Klin Wochenschr (1989) 67:813-817 Klinische

Wochen- schrift

© Springer-Verlag 1989

Mucin-Like Cancer-Associated Antigen (MCA) Compared with CA 15-3 in Advanced Breast Cancer

G.G. Steger ~, R. Mader ~, K. Derfler 2, K. Moser 1, and C. Dittrich 1 1 Universit/itsklinik fiir Chemotherapie, 2 I. Medizinische Universit/itsklinik, Universit/it Wien, ~sterreich

Summary. Mucin-like cancer-associated antigen (MCA), a new tumor marker using the mouse monoclonal antibody b-i 2 is thought to be of value in the management of patients with breast cancer. In this study sera from 191 female patients with breast cancer (112 with progressive disease [PD] and 79 with no evidence of disease [NED]) were analyzed for MCA levels and compared with those of cancer antigen 15-3 (CA 15-3) in single determi- nation and in combination with carcinoembryonic antigen (CEA) and tissue polypeptide antigen (TPA). A cut-off level of 14 U/ml for MCA seems to be more appropriate than the recommended 11 U/ml to distinguish between PD and N ED in patients with breast cancer. Although there was a fairly good correlation of MCA to CA 15-3, MCA was inferior in sensitivity and specificity to CA 15-3. Patients with osseous metastases and those with more than one metastatic site showed higher MCA levels than patients with visceral or soft tissue metastases, a fact which was comparable to CA 15-3. Combining MCA and CA 15-3 re- sulted in a gain in specificity but marked loss of sensitivity. The combination of MCA and CEA results also in a loss of sensitivity whereas the com- bination of CA 15-3 and CEA showed an increased specificity and only a negligible loss of sensitivity. The combination of MCA with TPA is of little value in the follow-up of breast cancer, as is the combination of CA 15-3 with TPA. The combina- tion of CA 15-3 with CEA can be still recom- mended tbr follow-up for early detection of metas- tases in breast cancer. Further investigations with

Abbreviations: MCA=mucin-like cancer-associated antigen; CA 15-3=cancer antigen 15-3; CEA=carcinoembrionic anti- gen; TPA=tissue polypeptide antigen; NED=no evidence of disease; PD=progressive disease; mab=monoclonal anti- body; RIA=radioimmunoassay; EIA: enzymimmunoassay; IRMA : immunoradiometric assay

MCA to determine the lead time of this new marker as well as its role as a possible monitor of the response to therapy are recommended.

Key words: Tumor marker - MCA - CA 15-3 -- Breast cancer

The development ofmonoclonal antibodies (mabs) against tumor-associated antigens represents a ma- jor advance for diagnosis and monitoring of thera- py in clinical oncology. For breast cancer the can- cer antigen 15-3 (CA 15-3), a marker determined by two mabs (115 D8, D F 3) directed against two different epitopes of the same molecule [9-11, 18] has been shown to have more specificity as well as sensitivity in advanced stages of the disease than the formerly used markers CEA and TPA. The combination of CA 15-3 and CEA shows even a greater sensitivity and an acceptable specificity than for each marker alone [1, 14, 23, 24], so that this combination is recommended as standard de- terminations in the follow-up of breast cancer pa- tients and it also appears to be of some value for early detection of metastases [15, 17, 19] as well as in monitoring the response to therapy for ad- vanced disease [5, 16]. However, there is a high rate of false-positive marker values in patients with NED so that there is a need for markers with a higher specificity and at least comparable sensitivi- ty.

One of these new markers which is presently under investigation is MCA. This marker, a mucin- like glycoprotein with a molecular weight of 350000 Daltons is recognised by mab b-12 [4, 20, 21]. Although it was detectable in the cytosol of breast cancer and normal breast tissue, MCA levels in the cytosol of carcinoma tissue were significantly higher than in those of normal tissue [25]. Early

814 G.G. Steger et al.: MCA in Advanced Breast Cancer

results of testing for MCA serum levels from pa- tients with primary, untreated breast cancer led to the suggestion that serum MCA levels correlate with serum CA 15-3 and also that MCA levels are correlated with the degree of axillary lymph node invasion [4].

The aim of our investigation was to determine the value of serum MCA compared with CA 15-3 in advanced breast cancer and for employing both markers simultaneously and also with CEA and TPA.

Materia l and Methods

Patients

Sera from 191 female patients (median age, 57 years; range, 27-85 years) with histologically confirmed breast cancer were examined for the markers MCA, CA 15-3, CEA and TPA. Of these patients 112 (median age, 57.5years; range, 27- 85 years) had metastatic disease. PD was evaluated clinically and confirmed using radiological meth- ods. The control group consisted of 79 patients (median age, 56 years; range, 29-80 years) with clinically and radiologicaUy confirmed NED who had undergone modified radical mastectomy at least 3 months prior to entry into the study.

Patients with NED and raised marker levels were reevaluated 6 to 8 weeks after the first marker determination and in case of signs of metastases excluded from the study; this resulted in three pa- tients being excluded.

Materials

Commercially available marker kits were used. The cut-off values determined were as follows: 26 U/ml for CA 15-3 (RIA, Centocor; detection range, 0-200 U/ml), 5 ng/ml for CEA (EIA-Monoclonal, Abbott Laboratories; detection range, 0-60 ng/ ml), and 55 U/1 for TPA (Prolifigen TPA IRMA, Sangtec Medical; detection range, 0-2000 U/l). For MCA (EIA, Roche Diagnostica; detection range, 0-30 U/ml), a two-step, solid-phase enzy- mimmunoassay, a cut-off point of 14 U/ml was used. Marker levels were determined in duplicate. Sera were stored at - 2 0 ° C until assayed and were not diluted if values were out of range of the test system. Sensitivity, specificity, and the positive pre- dictive values were calculated as published [7, 22].

Sensitivity-Specificity Diagrams

For demonstrating the results and to compare the tested markers sensitivity-specificity diagrams were

used. In these diagrams the sensitivity rates and specificity rates are plotted using different cut-off values for each marker. Ideally the resulting curve should extend as far as possible to the upper left- hand corner (100% sensitivity/100% specificity) of the diagram [12, 13]. For using this method for marker combinations a by us developed and re- cently published score was established [23, 24].

Statistics

Linear regression and the Wilcoxon ranked sum test were used for statistical analysis.

Results

The rates for sensitivity and specificity of the markers tested in this investigation can be seen in Table 1. For MCA 14 U/ml was employed as the cut-off value in this investigation because this value was more appropriate to distinguish between patients with PD and NED. Additionally at this cut-off MCA showed similar specificity to CA 15-3 which is necessary to compare sensitivity rates of two different markers.

There was a fairly good overall correlation of MCA with CA 15-3 (r=0.84, P<0.0001; n=191; Fig. 1). In patients with PD correlation (r=0.78,

Table 1. Rates for sensitivity and specificity of CA 15-3 and MCA at various cut-off values

Marker Cut-off Sensitivity Specificity (U/mt) (%) (%)

MCA 11 86 71 12 85 80 13 81 82 14 80 87 15 75 95

CA 15-3 26 87 89

CA 15- 3 ( U/ml ) n = 191 r = 0 8 4

250- p < 0 0001

200- • " . ' . - - "

150 ,

100-

°i "o.'2" .; " " : "

0 5 10 15 20 25 30 35 MCA ( U / m l )

Fig. 1. Correlation of CA 15-3 and MCA serum levels in pa. tients with PD and NED

G.G. Steger et al.: MCA in Advanced Breast Cancer 815

200--

150 -

E

0 3

,2-,_ lOO- (..P

75--

F- 1

t.

I

i i

!

l l t

ds vis Nb (n = 36)(n = 20)(n= 28) (n = 28)(n = 79)

Fig. 2. CA 15-3 serum levels and metastatic site (os, osseous; vis, visceral; no, lymph nodes; m, mixed; NED, no evidence of disease)

30-

25-

20-

E

v

~E 10-

5-

!

4t-

I t E

OS VIS NO

SPECIFICITY % 1OO 80 60 50

100 t i

t--- c~

uJ .3 t/)

60 - . ~ A l CEA ~ D.--~ TPA

50 1' / r :t91 Fig. 4. Sensitivity-specificity diagram comparing CA 15-3, MCA, CEA, and TPA

SPECIFICITY

~Z i00 8O 60 50

l l I l 100

' NIED M

6 0 ~ - ' - ' ~ MCA + CA 15,3

n = 191 5 0

Fig. & Sensitivity-specificity diagram comparing CA 15-3, MCA, and their combination

3). Sensitivity and specificity rates of the markers ~4 NED tested are seen in Fig. 4. The positive predictive

(n :36) (n=20)(n =28) (n=28) (n=79)

Fig. 3. MCA serum levels and metastatic site (os, osseous; vis, visceral; no, lymph nodes; m, mixed; NED, no evidence of disease)

P<0 .0001 ; n = 112) was slightly better than it was found to be with primary breast cancer (r=0.66, P<0 .00001 ; n = 145 [4]). Comparable were the rel- atively higher marker levels o f MCA and CA 15-3 in patients with osseous metastases and in those with more than one metastatic site (Figs. 2 and

values, as an index for the probability of a marker value above the cut-off to be a true positive, were 0.90 for CA 15-3 and 0.88 for MCA, and they are therefore comparable.

Using the T-score the combination of CA 15-3 and MCA resulted in a gain in specificity but no higher sensitivity as compared with using CA 15-3 alone at a cut-off value of 26 U/ml (sensitivity 87%, specificity 89%; Fig. 5). This combination with a sensitivity of 84% and specificity of 93% also showed no significant benefit compared with the standard combination of CA 15-3 with CEA

816 G.G. Steger et al. : MCA in Advanced Breast Cancer

S P E C I F I C I T Y

I00

8 0 -

H

H

Z

6 0 -

1 O0 80 60 50

I I I I

~,-~& 7- Z~----A CEA + CA 15.3

/?T/ ; % k . ..... o C A+.CA I~." eft" l r - - - -A TPA +MCA

n = 191

5 0

Fig. 6. Sensitivity-specificity diagram comparing CA 15-3 alone and MCA and CA 154 in combination with CEA and TPA

(sensitivity 86%, specificity 90%; Fig. 6). The si- multaneous determination of MCA and CEA re- sults in an increase in specificity, but this is accom- panied by a marked loss of sensitivity. The combi- nation of MCA and TPA or CA 15-3 with TPA showed no particular advantage.

Discussion

In this study the tumor markers tested were com- pared in patients with metastasized breast cancer and patients after mastectomy for malignant breast disease who served as control group. The patients with NED were chosen because this situation re- flects clinical reality in an oncologic outpatient ward and only in such patients can a new marker prove to be of value in the follow-up of breast cancer patients.

In the patient groups tested MCA shows a good correlation with CA 15-3 and also the rela- tively higher marker levels in patients with osseous and mixed metastases are comparable. These data are in accordance with those reported previously by other authors [2]. A good correlation of these two markers was also demonstrated by Bombar- dieri et al. [4], who reported a correlation of r = 0.663 ( P < 0.00001) in the sera of 145 patients with primary breast cancer with or without lymph node involvement. However, in our study the percentage

of correct positive marker values of MCA in the sera of patients with metastasized breast cancer is lower than that of CA 15-3, which was the only marker reaching the 85%/85% quadrant of the sensitivity/specificity diagram. These findings lead to the conclusion that MCA might be inferior to CA 15-3 for screening of distant metastases in pa- tients after mastectomy for mammary carcinoma and therefore the determination of MCA alone in the sera of breast cancer patients can not be recom- mended for routine follow-up programs.

The combination of MCA and CA 15-3 results in no gain in sensitivity compared with the determi- nation of CA 15-3 alone. Also simultaneous deter- mination of MCA and CEA results in no benefit. In contrast simultaneous determination of CA 15- 3 and CEA, allthough again discussed controver- sially [6], seem to be superior in sensitivity and specificity compared with CA 15-3 alone, so that this marker combination should be used as an ad- ditional screening method in the follow-up of breast cancer patients.

To avoid obvious false-positive marker values, which would be correctly positive in truth, three patients had to be excluded from evaluation be- cause of the detection of metastases shortly after the first determination of the tumor markers. Re- trospective analyses of the sera from these three patients drawn at several control visits showed that in two progression of disease was first indicated by MCA and in one by CA 15-3. This observation of a possibly longer lead time for MCA has also been made by others [2, 3], but in this study also patients at "high risk," defined as patients in com- plete remission after radiotherapy for metastases or local recurrences, served as control group for a group of patients with generalized disease (n = 74). Therefore, the control group is too heteroge- neous to answer the question whether there exists a longer lead time for MCA in patients after mas- tectomy for primary breast cancer than is known for CA 15-3, which is about 3-6 months.

Allthough MCA seems to be less sensitive than CA 15-3 it might add to CA 15-3 and CEA in early detection of metastases from breast cancer because of that possibly longer lead time. Additionally, the fact that MCA recognises a different antigen than CA 15-3 [21], points to a role of MCA to be of advantage in monitoring therapy or early detection of secondary resistance of the tumor to therapy. To answer these questions further prospective stu- dies with serial determinations of MCA and CA 15-3 in patients during follow-up programs after mastectomy for breast cancer and during therapy for advanced disease are needed.

G.G. Steger et al. : MCA in Advanced Breast Cancer 817

References

1. Bieghnayer C, Szepesi T (1985) First experiences with CA 15-3 a tumor marker for breast cancer. In: Bastert G (ed) International symposium of monoclonal antibodies in clinical oncology. University of Homburg/Saar S 4 (Ab- stract)

2. Bieglmayer C, Szepesi T, Neunteufel W (1988) Follow-up of metastatic breast cancer patients with a mucin-like carci- noma-associated antigen: comparison to CA 15-3 and car- cinoembryonic antigen. Cancer Lett 42:199---206

3. Bieglmayer C, Szepcsi T, Neunteufel W (1988) MCA - ein neuer Tumormarker. Gynfik Rdsch 28 [Suppl 2] : 203-205

4. Bombardieri E, Gion M, Mione R, Dittadi R, Bruscagnin, Buraggi G (1989) A mucinous-like carcinoma-associated antigen (MCA) in the tissue and blood of patients with primary breast cancer. Cancer 3:490-495

5. Colomer R, Sole LA, Navarro M, Encabo G, Ruibal A, Salvador L (1986) CA 15-3: early results of a new breast cancer marker. Anticancer Res 6 : 683-684

6. Delarue JC, Mouriesse H, Dubois F, Friedmann S, May- Levin F (1988) Markers in breast cancer: does CEA add to the detection by CA 15-37 Breast Cancer Res Treat 11 : 273-276

7. Fleig JL (1973) Statistical methods for rates and propor- tions 3. Wiley, New York

8. Eskelinen M, Tikanoja S, Collan Y (1988) A new tumor marker MCA in breast cancer diagnosis. Anticancer Res 8 : 665-668

9. Hilkens J, Hilger J, Buijs F (1983) Monoclonal antibodies against human milk-fat globulin membranes useful in carci- noma research. In: Peeters H (ed) Protides of the biological fluids 31. Pergamon, Oxford, pp 1013-1016

10. Kufe DW, Nadler L, Sargent L (1983) Biological behaviour of human breast carcinoma-associated antigens expressed during cellular proliferation. Cancer Res 43 : 851-857

11. Kufe DW, Inghirami G, Abe M, Hayes D, Just-Wheeler H, Schlom J (1984) Differential reactivity of a monoclonal antibody (DF 3) with human malign versus benign breast tumors. Hybridoma 3 : 223-232

12. Makuch RW, Muenz LR (1987) Evaluating the adequacy of tumor markers to discriminate among distinct popula- tions. Semin Oncol 2:8%101

13. Oehr P, Derigs G, Altmann R (1981) Evaluation and char- acterization of tumor-associated antigens by conversion of inverse distribution function values into specificity-sensitivi- ty diagrams. Tumor Diagnostic 2:283-290

14. Paulick R, Caffier H, Kaesemann H (1986) Erste Erfahrun- gen mit dem monoklonalen Markersystem CA 15-3 bei Mammakarzinompatientinnen. Tumor Diagnostik Thera- pie 7:85-87

15. Pons-Anicet DMF, Krebs BP, Mira R, Namer M (1987) Value of CA 15-3 in the follow-up of breast cancer patients. Br J Cancer 55 : 567569

16. Ruibal A, Encabo G, Colomer R, et al. (I985). CA 15-3 serum levels in clinical practice. Our experience in 1177 cases. XI. Congresso nazionale di oncologia (SIPDTT), 12-14 December 1985, Bologna, Italy

17. Sacks MPM, Stacker SA, Thompson CH (1987) Compari- son of mammary serum antigen (MSA) and CA 15-3 in the serum of patients with breast cancer. Br J Cancer 56 : 820-824

18. Schlom J, Greiner J, Horan-Hand P (1984) Monoclonal antibodies to breast cancer-associated antigens as potential reagents in the management of breast cancer. Cancer 54: 2777-2794

19. Schmidt-Rhode P, Schulz KD, Sturm G, Raab-Frick A, Prinz H (1987) CA I5-3 in breast cancer: first experiences with a monoclonal test system. Med Sci Res 15 : 765-766

20. St/ihli C, Takacs B, Miggiano V, Staehlin T, Carman H (1985) Monoclonal antibodies against antigens on breast cancer cells. Experienfia 41 : 1377-1381

21. Stfihli C, Caravatti M, Aeschbacher M, Kocyba C, Takacs B, Carmann H (1988) Mucin-like carcinoma-associated an- tigen defined by three monoclonal antibodies against differ- ent epitopes. Cancer Res 48 : 6799-6802

22. Staquet M, Rosencweig M, Lee YJ, Muggia FM (1981) Methodology for the assessment of new dichotomous diag- nostic tests. J Chronoc Dis 34:599-605

23. Steger GG, Mader R, Dittrich Ch, Moser K (1987) Conse- quences of the combination of various tumour markers with regard to the sensitivity and specificity in cases of breast carcinoma. In: Klapdor R (ed) New turnout markers and their monoclonal antibodies. Thieme, Stuttgart-New York, pp 52-56

24. Steger GG, Mader R, Dittrich Ch, Janata O, Eichler HG, Moser K (1987) CA 15-3, CA 50, CEA, TPA and a newly developed tumour marker score in breast cancer. J Tumor Marker Oncol 3:261 267

25. Zenklusen HR, yon Overbeck J, St~ihli C, Gudat F, Rolink J, Heitz PU (1988) The immunohistochemical reactivity of a new anti-epithelial monoclonat antibody (MAb b-12) against breast carcinoma and other normal and neoplastic human tissue. Virchows Arch (Pathol Anat) 413:3-10

Received: December 14, 1988 Returned for revision: February 2, 1989 Accepted: June 1, 1989

Dr. Gfinther G. Steger Univ.-KIinik f/ir Chemotherapie Universitfit Wien Lazarettgasse 14 A-1090 Wien Osterreich