- Email: [email protected]
The role of cancer antigen 125 (CA 125) in the management of ovarian epithelial carcinomas
GYNECOLOGIC
ONCOLOGY
30, 26-34 (1988)
The Role of Cancer Antigen 125 (CA 125) in the Management of Ovarian Epithelial Carcinomas MARKO M. ALTARAS, M.D.,*,’ GARY L. GOLDBERG, M.B.CH.B., M.R.C.O.G.,*,* WILFRED LEVIN, M.B.CH.B., F.F. RADIO (T), S.A. M.Med. Rad(T),t BASIL BLOCH, M.B.CH.B., F.R.C.O.G., M. Med. (0 & G),* LYNNE DARGE, B. Sc.,$ AND JAMES A. SMITH, M.B.CH,B., M.Med. RAD (D), M.Med.S *Division of Gynaecological Oncology, Department of Obstetrics and Gynaecology, tDivision Gynaecological Oncology, Department of Radiotherapy, and iDepartment of Nuclear Medicine, Groote Schuur Hospital and University of Cape Town, Rondebosch 7700 South Africa
of
Received November 6. 1986 From June 1, 1984, to May 31, 1985, 98 cases of epithelial ovarian carcinomas were assessed and followed prospectively using a new murine monoclonal antibody OC 125 which detects the antigen CA 125. Serous tumors comprised 43.7% of cases, mutinous tumors 20.4%, endometrioid tumors 16%, and other epithelial tumors 19.4%. Tumors of low malignant potential and benign epithelial cystadenomas were not included. For this study the upper limit of normal for CA I25 was 20 U/ml. Thirty-six were new cases. In this group the initial CA 125 levels >20 U/ml, >35 U/ml, and >65 U/ml were 97.2, 94.4, and 86. I%, respectively. When mutinous types were excluded the specificity rate did not change significantly. There was no significant difference in initial CA 125 levels between early stages I and II and late stages III and IV. No correlation between tumor bulk and the serum level of antigen was observed. The remaining 62 patients were being followed and in this group 50 were considered to be in remission. Six cases in the remission group had elevated CA 125 levels >20 U/ml and 5 of these developed clinical recurrence. The correlation between the clinical status and concordant fluctuations in the serum levels of CA 125 in all histological types was 87.8 and 93.5% when 10 cases of mutinous tumors were excluded. The contingency coefficient was 0.746. Seven SLOs were performed. All had CA 125 levels ~20 U/ml and the mean was 6.9 U/ml. Only 1 case was positive with microscopic disease. In our experience CA 125 was invaluable in the management and follow-up of patients with ovarian carcinoma especially for the early detection of recurrent disease and for the monitoring of patients on therapy. Q 1988 Academic PXSS, IX.
INTRODUCTION
In 1956 Witebsky et al. described and characterized cancer-associated antigens including ovarian cancer-associated antigens [I]. In the past 15 years attempts ’ To whom reprint requests should be addressed at Department of Obstetrics and Gynecology “A”, Gynaecological Oncology Unit, Sapir Medical Center, Meir Hospital, (SacPer Medical School, University of Tel Aviv), Kfar Saba, Israel. ’ To whom all correspondence should be addressed at Albert Einstein College of Medicine, Department of Gynecology, Division of Gynecologic Oncology, Pelham Parkway and Eastchester Road, Bronx, NY 10461. OO90-8258/88$1.50 Copyright All rights
0 1988 by Academic Press. Inc. of reproduclion in any form reserved.
26
OVARIAN
EPITHELIAL
CARCINOMA
21
TABLE 1 BREAKDOWN
OF PATIENTS
IN THE STUDY
Group 1
Group II
29 on treatment 3 lost to follow-up 4 no adjuvant therapy (surgery alone) Total 36 Total 98
12 on treatment 50 remission
62
have been made to isolate polyclonal antibodies derived from different types of epithelial carcinomas [2]. None of these have become commercially available. In 1981, Bast and co-workers developed the first monoclonal antibody OC 125 derived from an epithelial ovarian carcinoma cell line (OVCA 433) in order to detect ovarian tumors [3]. OC 125 reacts with an antigen, cancer antigen 125 (CA 125), which is common to most nonmucinous epithelial ovarian carcinomas and an assay was developed to detect CA 125 in serum [4]. This antigen is not expressed in fetal or normal adult ovarian tissue [5]. It has been demonstrated in fetal serosal surface epithelia, fallopian tube, endometrium, and cervix using the biotin-avidin peroxidase technique [6]. The same authors described serum levels of the antigen ~65 U/ml in 0.2% of apparently healthy males, 0% in healthy females, 2.1% in patients with benign disease, and 22% in patients with nongynecological cancer. In patients with histologically proven carcinoma of the ovary 74.3% had levels ~65 U/ml [4]. This antigen was detected in cryostat sections of nonmucinous epithelial ovarian tumors, but was not demonstrated in mutinous, Brenner, sex cord, and germ cell tumors of the ovary [5]. The purpose of this study was to evaluate the usefulness of CA 125 in the management and follow-up of patients with epithelial ovarian tumors of all histological types. PATIENTS, MATERIALS,
AND METHODS
In the 12-month period from June 1984, 98 consecutive cases of carcinoma of the ovary were prospectively evaluated. The patients were divided into two groups (Table 1). Group I comprised 36 new patients who had CA 125 serum levels evaluated prior to initial surgery or in the early postoperative period and Group II consisted of 62 patients seen at the follow-up clinic. The mean age of the patients was 51.2 years and the range was 23 to 79 years. The breakdown according to stage was 35 cases Stage I, 11 cases Stage II, 37 cases Stage III, and 15 cases Stage IV. The histological types included 43.7% serous, 20.4% mutinous, 16.3% endometrioid, and 19.4% other types which included clear cell (2), malignant Brenner (l), mixed type (7), and undifferentiated (9). Tumors of low malignant potential were not included in this study.
28
ALTARAS
ET AL.
During the study period the shortest follow-up was 1 month in 3 cases and the longest period of follow-up was 11 months in 5 cases. All new patients with histologically proven ovarian cancer were registered and followed up at the Combined Gynaecology Radiotherapy Assessment Clinic. Routine follow-up procedures included examination, chest radiography, ultrasound of the abdomen, CT scan, and aspiration cytology of the cul-de-sac. Routine hematological and biochemical parameters were assessed. The assay for CA 125 was introduced in our hospital in June 1984 and it became a routine investigation for patients on treatment and on routine followup. Patients suspected of having ovarian carcinoma or a pelvic mass had CA 125 evaluated preoperatively. We used the following terms to define the clinical response: Remission: where the patient had completed therapy and was clinically free of disease. Regression: where surgery reduced the tumor burden by over 50% and this clinical situation was maintained for at least 3 months by the administration of chemotherapy or in instances where chemotherapy produced marked objective and subjective improvement. This was assessed using clinical, radiological, and ultrasonic examinations. Stabilization: where the detectable tumor masses were neither progressing nor regressing and the patient’s general condition was stable. Progression: where new lesions appeared and clinical examinations related to the disease occurred. The biochemical counterparts were Remission: CA 125 <20 U/ml. Regression: ~50% decrease in serial measurements of CA 125. Stabilization: where CA 125 levels varied ~50% in serial values. Progression: >50% increase in serial measurements of CA 125. Cul-de-sac aspiration cytology was performed as described by Goldberg et al. [71. CA 125 was determined by a monoclonal antibody assay. Each sample was determined by an immunoradiomanometric assay (Centocor Inc., Malvern, PA 19355). The system is a simultaneous sandwich solid phase RIA. The patient’s plasma was stored at -4°C and defrosted prior to the assay. Polystyrene beads coated with mouse monoclonal antibody to CA 125 were incubated with the serum specimen, standards, and tracer. The tracer was ‘*‘Ilabeled monoclonal anti-CA 125 antibody. During the incubation the CA 125 formed sandwich complexes with the monoclonal antibodies. Unbound materials present in the specimen were removed by aspiration of the fluid and washing of beads. The bound radioactivity was determined by counting beads in the gammascintillation counter which was proportional to the concentration of the CA 125 in the specimen within the working range of the assay. A linear standard curve
OVARIAN
EPITHELIAL
29
CARCINOMA
TABLE 2 SPECIFICITY
All patients Nonmucinous Mutinous
RATES FOR
CA 125
No.
>20 U/ml
>35 U/ml
>65 U/ml
36 29 7
35 (97.2%) 29 (100%) 6 (85.7%)
34 (94.4%) 29 (100%) 5 (70.1%)
31 (86.1%)* 27 (93.1%)** 4 (57.1%)***
*,** P, NS. **.*** P < 0.04 (x2).
was obtained by plotting the CA 125 concentration of the standards versus bound radioactivity. The CA 125 concentration of the unknowns and the control ran concurrently with the standards and could be determined from the standard curve. The observed range for the concentration of CA 125 in serum or plasma of healthy females has been determined by the manufacturer as 10.7 + 9.6 U/ml (mean + SD) and the lower limit of sensitivity is 7 U/ml. Levels below this are determined by extrapolation. In view of these data we used 20 U/ml as the upper limit of normal. We considered a deviation of 50% in antigen level significant. Levels >500 U/ml were estimated by extrapolation although a dilutional technique was available if necessary, The interassay coefficient of variation calculated according to control provided by the manufacturer was 10.1%. A total of 367 assays were performed in this study group. In 23 patients being followed only one serum CA 125 assay was available. Most of these cases had Stage I disease and were in remission. A second-look operation (SLO) was performed in 7 cases. All patients had at least two CA 125 levels prior to SLO. Statistical methods used were chi square (x2), Fisher’s exact test, and contigency coefficient [S]. RESULTS Table 2 depicts the specificity rate of CA 125 in the 36 new patients with respect to histological type and various serum levels. Table 3 shows the correlation between the stage of the disease and the initial levels of CA 125 in the 36 new cases. Nine out of 10 patients (90%) who presented
CLINICAL
Clinical stage I II III
TABLE STAGE AND
3
CA 125 LEVELS
20 U/ml
<20-35 U/ml
36-65 U/ml
-
-
1 2
IV
1
1
Total
1
1
-
3
>65 U/ml 8 1 11
Number 9 1 13
11
13
31
36
30
ALTARAS
ET AL.
TABLE 4 HISTOLOGICAL
Histology Serous Mutinous Endometrioid Others Mixed (4) Brenner (1) Undifferentiated (3) Total %
TYPE IN RELATION
Number of cases
TO
CA 12.5LEVELS
<20 U/ml
20-35 U/ml
-
-
15 7 6 8
1
36
36-65 U/ml
1
>65 U/ml
1 1 1
14 4 5 8
31
-
-
-
1 2.8
1 2.8
3 8.3
with Stage I or II disease and 22 out of 26 (84%) of patients with Stages III and IV had CA 125 levels >65 U/ml. There was no statistically significant difference between these groups. Table 4 depicts the different levels of CA 125 at the time of diagnosis and shows its relationship to the histological type of tumor. Of the 62 patients in Group II being followed, 50 were in remission. In 44 of these patients the serum levels of CA 125 were <20 U/ml and in the 6 remaining patients the serum levels were >20 U/ml. The 44 patients with levels ~20 U/ml have remained clinically free of disease during the study period. Their serum levels ranged from 0 to 15.8 U/ml in 110 assays. The mean was 5.69 k 7.39 U/ml. Table 5 summarizes the clinical status of the disease in relation to fluctuations in the CA 125 levels. This analysis comprised 29 new patients and 12 patients from the follow-up group who were on treatment during the study period (see Table 1). The variance in CA 125 levels correlated with the clinical status in 36 of the 41 patients (87.8%). When the 10 patients with mutinous tumors were excluded the correlation was correct in 29 of 31 patients (93.5%). In 7 out of 10 patients with mutinous tumors there was a positive correlation between clinical status of the disease and the antigen levels. The contingency coefficient was obtained among the regression, stabilization, and progression groups using three cutoff levels of the antigen. The result was found to be 0.746 which is almost
CORRELATION
CA 125 >50% decrease <50% variation >50% increase Total
OF
TABLE 5 CA 125 LEVELS AND
CLINICAL
STATUS
Regression
Stabilization
Progression
21 1 0 22
0 4 0 4
0 4 11 15
OVARIAN
EPITHELIAL
CARCINOMA
31
perfect concordance. There were 5 inconsistent results. In 2 cases there was clinical progression of the disease which was not reflected with serial CA 125 assays. One of these patients had a malignant Brenner tumor and the other had an endometrioid carcinoma. Although the levels of CA 125were high, the increments were not >50% and did not reflect clinical progression of disease. In 2 cases there was clinical progression of disease but the CA 125 levels did not increase significantly. Both these patients had mutinous tumors. In 1 case the CA 125 level plateaued for 2 months and the patient was found to have bone metastases. We performed seven SLOs during this time. All were clinically assessed to be free of disease and had negative chest radiology, CT scan, and cul-de-sac aspiration cytology and preoperative CA 125 assays ~20 U/ml. One laparotomy was positive revealing microscopic disease only. DISCUSSION Over the past 5 to 10 years there has been an improvement in the survival of patients with epithelial ovarian carcinomas [9]. This is probably due to aggressive cytoreductive surgery, new combination chemotherapy protocols, and adjuvant radiotherapy [lo-141. However, no significant further improvement in survival can be expected without the availability of a reliable, noninvasive means for monitoring the clinical course of the patient. The major problems facing the clinician are the optimal duration of chemotherapy, the timing and need for second-look operation, and the reliable assessment of patients in remission [15-181. Another problem is the early detection of recurrent disease, especially tumor nodules less than 2 cm in diameter, and disease in the upper abdomen [18]. These recurrences are only diagnosed once the tumors become clinically evident. The immunoradiometric assay for CA 125 was originally described for the detection of nonmucinous epithelial ovarian cancers [4]. In two studies the presence of CA 125 could not be demonstrated in cryostat sections of mutinous and Brenner tumors [5,19]. We were thus surprised to find markedly elevated CA 125 levels in mutinous tumors at the start of this investigation. We continued to monitor CA 125 in all suspected cases of ovarian tumors and in all patients being followed up irrespective of the histological type. We determined the specificity rate of CA 125 in 36 new cases of ovarian epithelial cancer of all histological types using 20 U/ml as the upper limit of normal (see Table 3). The specificity rate was found to be 97.2% at levels ~20 U/ml, 94.4% at levels >35 U/ml, and 86.1% for values ~65 U/ml when all histological types were evaluated. When mutinous tumors were excluded the specificity rate was 93.1% for levels >65 U/ml and 100% for levels ~35 U/ml. The high specificity rate in our study is similar to that reported by Bast et al. and predicts its usefulness in clinical practice. It has allowed us to explore its use as an indicator of the response to treatment. Our results show the specificity rate for mutinous tumors at levels 20-65 and >65 U/ml to be 85.7 and 57.1% compared to 100 and 93.1% for nonmucinous tumors. These results suggest that mutinous tumors possess CA 125 antigen but
32
ALTARAS
ET AL.
at levels ~65 U/ml the pickup rate in nonmucinous tumors is statistically different from mutinous tumors (P < 0.04). The overall correlation between clinical status and CA 125levels in all histological types was 87.8 and 93.5% when mutinous tumors were excluded. Thus, regardless of the histology of the tumor the correlation between CA 125 and the actual clinical status was excellent. The calculated contingency coefficient was 0.746 and this reflects excellent correlation. There was no significant difference in the initial CA 125 levels between early Stage I and II and late Stage III and IV. This suggests that the assay is not related to tumor bulk. This may be important in the early diagnosis of asymptomatic patients with ovarian carcinoma. There were 50 patients in the remission group. Six cases in this group were found to have CA 125 levels ~20 U/ml and the range was 29-259 U/ml. All 6 patients had previously had SLOs and were clinically and histologically disease free for 14-32 months. To date 4 of these patients have surgically proven recurrence, 1 patient was CT evidence of tumor, and the sixth patient is clinically disease free but has had rising CA 125 levels over the past 8 months. These elevated levels preceded clinical appearance of the disease by months and demonstrate the usefulness of CA 125 in the detection of recurrence of disease before this is clinically apparent. This will obviate the continuation of ineffective chemotherapy and suggests that an alteration in the management might be considered. All patients with levels <20 U/ml remained in complete remission during the study period. Seven patients receiving combination chemotherapy, after primary cytoreductive surgery, were eligible for SLO according to generally acceptable criteria and had CA 125 levels ~20 U/ml. The mean CA 125 level prior to SLO was 6.3 U/ml. The only patient who had residual disease at SLO had a mean CA 125 of 9.6 U/ml calculated from seven previous values. A recent report by Gershenson et al. [20] showed that only 85 out of 246 patients who underwent SLO were in fact free of disease. Our study utilizing the added parameter of CA 125 levels <20 U/ml had six out of seven negative SLOs (P < 0.05 (x2)). Despite the small group of patients in the study, the predictive value of CA 125 in assessment of the disease status prior to second-look operation appears to be encouraging. Recently, Niloff et al. [21] reported on 44 patients undegoing second- and thirdlook procedures. They did not obtain satisfactory correlation between CA 125 levels and actual disease status at operation. The lower limit of normal for CA 125 was 35 U/ml. As a result of the paucity of data relating to the predictive value of CA 125 in patients who are clinically free of disease, additional information pertaining to this subject is essential. All patients on chemotherapy had serial CA 125 levels charted against time. In 3 patients in regression, the CA 125 levels plateaued and these were the first indicator of resistance to chemotherapy while this was not clinically suspected. In these cases the CA 125 levels started rising about 2-3 months after the initial plateau. Tumor progression was noted to have occurred. We believe that a plateau at levels above the upper limit of normal is an ominous sign probably signifying resistance to therapy. We found CA 125 to be invaluable in the following situations:
OVARIAN
EPITHELIAL
CARCINOMA
33
(1) Preoperative assessment of patients with suspected ovarian tumors because of its high specificity rate at levels >65 U/ml in the absence of cardiac or hepatic dysfunction. (2) Monitoring of patients on chemotherapy because of the good correlation with clinical status. Patients with elevated levels of CA 125 probably have cancer or disease resistant to chemotherapy. (3) Negative CA 125 levels do not mean absence of tumor. (4) Timing of second-look operations because of the 85.7% correlation between CA 125 levels (20 U/ml) and negative SLO. More data are necessary; however, we believe that SLO is obviated in patients with CA 125 levels ~20 U/ml. (5) The early detection of recurrent disease in patients being followed in remission. In patients with normal levels, rising CA 125 levels have or will develop tumor. The CA 125 assay has made a profound contribution to the management of patients with carcinoma of the ovary by providing the clinician with a semiquantitative method for assessing the clinical status of the patient. In association with other tumor markers it may prove to be valuable in the diagnosis of ovarian cancer at an earlier or preclinical stage. ACKNOWLEDGMENTS We thank Mr. Sedick Isaacs for the statistical analysis, the Medical Superintendent of Groote Schuur Hospital for allowing us to publish these results, and Miss Nuja Baboo and Mrs. Audrey Mack for the preparation of the manuscript.
REFERENCES 1. Witebsky, E., Rose, N. R., and Shulman, S. Studies of normal and malignant tissue antigens, Cuncer Res. 16, 831 (1956). 2. Lloyd, S. H., and Richard, 0. H. Detection of malignant ovarian neoplasms: A review of the literature. III. Immunological detection and ovarian cancer-associated antigens, Obstet. Gynecol. Surv. 39(6), 352-355 (1984). 3. Bast, R. C., Jr., Feeney, M., Lazarus, H., Nadler, L. M., Colvin, R. B., and Knapp, R. C. Reactivity of a monoclonal antibody with human ovarian carcinoma, J. C/in. Invest. 68, 13311337 (1981). 4. Bast, R. C., Jr., Klug, T. L., St. John, E., Jenison, E., Niloff, J. M., Lazarus, H., Berkowitz, R. S., Leavitt, T., GriBiths, T., Parker, L., Zuranski, V., and Knapp, R. C. A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer, N. Engl. .Z. Med. 309, 883-887 (1983). 5. Kabawat, S. E., Bast, R. C., Jr., Welch, W. R., Knapp, R. C., and Colwin, R. B. Immunopathologic characterization of a monoclonal antibody that recognizes common surface antigens in human ovarian tumours of serous, endometriosis and clear cell type, Amer. J. C/in. Pathol. 79, 98104 (1983). 6. Niloff, J. M., Knapp, R. C., and Bast, R. C., Jr. Are CA 125 and placental alkaline phosphatase the same antigen? Amer. .Z. Obstet Gynecol. 151(3), 419-420 (1985). 7. Goldberg, G. L., Learmonth, G., Bloch, B., and Levin, W. L. The role of cul-de-sac aspiration cytology in the management and follow-up of patients with ovarian carcinoma, .I. Reprod. Med. 30, 867-870 (1985). 8. Siegal, S. Non parametric statistics for the behaviour sciences, McGraw-Hill, New York, pp. l%-202 (1956). 9. Richardson, G. S., Scully, R. E., and Nikrui, N. Common epithelial cancer of the ovary, N. Engl. J. Med. 312(8), 474-483 (1985).
34
ALTARAS ET AL.
10. Wharton, J. T., and Herson, J. Surgery for common epithelial tumours of the ovary, Cancer 48, 158-165 (1981). 11. Ehrlich, C., Einhorn, L., Stehman, F. B., and Blessing, J. Treatment of advanced epithelial ovarian cancer using cisplatin, Adriamycin and cytoxan-The Indiana University experience, Clin. Obstet. Gynecol. 10(2), 325-335 (1983). 12. Katz, M. E., Schwartz, P. E., and Kapp, E. Epithelial carcinoma of the ovary: Current strategies, Ann. Znt. Med. 95, 98-111 (1981). 13. Dembo, A. J., Sturgeon, J. F. G., and Bean, H. A. The effectiveness of adjuvant abdominopelvic irradiation in ovarian cancer, in Adjuvant therapy of cancer ZZ(S. E. Jones and S. E. Salmon, Eds.) Grune & Stratton, New York, pp. 475-482 (1979). 14. Dembo, A. J., Bush, R. S., and Brown, T. C. Clinicopathological correlates in ovarian cancer, Bull. Cancer 69, 292-297 (1982). 15. Greco, F. A., Julain, C. G., Richardson, R. L., Burnett, L., Hande, R. R., and Oldham, R. K. Advanced ovarian cancer: Brief intensive combination chemotherapy and second-look operation, Obstet. Gynecol. 138, 199-205 (1981). 16. Schwartz, P. E., and Smith, J. P. Second-look operations in ovarian cancer, Amer. .Z. Obstet. Gynecol. 138, 1124-l 130 (1981). 17. Decker, D. G. Mayo clinic experience with epithelial ovarian cancer, C/in. Obstet. Gynecol. 10(2), 337-356 (1983). 18. Ozols, R. F., Fisher, R. I., and Anderson, T. Peritoneoscopy in the management of ovarian cancer, Amer. .Z. Obstet. Gynecol. 140, 611-619 (1981). 19. Kabawat, S. E., Bast, R. C., Jr., Bhan, A. K., Welch, W. R., Knapp, R. C., and Colvin, R. B. Tissue distribution of a coelomic epithelial related antigen recognized by the monoclonal antibody OC 125, Lab. Invest. 48, 42A (1983). 20. Gershenson, D. M., Copeland, L. J., and Wharton, J. T. Prognosis of surgically determined complete responders in advanced ovarian cancer, Cancer 55, 1129-1135 (1985). 21. Niloff, J. M., Bast, R. C., Schaetzel, E. M., and Knapp, R. C. Predictive value of CA 125 antigen levels in second-look procedures for ovarian cancer, Amer. J. Obstet. Gynecol. 151, 981-986 (1985).
ONCOLOGY
30, 26-34 (1988)
The Role of Cancer Antigen 125 (CA 125) in the Management of Ovarian Epithelial Carcinomas MARKO M. ALTARAS, M.D.,*,’ GARY L. GOLDBERG, M.B.CH.B., M.R.C.O.G.,*,* WILFRED LEVIN, M.B.CH.B., F.F. RADIO (T), S.A. M.Med. Rad(T),t BASIL BLOCH, M.B.CH.B., F.R.C.O.G., M. Med. (0 & G),* LYNNE DARGE, B. Sc.,$ AND JAMES A. SMITH, M.B.CH,B., M.Med. RAD (D), M.Med.S *Division of Gynaecological Oncology, Department of Obstetrics and Gynaecology, tDivision Gynaecological Oncology, Department of Radiotherapy, and iDepartment of Nuclear Medicine, Groote Schuur Hospital and University of Cape Town, Rondebosch 7700 South Africa
of
Received November 6. 1986 From June 1, 1984, to May 31, 1985, 98 cases of epithelial ovarian carcinomas were assessed and followed prospectively using a new murine monoclonal antibody OC 125 which detects the antigen CA 125. Serous tumors comprised 43.7% of cases, mutinous tumors 20.4%, endometrioid tumors 16%, and other epithelial tumors 19.4%. Tumors of low malignant potential and benign epithelial cystadenomas were not included. For this study the upper limit of normal for CA I25 was 20 U/ml. Thirty-six were new cases. In this group the initial CA 125 levels >20 U/ml, >35 U/ml, and >65 U/ml were 97.2, 94.4, and 86. I%, respectively. When mutinous types were excluded the specificity rate did not change significantly. There was no significant difference in initial CA 125 levels between early stages I and II and late stages III and IV. No correlation between tumor bulk and the serum level of antigen was observed. The remaining 62 patients were being followed and in this group 50 were considered to be in remission. Six cases in the remission group had elevated CA 125 levels >20 U/ml and 5 of these developed clinical recurrence. The correlation between the clinical status and concordant fluctuations in the serum levels of CA 125 in all histological types was 87.8 and 93.5% when 10 cases of mutinous tumors were excluded. The contingency coefficient was 0.746. Seven SLOs were performed. All had CA 125 levels ~20 U/ml and the mean was 6.9 U/ml. Only 1 case was positive with microscopic disease. In our experience CA 125 was invaluable in the management and follow-up of patients with ovarian carcinoma especially for the early detection of recurrent disease and for the monitoring of patients on therapy. Q 1988 Academic PXSS, IX.
INTRODUCTION
In 1956 Witebsky et al. described and characterized cancer-associated antigens including ovarian cancer-associated antigens [I]. In the past 15 years attempts ’ To whom reprint requests should be addressed at Department of Obstetrics and Gynecology “A”, Gynaecological Oncology Unit, Sapir Medical Center, Meir Hospital, (SacPer Medical School, University of Tel Aviv), Kfar Saba, Israel. ’ To whom all correspondence should be addressed at Albert Einstein College of Medicine, Department of Gynecology, Division of Gynecologic Oncology, Pelham Parkway and Eastchester Road, Bronx, NY 10461. OO90-8258/88$1.50 Copyright All rights
0 1988 by Academic Press. Inc. of reproduclion in any form reserved.
26
OVARIAN
EPITHELIAL
CARCINOMA
21
TABLE 1 BREAKDOWN
OF PATIENTS
IN THE STUDY
Group 1
Group II
29 on treatment 3 lost to follow-up 4 no adjuvant therapy (surgery alone) Total 36 Total 98
12 on treatment 50 remission
62
have been made to isolate polyclonal antibodies derived from different types of epithelial carcinomas [2]. None of these have become commercially available. In 1981, Bast and co-workers developed the first monoclonal antibody OC 125 derived from an epithelial ovarian carcinoma cell line (OVCA 433) in order to detect ovarian tumors [3]. OC 125 reacts with an antigen, cancer antigen 125 (CA 125), which is common to most nonmucinous epithelial ovarian carcinomas and an assay was developed to detect CA 125 in serum [4]. This antigen is not expressed in fetal or normal adult ovarian tissue [5]. It has been demonstrated in fetal serosal surface epithelia, fallopian tube, endometrium, and cervix using the biotin-avidin peroxidase technique [6]. The same authors described serum levels of the antigen ~65 U/ml in 0.2% of apparently healthy males, 0% in healthy females, 2.1% in patients with benign disease, and 22% in patients with nongynecological cancer. In patients with histologically proven carcinoma of the ovary 74.3% had levels ~65 U/ml [4]. This antigen was detected in cryostat sections of nonmucinous epithelial ovarian tumors, but was not demonstrated in mutinous, Brenner, sex cord, and germ cell tumors of the ovary [5]. The purpose of this study was to evaluate the usefulness of CA 125 in the management and follow-up of patients with epithelial ovarian tumors of all histological types. PATIENTS, MATERIALS,
AND METHODS
In the 12-month period from June 1984, 98 consecutive cases of carcinoma of the ovary were prospectively evaluated. The patients were divided into two groups (Table 1). Group I comprised 36 new patients who had CA 125 serum levels evaluated prior to initial surgery or in the early postoperative period and Group II consisted of 62 patients seen at the follow-up clinic. The mean age of the patients was 51.2 years and the range was 23 to 79 years. The breakdown according to stage was 35 cases Stage I, 11 cases Stage II, 37 cases Stage III, and 15 cases Stage IV. The histological types included 43.7% serous, 20.4% mutinous, 16.3% endometrioid, and 19.4% other types which included clear cell (2), malignant Brenner (l), mixed type (7), and undifferentiated (9). Tumors of low malignant potential were not included in this study.
28
ALTARAS
ET AL.
During the study period the shortest follow-up was 1 month in 3 cases and the longest period of follow-up was 11 months in 5 cases. All new patients with histologically proven ovarian cancer were registered and followed up at the Combined Gynaecology Radiotherapy Assessment Clinic. Routine follow-up procedures included examination, chest radiography, ultrasound of the abdomen, CT scan, and aspiration cytology of the cul-de-sac. Routine hematological and biochemical parameters were assessed. The assay for CA 125 was introduced in our hospital in June 1984 and it became a routine investigation for patients on treatment and on routine followup. Patients suspected of having ovarian carcinoma or a pelvic mass had CA 125 evaluated preoperatively. We used the following terms to define the clinical response: Remission: where the patient had completed therapy and was clinically free of disease. Regression: where surgery reduced the tumor burden by over 50% and this clinical situation was maintained for at least 3 months by the administration of chemotherapy or in instances where chemotherapy produced marked objective and subjective improvement. This was assessed using clinical, radiological, and ultrasonic examinations. Stabilization: where the detectable tumor masses were neither progressing nor regressing and the patient’s general condition was stable. Progression: where new lesions appeared and clinical examinations related to the disease occurred. The biochemical counterparts were Remission: CA 125 <20 U/ml. Regression: ~50% decrease in serial measurements of CA 125. Stabilization: where CA 125 levels varied ~50% in serial values. Progression: >50% increase in serial measurements of CA 125. Cul-de-sac aspiration cytology was performed as described by Goldberg et al. [71. CA 125 was determined by a monoclonal antibody assay. Each sample was determined by an immunoradiomanometric assay (Centocor Inc., Malvern, PA 19355). The system is a simultaneous sandwich solid phase RIA. The patient’s plasma was stored at -4°C and defrosted prior to the assay. Polystyrene beads coated with mouse monoclonal antibody to CA 125 were incubated with the serum specimen, standards, and tracer. The tracer was ‘*‘Ilabeled monoclonal anti-CA 125 antibody. During the incubation the CA 125 formed sandwich complexes with the monoclonal antibodies. Unbound materials present in the specimen were removed by aspiration of the fluid and washing of beads. The bound radioactivity was determined by counting beads in the gammascintillation counter which was proportional to the concentration of the CA 125 in the specimen within the working range of the assay. A linear standard curve
OVARIAN
EPITHELIAL
29
CARCINOMA
TABLE 2 SPECIFICITY
All patients Nonmucinous Mutinous
RATES FOR
CA 125
No.
>20 U/ml
>35 U/ml
>65 U/ml
36 29 7
35 (97.2%) 29 (100%) 6 (85.7%)
34 (94.4%) 29 (100%) 5 (70.1%)
31 (86.1%)* 27 (93.1%)** 4 (57.1%)***
*,** P, NS. **.*** P < 0.04 (x2).
was obtained by plotting the CA 125 concentration of the standards versus bound radioactivity. The CA 125 concentration of the unknowns and the control ran concurrently with the standards and could be determined from the standard curve. The observed range for the concentration of CA 125 in serum or plasma of healthy females has been determined by the manufacturer as 10.7 + 9.6 U/ml (mean + SD) and the lower limit of sensitivity is 7 U/ml. Levels below this are determined by extrapolation. In view of these data we used 20 U/ml as the upper limit of normal. We considered a deviation of 50% in antigen level significant. Levels >500 U/ml were estimated by extrapolation although a dilutional technique was available if necessary, The interassay coefficient of variation calculated according to control provided by the manufacturer was 10.1%. A total of 367 assays were performed in this study group. In 23 patients being followed only one serum CA 125 assay was available. Most of these cases had Stage I disease and were in remission. A second-look operation (SLO) was performed in 7 cases. All patients had at least two CA 125 levels prior to SLO. Statistical methods used were chi square (x2), Fisher’s exact test, and contigency coefficient [S]. RESULTS Table 2 depicts the specificity rate of CA 125 in the 36 new patients with respect to histological type and various serum levels. Table 3 shows the correlation between the stage of the disease and the initial levels of CA 125 in the 36 new cases. Nine out of 10 patients (90%) who presented
CLINICAL
Clinical stage I II III
TABLE STAGE AND
3
CA 125 LEVELS
20 U/ml
<20-35 U/ml
36-65 U/ml
-
-
1 2
IV
1
1
Total
1
1
-
3
>65 U/ml 8 1 11
Number 9 1 13
11
13
31
36
30
ALTARAS
ET AL.
TABLE 4 HISTOLOGICAL
Histology Serous Mutinous Endometrioid Others Mixed (4) Brenner (1) Undifferentiated (3) Total %
TYPE IN RELATION
Number of cases
TO
CA 12.5LEVELS
<20 U/ml
20-35 U/ml
-
-
15 7 6 8
1
36
36-65 U/ml
1
>65 U/ml
1 1 1
14 4 5 8
31
-
-
-
1 2.8
1 2.8
3 8.3
with Stage I or II disease and 22 out of 26 (84%) of patients with Stages III and IV had CA 125 levels >65 U/ml. There was no statistically significant difference between these groups. Table 4 depicts the different levels of CA 125 at the time of diagnosis and shows its relationship to the histological type of tumor. Of the 62 patients in Group II being followed, 50 were in remission. In 44 of these patients the serum levels of CA 125 were <20 U/ml and in the 6 remaining patients the serum levels were >20 U/ml. The 44 patients with levels ~20 U/ml have remained clinically free of disease during the study period. Their serum levels ranged from 0 to 15.8 U/ml in 110 assays. The mean was 5.69 k 7.39 U/ml. Table 5 summarizes the clinical status of the disease in relation to fluctuations in the CA 125 levels. This analysis comprised 29 new patients and 12 patients from the follow-up group who were on treatment during the study period (see Table 1). The variance in CA 125 levels correlated with the clinical status in 36 of the 41 patients (87.8%). When the 10 patients with mutinous tumors were excluded the correlation was correct in 29 of 31 patients (93.5%). In 7 out of 10 patients with mutinous tumors there was a positive correlation between clinical status of the disease and the antigen levels. The contingency coefficient was obtained among the regression, stabilization, and progression groups using three cutoff levels of the antigen. The result was found to be 0.746 which is almost
CORRELATION
CA 125 >50% decrease <50% variation >50% increase Total
OF
TABLE 5 CA 125 LEVELS AND
CLINICAL
STATUS
Regression
Stabilization
Progression
21 1 0 22
0 4 0 4
0 4 11 15
OVARIAN
EPITHELIAL
CARCINOMA
31
perfect concordance. There were 5 inconsistent results. In 2 cases there was clinical progression of the disease which was not reflected with serial CA 125 assays. One of these patients had a malignant Brenner tumor and the other had an endometrioid carcinoma. Although the levels of CA 125were high, the increments were not >50% and did not reflect clinical progression of disease. In 2 cases there was clinical progression of disease but the CA 125 levels did not increase significantly. Both these patients had mutinous tumors. In 1 case the CA 125 level plateaued for 2 months and the patient was found to have bone metastases. We performed seven SLOs during this time. All were clinically assessed to be free of disease and had negative chest radiology, CT scan, and cul-de-sac aspiration cytology and preoperative CA 125 assays ~20 U/ml. One laparotomy was positive revealing microscopic disease only. DISCUSSION Over the past 5 to 10 years there has been an improvement in the survival of patients with epithelial ovarian carcinomas [9]. This is probably due to aggressive cytoreductive surgery, new combination chemotherapy protocols, and adjuvant radiotherapy [lo-141. However, no significant further improvement in survival can be expected without the availability of a reliable, noninvasive means for monitoring the clinical course of the patient. The major problems facing the clinician are the optimal duration of chemotherapy, the timing and need for second-look operation, and the reliable assessment of patients in remission [15-181. Another problem is the early detection of recurrent disease, especially tumor nodules less than 2 cm in diameter, and disease in the upper abdomen [18]. These recurrences are only diagnosed once the tumors become clinically evident. The immunoradiometric assay for CA 125 was originally described for the detection of nonmucinous epithelial ovarian cancers [4]. In two studies the presence of CA 125 could not be demonstrated in cryostat sections of mutinous and Brenner tumors [5,19]. We were thus surprised to find markedly elevated CA 125 levels in mutinous tumors at the start of this investigation. We continued to monitor CA 125 in all suspected cases of ovarian tumors and in all patients being followed up irrespective of the histological type. We determined the specificity rate of CA 125 in 36 new cases of ovarian epithelial cancer of all histological types using 20 U/ml as the upper limit of normal (see Table 3). The specificity rate was found to be 97.2% at levels ~20 U/ml, 94.4% at levels >35 U/ml, and 86.1% for values ~65 U/ml when all histological types were evaluated. When mutinous tumors were excluded the specificity rate was 93.1% for levels >65 U/ml and 100% for levels ~35 U/ml. The high specificity rate in our study is similar to that reported by Bast et al. and predicts its usefulness in clinical practice. It has allowed us to explore its use as an indicator of the response to treatment. Our results show the specificity rate for mutinous tumors at levels 20-65 and >65 U/ml to be 85.7 and 57.1% compared to 100 and 93.1% for nonmucinous tumors. These results suggest that mutinous tumors possess CA 125 antigen but
32
ALTARAS
ET AL.
at levels ~65 U/ml the pickup rate in nonmucinous tumors is statistically different from mutinous tumors (P < 0.04). The overall correlation between clinical status and CA 125levels in all histological types was 87.8 and 93.5% when mutinous tumors were excluded. Thus, regardless of the histology of the tumor the correlation between CA 125 and the actual clinical status was excellent. The calculated contingency coefficient was 0.746 and this reflects excellent correlation. There was no significant difference in the initial CA 125 levels between early Stage I and II and late Stage III and IV. This suggests that the assay is not related to tumor bulk. This may be important in the early diagnosis of asymptomatic patients with ovarian carcinoma. There were 50 patients in the remission group. Six cases in this group were found to have CA 125 levels ~20 U/ml and the range was 29-259 U/ml. All 6 patients had previously had SLOs and were clinically and histologically disease free for 14-32 months. To date 4 of these patients have surgically proven recurrence, 1 patient was CT evidence of tumor, and the sixth patient is clinically disease free but has had rising CA 125 levels over the past 8 months. These elevated levels preceded clinical appearance of the disease by months and demonstrate the usefulness of CA 125 in the detection of recurrence of disease before this is clinically apparent. This will obviate the continuation of ineffective chemotherapy and suggests that an alteration in the management might be considered. All patients with levels <20 U/ml remained in complete remission during the study period. Seven patients receiving combination chemotherapy, after primary cytoreductive surgery, were eligible for SLO according to generally acceptable criteria and had CA 125 levels ~20 U/ml. The mean CA 125 level prior to SLO was 6.3 U/ml. The only patient who had residual disease at SLO had a mean CA 125 of 9.6 U/ml calculated from seven previous values. A recent report by Gershenson et al. [20] showed that only 85 out of 246 patients who underwent SLO were in fact free of disease. Our study utilizing the added parameter of CA 125 levels <20 U/ml had six out of seven negative SLOs (P < 0.05 (x2)). Despite the small group of patients in the study, the predictive value of CA 125 in assessment of the disease status prior to second-look operation appears to be encouraging. Recently, Niloff et al. [21] reported on 44 patients undegoing second- and thirdlook procedures. They did not obtain satisfactory correlation between CA 125 levels and actual disease status at operation. The lower limit of normal for CA 125 was 35 U/ml. As a result of the paucity of data relating to the predictive value of CA 125 in patients who are clinically free of disease, additional information pertaining to this subject is essential. All patients on chemotherapy had serial CA 125 levels charted against time. In 3 patients in regression, the CA 125 levels plateaued and these were the first indicator of resistance to chemotherapy while this was not clinically suspected. In these cases the CA 125 levels started rising about 2-3 months after the initial plateau. Tumor progression was noted to have occurred. We believe that a plateau at levels above the upper limit of normal is an ominous sign probably signifying resistance to therapy. We found CA 125 to be invaluable in the following situations:
OVARIAN
EPITHELIAL
CARCINOMA
33
(1) Preoperative assessment of patients with suspected ovarian tumors because of its high specificity rate at levels >65 U/ml in the absence of cardiac or hepatic dysfunction. (2) Monitoring of patients on chemotherapy because of the good correlation with clinical status. Patients with elevated levels of CA 125 probably have cancer or disease resistant to chemotherapy. (3) Negative CA 125 levels do not mean absence of tumor. (4) Timing of second-look operations because of the 85.7% correlation between CA 125 levels (20 U/ml) and negative SLO. More data are necessary; however, we believe that SLO is obviated in patients with CA 125 levels ~20 U/ml. (5) The early detection of recurrent disease in patients being followed in remission. In patients with normal levels, rising CA 125 levels have or will develop tumor. The CA 125 assay has made a profound contribution to the management of patients with carcinoma of the ovary by providing the clinician with a semiquantitative method for assessing the clinical status of the patient. In association with other tumor markers it may prove to be valuable in the diagnosis of ovarian cancer at an earlier or preclinical stage. ACKNOWLEDGMENTS We thank Mr. Sedick Isaacs for the statistical analysis, the Medical Superintendent of Groote Schuur Hospital for allowing us to publish these results, and Miss Nuja Baboo and Mrs. Audrey Mack for the preparation of the manuscript.
REFERENCES 1. Witebsky, E., Rose, N. R., and Shulman, S. Studies of normal and malignant tissue antigens, Cuncer Res. 16, 831 (1956). 2. Lloyd, S. H., and Richard, 0. H. Detection of malignant ovarian neoplasms: A review of the literature. III. Immunological detection and ovarian cancer-associated antigens, Obstet. Gynecol. Surv. 39(6), 352-355 (1984). 3. Bast, R. C., Jr., Feeney, M., Lazarus, H., Nadler, L. M., Colvin, R. B., and Knapp, R. C. Reactivity of a monoclonal antibody with human ovarian carcinoma, J. C/in. Invest. 68, 13311337 (1981). 4. Bast, R. C., Jr., Klug, T. L., St. John, E., Jenison, E., Niloff, J. M., Lazarus, H., Berkowitz, R. S., Leavitt, T., GriBiths, T., Parker, L., Zuranski, V., and Knapp, R. C. A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer, N. Engl. .Z. Med. 309, 883-887 (1983). 5. Kabawat, S. E., Bast, R. C., Jr., Welch, W. R., Knapp, R. C., and Colwin, R. B. Immunopathologic characterization of a monoclonal antibody that recognizes common surface antigens in human ovarian tumours of serous, endometriosis and clear cell type, Amer. J. C/in. Pathol. 79, 98104 (1983). 6. Niloff, J. M., Knapp, R. C., and Bast, R. C., Jr. Are CA 125 and placental alkaline phosphatase the same antigen? Amer. .Z. Obstet Gynecol. 151(3), 419-420 (1985). 7. Goldberg, G. L., Learmonth, G., Bloch, B., and Levin, W. L. The role of cul-de-sac aspiration cytology in the management and follow-up of patients with ovarian carcinoma, .I. Reprod. Med. 30, 867-870 (1985). 8. Siegal, S. Non parametric statistics for the behaviour sciences, McGraw-Hill, New York, pp. l%-202 (1956). 9. Richardson, G. S., Scully, R. E., and Nikrui, N. Common epithelial cancer of the ovary, N. Engl. J. Med. 312(8), 474-483 (1985).
34
ALTARAS ET AL.
10. Wharton, J. T., and Herson, J. Surgery for common epithelial tumours of the ovary, Cancer 48, 158-165 (1981). 11. Ehrlich, C., Einhorn, L., Stehman, F. B., and Blessing, J. Treatment of advanced epithelial ovarian cancer using cisplatin, Adriamycin and cytoxan-The Indiana University experience, Clin. Obstet. Gynecol. 10(2), 325-335 (1983). 12. Katz, M. E., Schwartz, P. E., and Kapp, E. Epithelial carcinoma of the ovary: Current strategies, Ann. Znt. Med. 95, 98-111 (1981). 13. Dembo, A. J., Sturgeon, J. F. G., and Bean, H. A. The effectiveness of adjuvant abdominopelvic irradiation in ovarian cancer, in Adjuvant therapy of cancer ZZ(S. E. Jones and S. E. Salmon, Eds.) Grune & Stratton, New York, pp. 475-482 (1979). 14. Dembo, A. J., Bush, R. S., and Brown, T. C. Clinicopathological correlates in ovarian cancer, Bull. Cancer 69, 292-297 (1982). 15. Greco, F. A., Julain, C. G., Richardson, R. L., Burnett, L., Hande, R. R., and Oldham, R. K. Advanced ovarian cancer: Brief intensive combination chemotherapy and second-look operation, Obstet. Gynecol. 138, 199-205 (1981). 16. Schwartz, P. E., and Smith, J. P. Second-look operations in ovarian cancer, Amer. .Z. Obstet. Gynecol. 138, 1124-l 130 (1981). 17. Decker, D. G. Mayo clinic experience with epithelial ovarian cancer, C/in. Obstet. Gynecol. 10(2), 337-356 (1983). 18. Ozols, R. F., Fisher, R. I., and Anderson, T. Peritoneoscopy in the management of ovarian cancer, Amer. .Z. Obstet. Gynecol. 140, 611-619 (1981). 19. Kabawat, S. E., Bast, R. C., Jr., Bhan, A. K., Welch, W. R., Knapp, R. C., and Colvin, R. B. Tissue distribution of a coelomic epithelial related antigen recognized by the monoclonal antibody OC 125, Lab. Invest. 48, 42A (1983). 20. Gershenson, D. M., Copeland, L. J., and Wharton, J. T. Prognosis of surgically determined complete responders in advanced ovarian cancer, Cancer 55, 1129-1135 (1985). 21. Niloff, J. M., Bast, R. C., Schaetzel, E. M., and Knapp, R. C. Predictive value of CA 125 antigen levels in second-look procedures for ovarian cancer, Amer. J. Obstet. Gynecol. 151, 981-986 (1985).