients with MTC. Plasma calcitonin levels unlike plasma CEA levels were more often elevated in patients with occult disease and were increased above basal following the intravenous administration of either calcium gluconate or pentagastrin.

Comparison of sex steroid receptor analyses and carcinoembryonic antigen with clinical response to hormone therapy

This study corroborates previous reports which suggested the efficacy of estrogen receptor (ER) analysis in predicting responses of patients with metastatic mammary carcinoma to hormonal therapeutic manipulation. The predictive value of multiconcentration titration and sucrose density gradient analyses of ERs and progesterone receptors (PRs) are compared. The predictive value of ER analyses can be improved by the discrimination of 8S versus 4S binding species or by the use of PR analysis in combination with ER analysis. The tumor‐associated antigen, carcinoembryonic antigen (CEA), is evolving as an important quantitative aid in evaluating the clinical responses of patients receiving hormonal therapy.

Comparative evaluation of carcinoembryonic antigen and gross cystic disease fluid protein as plasma markers for human breast carcinoma

A comparative evaluation of carcinoembryonic antigen (CEA) and gross cystic disease fluid protein (GCDFP) as plasma markers for human breast carcinoma has been performed. Both assays appear to be useful in patients with metastatic breast carcinoma. Of 216 patients under treatment for metastatic disease, 111 (51%) had abnormal plasma levels of CEA and/or GCDFP. Abnormal plasma levels of CEA were present in 73 patients whereas abnormal GCDFP levels were present in 67. Twenty‐nine patients had increased plasma levels of both markers simultaneously, 44 patients had increased CEA levels only and 38 patients had increased GCDFP levels only. Thus, of the 111 patients with elevated levels of either CEA or GCDFP, the two markers varied independently of each other in 74%. Utilizing both assays, abnormal plasma levels were present in 79% of patients with osseous metastasis, 53% of patients with visceral metastasis and 26% with soft tissue metastasis. Both assays, when performed serially in patients treated for metastatic breast carcinoma, were found to have utility in monitoring responsiveness; an increasing CEA or GCDFP plasma level indicated disease progression and a decreasing plasma level indicated regression.

Immunoperoxidase localization of GCDFP-15 with mouse monoclonal antibodies versus rabbit antiserum.

Five monoclonal antibodies (Mabs) raised against separate determinants on a breast gross cystic disease fluid protein of 15 KD (GCDFP-15) were compared to one another and to a rabbit antiserum (Rb) against GCDFP-15 by radioimmunoassay (RIA) and by immunoperoxidase localization in paraffin-embedded tissues. All five Mabs and the Rb were equivalent in recognition of GCDFP-15 in solution, as determined by RIA. However, two of the Mabs (A5, B15) showed only minimal binding to GCDFP-15 in paraffin-embedded tissues, whereas the other three Mabs (B1, B4, D6) were equivalent to the Rb in staining intensity. These latter three Mabs and the Rb were evaluated by the immunoperoxidase technique on a variety of benign and malignant neoplasms as well as normal tissues (150 specimens) for staining specificity. Immunoperoxidase staining by the three Mabs vs the Rb was equivalent in apocrine glands, metaplastic apocrine epithelium of breast, and breast carcinomas with apocrine features. No staining of the Mabs or Rb was seen in the other tissue specimens.

Quantitation of response to therapy in patients with metastatic breast carcinoma by serial analysis of plasma gross cystic disease fluid protein and carcinoembryonic antigen

The maximum percent change (MPC) of plasma carcinoembryonic antigen (CEA) and gross cystic disease fluid protein (CDP) were correlated with response to therapy in 92 metastatic breast carcinoma patients. In patients treated with hormone therapy MPC values were significantly different between patients with disease progression (Prog) and regression (Reg): MPC‐CEA for Reg = −72 ± 7%, for Prog = 396 ± 150%; MPC‐CDP for Reg = −86 ± 6%, for Prog = 702 ± 330%, P > 0.001 in a one‐way ANOVA for CEA and CDP. Similar differences were noted in patients treated with chemotherapy. Decreased (>50%) plasma CEA levels were observed in 24/29 (83%) of Reg, 18/35 (51%) stable and 0/49 (0%) of Prog; decreased (>50%) plasma CDP levels were noted in 19/24 (79%) of Reg, 21/28 (75%) of stable and 2/35 (6%) of Prog. Patients with plasma marker decreases > 50% had significantly longer responses to therapy (14.2 months for CEA, 14.1 months for CDP) compared to patients with < 20% decrease (2.0 months for CEA, 0.8 months for CDP), P < 0.001 in a one‐way ANOVA. Decreasing marker levels during the initial six weeks of therapy (negative slope) accurately identified Reg or stable patients: the predictive value of a negative slope was 92% for CEA and 86% for CDP. Rising marker values correctly identified treatment failures (Prog): the predictive value of a positive slope was 90% for CEA and 76% for CDP. These data indicated that changes in plasma CEA and CDP levels reflected increasing or decreasing tumor burden during hormone or chemotherapy treatment of metastatic breast carcinoma. Criteria have been established to predict therapeutic outcome based on the slope of CEA or CDP after six weeks of treatment.

Fluoxymesterone stimulation of tumor marker secretion in patients with breast carcinoma

The gross cystic disease fluid protein of 15,000 MW (GCDFP-15) has been demonstrated to be a circulating glycoprotein tumor marker for breast carcinoma in approximately 40% of patients with advanced disease. A recent retrospective analysis of plasma GCDFP-15 levels in patients with advanced breast cancer suggested that androgen therapy could cause significant increases in plasma levels in the absence of disease progression. In order to evaluate the frequency, time course, and intensity of the androgen effect on GCDFP-15 production, a prospective study was initiated.Twenty-nine patients with stage IV breast carcinoma were treated with fluoxymesterone (20 or 30 mg/d). Plasma levels of GCDFP-15 and carcinoembryonic antigen (CEA) were measured by radioimmunoassay before and at various times during therapy. By day 6 of therapy, plasma GCDFP-15 had increased significantly (p = 0.03) from a mean basal level of 58 ± 12 ng/ml to 160 ± 60 ng/ml. By contrast, the mean CEA levels in the same patients increased only from 36 ± 14 ng/ml to 38 ± 13 ng/ml. The distribution of percent increases in plasma GCDFP-15 was not uniform, but patients with high (>82 ng/ml) basal levels had marked (⩾75%) increases in 6/6 (100%) cases, whereas patients with low (<30 ng/ml) basal levels had similar increases in only 2/15 (13%) cases. Urinary excretion of GCDFP-15 usually paralleled the increases in plasma levels of the glycoprotein during the first six days of therapy. A linear correlation between percent change in plasma and percent change in urinary GCDFP-15 was demonstrated. A permanent cell line of human breast carcinoma, T47-D, was stimulated to secrete GCDFP-15in vitro by androgen, but not by estrogen.From these data, we conclude that androgens can specifically stimulate secretion of GCDFP-15 by breast carcinoma tissue in most patients with elevated plasma levels of GCDFP-15, and in some patients with normal levels. The stimulation occurs within days and is not associated with clinical signs of tumor growth.