Comparison of immunocytochemical and biochemical assays for estrogen receptor in fine needle aspirates and histologic sections from breast carcinomas

An immunocytochemical assay using a monoclonal antibody specific for estrogen receptor (ER-ICA) was performed on needle aspirates and on histologic sections (mastectomy and biopsy specimens) from 55 patients with breast cancer. A total of 82 ER-ICAs were performed, with matched cytologic and histologic specimens in 27 patients, cytology alone in 15, and histology alone in 13. ER-ICA results were described by a histochemical score (H score) based on intensity-weighted percentages of staining cells. The H scores were compared with results of sucrose density gradient (SDG) analysis of histologic specimens (mastectomy, resection, or biopsy). An H score greater than or equal to 10 and an SDG value greater than or equal to 10 fmol/mg protein were considered positive. The sensitivity of cytologic ER-ICA was 94%, the specificity 100%. The sensitivity of histologic ER-ICA was 67%, the specificity 90%. Correlating cytologic H score with Black's nuclear grade showed that grade 1 (the most anaplastic) carcinomas demonstrated the lowest H scores (mean, 7.3 +/- 29.8), whereas the highest H scores were noted in grade 3 tumors (mean, 150.0 +/- 88.1). Both SDG and ER-ICA showed ER values to be lower in premenopausal than postmenopausal women. There was no correlation between H score and presence of axillary nodal metastases or tumor size. An overall good correlation was demonstrated between immunohistochemical methods and biochemical analysis.

Screening antisera for subtype specificity based on immunoblotting of lectin-affinity electrophoresis: application toward alpha-fetoprotein subfractions

We adapted a method for evaluating monoclonal antibody specificity toward isoforms and examined 29 alpha-fetoprotein (AFP) antibodies from Hybritech. These antibodies were separated by affinity electrophoresis on agarose-containing erythroagglutinating phytohemagglutinin (PHA-E) followed by immunoblotting onto nitrocellulose. Under these conditions, AFP separates into four subfractions: NR (nonreactive with PHA-E), WR (reactive weakly), RS2 (strongly reactive), and RS1 (very strongly reactive). When polyclonal goat anti-AFP control was used as the test antibody, the distribution of amniotic fluid AFP was approximately 50%, 31%, 12%, and 7% for NR, WR, RS2, and RS1, respectively. The distribution for these same bands from the serum of a patient with hepatocellular carcinoma was 39%, 31%, 30%, and 0%; from the serum of a patient with embryonal carcinoma, it was 29%, 38%, 0%, and 33%, respectively. Twenty-six monoclonal antibodies, including the two used in the Hybritech Tandem-E assay for AFP, showed similar distribution, while the remaining three antibodies showed no reactivity to any of the AFP antigens in this system. Although we were not able to demonstrate any specificity of these antisera toward any single AFP isoform, use of this affinity electrophoresis system provides a model for studying isoform specificity of other AFP antibodies and other antibody-antigen systems.

Prostate-specific antigen and prostate gland volume: correlation and clinical application

We studied 103 patients seen in our Prostate Cancer Detection Clinic to determine whether a correlation exists between serum prostate-specific antigen (PSA) values and ultrasound-calculated prostate gland volume. Seventy men (68%) had a PSA value less than or equal to 4 ng/ml (our upper limit of normal). The men were subclassified by prostate gland volume at arbitrary break points. Twenty-five men (24%) had a prostate gland volume less than or equal to 25 cm3; in 96%, the PSA value was less than or equal to 4 mg/ml. Further analysis revealed that the percentage of men with a normal serum PSA value decreased as the prostate gland volume increased; 65.6% of the group with a gland volume between 25 and 50 cm3 (40 of 61) and 35.5% of the group whose prostate volume exceeded 50 cm3 (6 of 17) had PSA values less than or equal to 4 ng/ml. Four men had PSA values greater than 20 ng/ml; all had prostate cancer. Cancer was diagnosed in four additional patients, three with PSA values between 5 and 10 ng/ml and one with a PSA value less than 4 ng/ml. There appears to be a direct relationship between prostate gland volume and PSA value, as well as a cancer value threshold. The clinical implications of these findings are discussed.

Sequential cyclic combined hormonal therapy for metastatic breast cancer

Thirty postmenopausal patients who had evaluable estrogen receptor-positive or unknown metastatic breast cancer were treated with cyclic sequential combined hormonal therapy consisting of 50 micrograms of ethinylestradiol orally daily for 7 days followed by 400 mg of medroxyprogesterone acetate orally daily for 21 days, followed in turn by 7 days of rest. Cyclic administration was continued until progressive disease was detected. Patients who had had one previous chemotherapy regimen were included, but 63% of patients were previously untreated. Six patients achieved complete remission and 11, a partial remission, for an overall response rate of 57%. Median remission duration was 22 months; median time to disease progression for all 30 patients was 8 months. Toxicity consisted of cyclic vaginal bleeding, hot flashes, weight gain, irritability, and fluid retention. This cyclic, sequential hormonal regimen was effective and well tolerated.

Methods for determining "reference changes" from serial measurements: plasma lipid-bound sialic acid

Lipid-bound sialic (neuraminic) acid (LSA) was measured in EDTA-treated plasma of 26 healthy subjects at three-month intervals for up to one year. The change in LSA concentration for consecutive measurements ranged from -54 to 42 mg/L (mean, -2.1 mg/L; SD, 19.6 mg/L; n = 56). The "reference change" for plasma LSA (+/- 2 SD), calculated from distribution of the differences, was +/- 39 mg/L. The 88th percentile of the intra-individual variance was 338 mg2/L2 and the mean variance was 159 mg2/L2. Using the homeostatic, autoregressive time-series model, a reference change of +/- 51 mg/L between two consecutive measurements was determined to be statistically significant (i.e., expected by chance no more than 5% of the time) in 88% of the healthy subjects. Only 73% of the healthy subjects would have had intra-individual variances corresponding to the reference change of +/- 39 mg/L according to the autoregressive model. The concentration of LSA in plasma was significantly decreased upon surgery in five of 10 patients with colorectal adenocarcinomas of Dukes stages A-C when we used +/- 39 mg/L as the reference change, but in only two of the 10 when we used +/- 51 mg/L as the reference change.

Methods for determining "reference changes" from serial measurements: plasma lipid-bound sialic acid

Lipid-bound sialic (neuraminic) acid (LSA) was measured in EDTA-treated plasma of 26 healthy subjects at three-month intervals for up to one year. The change in LSA concentration for consecutive measurements ranged from -54 to 42 mg/L (mean, -2.1 mg/L; SD, 19.6 mg/L; n = 56). The "reference change" for plasma LSA (+/- 2 SD), calculated from distribution of the differences, was +/- 39 mg/L. The 88th percentile of the intra-individual variance was 338 mg2/L2 and the mean variance was 159 mg2/L2. Using the homeostatic, autoregressive time-series model, a reference change of +/- 51 mg/L between two consecutive measurements was determined to be statistically significant (i.e., expected by chance no more than 5% of the time) in 88% of the healthy subjects. Only 73% of the healthy subjects would have had intra-individual variances corresponding to the reference change of +/- 39 mg/L according to the autoregressive model. The concentration of LSA in plasma was significantly decreased upon surgery in five of 10 patients with colorectal adenocarcinomas of Dukes stages A-C when we used +/- 39 mg/L as the reference change, but in only two of the 10 when we used +/- 51 mg/L as the reference change.

IgG heavy-chain subclass typing of myeloma paraproteins by isoelectric focusing immunoblot analysis

The objective of this study was to develop a clinical laboratory method for subclass typing of human immunoglobulin G (IgG) paraproteins. Serum proteins were isoelectrically focused (IEF) in a mini-gel and passively blotted by capillary diffusion onto untreated nitrocellulose. Unreacted sites on the nitrocellulose were blocked with bovine serum albumin and the bound IgG was detected with peroxidase-conjugated anti-human IgG1-4 monoclonal antibodies from WHO/IUIS clones. The IEF immunoblot specificity was demonstrated by analysis of documented IgG, IgA, and IgM myeloma proteins of known subclass and light-chain composition. IEF immunoblots of sera from 18 myeloma patients who had an above-normal total IgG concentration produced IEF immunoblot patterns composed of five to 10 discrete bands (pI range 6.0 to 8.4). In contrast, no detectable IgG bands were observed with sera containing IgA and IgM paraproteins. The observed subclass frequencies of IgG paraproteins were 56% IgG1 (10/18), 28% IgG2 (5/18), 11% IgG3 (2/18), and 5% IgG4 (1/18). IEF immunoblot analysis permits the monitoring of changes in the pI and subclass of an IgG paraprotein over the course of a myeloma patient's therapy program.

Pancreatic oncofetal antigen and carbohydrate antigen 19-9 in sera of patients with cancer of the pancreas

Pancreatic oncofetal antigen (POA) and carbohydrate antigen 19-9 (CA 19-9) were measured in the sera of 23 patients with cancer of the pancreas to determine the true positive rates of these cancer markers. In one group of unselected pancreatic cancer patients (n = 9), both tests showed above-normal results in three patients and both gave values that were within reference limits in three other patients. Two of the three remaining patients had increased CA 19-9 but normal POA values, and one patient had increased POA but normal CA 19-9 concentrations in serum. In another group of 14 pancreatic cancer patients, selected on the basis of increased concentrations of POA in serum, the CA 19-9 values were increased in eight. In four patients who had progressive disease, the concentrations of both markers increased with time in one patient, only POA in one, and only CA 19-9 concentration in another. (The fourth patient had increased but stable concentrations of POA and CA 19-9 in serum.) These data suggest that serum POA and CA 19-9 measurements should be used in combination in the evaluation of patients with cancer of the pancreas.

IgG heavy-chain subclass typing of myeloma paraproteins by isoelectric focusing immunoblot analysis

The objective of this study was to develop a clinical laboratory method for subclass typing of human immunoglobulin G (IgG) paraproteins. Serum proteins were isoelectrically focused (IEF) in a mini-gel and passively blotted by capillary diffusion onto untreated nitrocellulose. Unreacted sites on the nitrocellulose were blocked with bovine serum albumin and the bound IgG was detected with peroxidase-conjugated anti-human IgG1-4 monoclonal antibodies from WHO/IUIS clones. The IEF immunoblot specificity was demonstrated by analysis of documented IgG, IgA, and IgM myeloma proteins of known subclass and light-chain composition. IEF immunoblots of sera from 18 myeloma patients who had an above-normal total IgG concentration produced IEF immunoblot patterns composed of five to 10 discrete bands (pI range 6.0 to 8.4). In contrast, no detectable IgG bands were observed with sera containing IgA and IgM paraproteins. The observed subclass frequencies of IgG paraproteins were 56% IgG1 (10/18), 28% IgG2 (5/18), 11% IgG3 (2/18), and 5% IgG4 (1/18). IEF immunoblot analysis permits the monitoring of changes in the pI and subclass of an IgG paraprotein over the course of a myeloma patient's therapy program.

The role of serum CEA as a prognostic indicator in stage II and III breast cancer patients treated with adjuvant chemotherapy

Carcinoembryonic antigen (CEA) values in 529 patients treated in two consecutive adjuvant chemotherapy protocols were analyzed to determine if CEA values correlated with disease-free status or prognostic utility. CEA values were evaluated preoperatively, before chemotherapy, at the conclusion of chemotherapy, and during postchemotherapy followup. The sensitivity of CEA for predicting disease recurrence was low; however, any abnormal CEA at the conclusion of chemotherapy and during followup significantly correlated with reduced disease-free and overall survival. A CEA value greater than or equal to 20 ng/ml at the end of chemotherapy or during followup was highly specific and a strongly positive predictor for the presence of metastases. Abnormal CEA values before chemotherapy that became normal at the conclusion of chemotherapy were associated with a significantly reduced recurrence rate. An abnormal CEA value obtained before or after adjuvant chemotherapy is clinically useful and can provide prognostic information.

Pancreatic oncofetal antigen and carbohydrate antigen 19-9 in sera of patients with cancer of the pancreas

Pancreatic oncofetal antigen (POA) and carbohydrate antigen 19-9 (CA 19-9) were measured in the sera of 23 patients with cancer of the pancreas to determine the true positive rates of these cancer markers. In one group of unselected pancreatic cancer patients (n = 9), both tests showed above-normal results in three patients and both gave values that were within reference limits in three other patients. Two of the three remaining patients had increased CA 19-9 but normal POA values, and one patient had increased POA but normal CA 19-9 concentrations in serum. In another group of 14 pancreatic cancer patients, selected on the basis of increased concentrations of POA in serum, the CA 19-9 values were increased in eight. In four patients who had progressive disease, the concentrations of both markers increased with time in one patient, only POA in one, and only CA 19-9 concentration in another. (The fourth patient had increased but stable concentrations of POA and CA 19-9 in serum.) These data suggest that serum POA and CA 19-9 measurements should be used in combination in the evaluation of patients with cancer of the pancreas.

The correlation of vinblastine pharmacokinetics to toxicity in testicular cancer patients

The clinical pharmacokinetics of vinblastine administered by continuous 5-day infusion (3 mg/m2/day) was studied in 12 patients with primary testicular cancer. Serum vinblastine concentrations were determined by radioimmunoassay on serum collected over a 10-day period. Steady-state vinblastine concentrations were achieved within 60 to 108 hours (median, 72 hours). Vinblastine pharmacokinetics were analyzed and correlated to hematologic and nonhematologic toxicity. Hematologic toxicity was severe (granulocytopenia of less than 500/microL) in all patients; however, no correlation of vinblastine pharmacokinetics to duration of granulocytopenia or nadir was noted. Nonhematologic toxicity, however, showed a direct correlation to steady-state vinblastine concentrations. Two distinct groups of patients were identified by a toxicity score evaluating nonhematologic toxicity: as low (group A) or high (group B). The toxicity score was calculated for each patient based on accumulated toxicity during the course of treatment. The mean toxicity score for all patients was 7.11 and for groups (A and B) it was 4.0 and 9.6, respectively (P = .02). Steady-state vinblastine concentration for each patient was compared with toxicity where the mean steady-state vinblastine concentration was 7.3 ng/mL for all patients, and 5.8 ng/mL and 8.5 ng/mL for groups A and B, respectively (P = .01). These steady-state vinblastine concentrations correlated directly with the mean toxicity scores revealing that patients with high steady-state vinblastine concentrations demonstrated more nonhematologic toxicity. Application of these data to pharmacokinetically directed studies are warranted to investigate this relationship and designate dosages of vinblastine to avoid excessive toxicity.

Use of carcinoembryonic antigen in small cell lung cancer. Prognostic value and relation to the clinical course 1

Carcinoembryonic antigen (CEA) was measured in 147 patients at diagnosis of small cell lung cancer; 17% of patients with limited disease and 51% with extensive disease had an abnormal CEA level (greater than 10 ng/ml). The median level was higher in extensive than in limited disease (11 ng/ml and 5.8 ng/ml, respectively; P less than 0.001). Multivariate analysis showed CEA level greater than or equal to 50 ng/ml to be an adverse prognostic factor (P = 0.02); median survival at this level was shorter than at less than 50 ng/ml (7 and 46 weeks, respectively; P = 0.002). No consistent directional changes of follow-up CEA values were observed in patients with initially normal CEA levels, but normalization of levels occurred in complete responders. We recommend that CEA be measured in this disease at diagnosis as an additional prognostic factor and that patients with abnormal initial CEA levels have follow-up measurements to aid in evaluating response.

Reassessment of creatine kinase BB as a marker for cancer of the prostate, breast, and lung

We used an enzyme-linked immunoabsorbent assay to measure creatine kinase (CK; EC 2.7.3.2) BB in the sera of 58 cancer patients. A pre-incubation step with an anti-CK-M antibody-coated bead removed M chain components, and CK-BB was quantified with use of an anti-CK-B antibody-coated tube. No cross reactivity was observed with mitochondrial CK or CK-MM; CK-MB cross reacted slightly (1.6%). Macro CK type 1 was measured as CK-BB. Average analytical recovery of purified CK-BB added to serum was 97.7%. Although the enzyme activity of CK-BB is labile, our studies show that this protein is antigenically stable for 12 months when stored frozen. The upper limit of normal for CK-BB concentration was 0.3 micrograms/L (95th percentile, n = 25). Of the 20 cases of breast cancer of various stages, none showed any increases of serum CK-BB. Only two of 18 patients with prostatic carcinoma (stage D), and two of 10 patients with oat-cell carcinoma of the lung had increased concentrations of CK-BB in the serum. Ten patients with squamous cell cancer of the lung had normal concentration of the enzyme. Thus the CK-BB isoenzyme is not frequently increased in cancers of the prostate, lung, and breast, and it has little apparent value as a tumor marker for these diseases.

Reassessment of creatine kinase BB as a marker for cancer of the prostate, breast, and lung

We used an enzyme-linked immunoabsorbent assay to measure creatine kinase (CK; EC 2.7.3.2) BB in the sera of 58 cancer patients. A pre-incubation step with an anti-CK-M antibody-coated bead removed M chain components, and CK-BB was quantified with use of an anti-CK-B antibody-coated tube. No cross reactivity was observed with mitochondrial CK or CK-MM; CK-MB cross reacted slightly (1.6%). Macro CK type 1 was measured as CK-BB. Average analytical recovery of purified CK-BB added to serum was 97.7%. Although the enzyme activity of CK-BB is labile, our studies show that this protein is antigenically stable for 12 months when stored frozen. The upper limit of normal for CK-BB concentration was 0.3 micrograms/L (95th percentile, n = 25). Of the 20 cases of breast cancer of various stages, none showed any increases of serum CK-BB. Only two of 18 patients with prostatic carcinoma (stage D), and two of 10 patients with oat-cell carcinoma of the lung had increased concentrations of CK-BB in the serum. Ten patients with squamous cell cancer of the lung had normal concentration of the enzyme. Thus the CK-BB isoenzyme is not frequently increased in cancers of the prostate, lung, and breast, and it has little apparent value as a tumor marker for these diseases.

A double-monoclonal immunoenzymometric assay for alpha 1-fetoprotein modified for increased analytical precision

We modified a one-step, two-site, double monoclonal immunoenzymometric assay (Abbott Laboratories) for serum alpha1-fetoprotein (AFP) to increase its sensitivity and improve test precision at the low end. We increased sample size, incubation interval, and reaction time and temperature, and decreased the final reaction volume. Interassay CVs for the modified method ranged from 6.2 to 8.0% for mean concentrations of AFP in serum of 5.2 to 34.2 micrograms/L--substantially better than those for the unmodified monoclonal method--and agreed well with those of the comparison method (modified Abbott polyclonal immunoenzymometric assay). AFP values by the modified monoclonal procedure (y) correlated well with results by the polyclonal method (x): y = 0.983x + 1.84 micrograms/L (r = 0.927, n = 59). The detection limit of the modified monoclonal test was 0.2 microgram/L, as compared with 1.0 and 1.4 micrograms/L, respectively, for the modified polyclonal and the unmodified monoclonal procedures. We recommend using the modified monoclonal method for monitoring cancer patients with low tumor burden.

Underestimation of monoclonal proteins by serum protein electrophoresis on cellulose acetate

Our study of 95 serum samples from 37 patients with monoclonal gammopathy revealed distorted irregular monoclonal (M) protein bands after serum protein electrophoresis (SPE) on cellulose acetate membrane. In 71 (75%) of the 95 sera, the M-protein was underestimated and the albumin concentration overestimated. Dilution of the serum sample before SPE eliminated the abnormality of the M-protein bands. By SPE, the mean albumin concentration in these 71 undiluted sera was 45.8 (SD 7.4) g/L vs 37.9 (SD 5.8) g/L for the diluted sera; moreover, this was true of individual samples: measured albumin concentration in each diluted serum sample was always less than in the undiluted serum. As measured by the bromcresol green dye-binding method, the albumin concentration was 32.8 (SD 5.9) g/L. Similarly, the M-protein concentration in SPE was 49.5 (SD 12.3) g/L for the undiluted sera vs 61.8 (SD 15.1) g/L for the diluted sera, and the M-protein concentration in each diluted serum sample always exceeded that in the undiluted serum. Underestimation of M-protein limits the usefulness of M-protein measurement in evaluating the patient's response to therapy and for early detection of disease progression. SPE strips should be carefully inspected visually, and sera with M-protein band abnormalities should be diluted and re-assayed if SPE is to quantify concentrations of M-protein and albumin accurately.

Underestimation of monoclonal proteins by serum protein electrophoresis on cellulose acetate

Our study of 95 serum samples from 37 patients with monoclonal gammopathy revealed distorted irregular monoclonal (M) protein bands after serum protein electrophoresis (SPE) on cellulose acetate membrane. In 71 (75%) of the 95 sera, the M-protein was underestimated and the albumin concentration overestimated. Dilution of the serum sample before SPE eliminated the abnormality of the M-protein bands. By SPE, the mean albumin concentration in these 71 undiluted sera was 45.8 (SD 7.4) g/L vs 37.9 (SD 5.8) g/L for the diluted sera; moreover, this was true of individual samples: measured albumin concentration in each diluted serum sample was always less than in the undiluted serum. As measured by the bromcresol green dye-binding method, the albumin concentration was 32.8 (SD 5.9) g/L. Similarly, the M-protein concentration in SPE was 49.5 (SD 12.3) g/L for the undiluted sera vs 61.8 (SD 15.1) g/L for the diluted sera, and the M-protein concentration in each diluted serum sample always exceeded that in the undiluted serum. Underestimation of M-protein limits the usefulness of M-protein measurement in evaluating the patient's response to therapy and for early detection of disease progression. SPE strips should be carefully inspected visually, and sera with M-protein band abnormalities should be diluted and re-assayed if SPE is to quantify concentrations of M-protein and albumin accurately.

Lacto-N-fucopentaose III activity in the serum of patients with ovarian carcinoma

The expression of the lacto-N-fucopentaose III (LNF III) epitope by tumor cells of the gastrointestinal tract and ovary has been observed in tissue sections with the use of the murine monoclonal antibody GA 29-1. The presence of the LNF III epitope in the circulation of patients with colorectal cancer has also been reported. In this preliminary study, we describe the presence of LNF III activity in the serum of patients with adenocarcinoma of the ovary. Twelve of 18 (66%) patients with stage I-IV disease demonstrated high reactivity to the GA 29-1 monoclonal antibody. This serum reactivity was independent of disease stage and histologic cell type. In contrast, only 1 of 6 control patients demonstrated a false-positive level of reactivity to GA 29-1. These preliminary results suggest that LNF III warrants further study of its potential application as a serum tumor marker test in patients with adenocarcinoma of the ovary.

Leukocyte interferon conjugated to beta-D-galactosidase for receptor studies: a preliminary report

In this report, we describe a simple procedure for labeling leukocyte interferon (IFN) with the enzyme beta-D-galactosidase (beta-gal), through a low-molecular-weight linking agent. The labeling procedure did not appear to alter the properties of the enzyme, the antiviral activity of the recombinant DNA-derived human clone A leukocyte IFN (rIFN alpha A), or the ability of the IFN to bind to Daudi cells. The binding of the putative IFN-enzyme complex to Daudi cells could be blocked by excess unlabeled IFN or by monoclonal antibody to IFN receptors. Our preliminary results indicate that enzyme labeling of IFN can result in a biologically active agent that has potential for use in the quantitation of circulating IFN and specific receptor proteins. In addition, this labeling procedure may be potentially useful with other immunomodulatory proteins.

Pharmacokinetic study of partially pure gamma-interferon in cancer patients

A pharmacokinetic study was performed with partially pure immune (gamma) interferon (IFN-gamma) in patients with metastatic cancer. Nine patients were given IFN-gamma by the i.m. route in doses ranging from 1.5 X 10(5) to 9.6 X 10(6) antiviral units. There was no detectable antiviral activity in patients' serum, and only minimal side effects were observed. Fifteen patients were given IFN-gamma by i.v. bolus infusion in doses ranging from 1.5 X 10(5) to 54 X 10(6) units. Serum clearance of antiviral activity was described by a monoexponential disappearance curve. The serum half-life was dose dependent (3 min at the lower doses and 34 min at the highest doses). There were few consistent biological effects observed in the patients. Based on these pharmacokinetic data, eight patients were treated by a 6-hr continuous infusion consisting of 3 X 10(6) units by i.v. bolus followed by 4 X 10(6) units/hr for 6 hr. This regimen resulted in consistent serum levels of IFN-gamma ranging from 40 to 60 units over the 6-hr period. Marked granulocytopenia occurred within 24 hr and was sustained during the 10-day infusion period. There was marked increase in serum beta 2-microglobulin. We conclude that, in order to induce consistent serum antiviral activity, partially pure IFN-gamma, because of its rapid serum disappearance curve, must be administered by continuous i.v. infusion.

Diagnostic and prognostic values of carcinoembryonic antigen in medullary carcinoma of the thyroid

This study was conducted to compare the diagnostic and prognostic values of calcitonin (CT) and carcinoembryonic antigen (CEA) measurements in patients with medullary carcinoma of the thyroid (MCT). Plasma CEA, basal CT (BCT), and peak CT (PCT) levels after pentagastrin stimulation were determined in 15 patients with occult familial MCT before treatment. CEA was elevated in 11 patients (73%), BCT in 8 patients (53%), and PCT in all patients. The prognostic values of serial CEA and BCT determinations were studied in 24 other MCT patients followed for a period of 1.5-8 yr. These patients were divided into 2 groups; group I consisted of 11 patients with normal serial CEA levels, and group II consisted of 13 patients with persistently elevated CEA levels. In group I, 2 patients (18%) developed metastases compared with 10 patients (77%) in group II (P less than 0.001). Elevated BCT levels were found in all patients with metastases as well as in 63% of patients with no evidence of disease. When CEA time curves were studied, three patterns could be identified: 1) a steep slope with rapidly rising CEA levels in patients with rapidly progressive disease, 2) a flat slope in patients with no metastasis or nonprogressive disease, and 3) an intermediate type of slope in patients with slowly progressive disease. Slope analysis of BCT time curves was not done because of marked fluctuations in BCT levels. PCT after pentagastrin stimulation remains the best diagnostic marker for MCT. However, CEA may be a better prognostic indicator, as it discriminated more efficiently between patients with and without metastasis. Slope analysis of CEA time curve and calculation of the doubling time are recommended, as they correlated well with the course of the disease.

Serial plasma carcinoembryonic antigen measurements during treatment of metastatic breast cancer

Serial plasma carcinoembryonic antigen (CEA) levels were measured in 167 patients with metastatic breast cancer treated with fluorouracil, doxorubicin hydrochloride, and cyclophosphamide (FAC). In 84 patients, the pretreatment CEA value was abnormal. Response rates and remission duration were similar in patients with normal and abnormal pretreatment levels. Carcinoembryonic antigen concentrations decreased in 94% of patients who responded to FAC therapy. The duration of response was 22 months for patients in whom the CEA levels normalized v nine months in those in whom it decreased but never returned to normal. Increasing CEA levels correlated with progressive disease in 87% of patients and preceded clinical evidence of progression in 77%. Serial CEA measurements can monitor response to chemotherapy, provide useful prognostic information of response duration, and detect progressive disease early.