- Email: [email protected]
Use of CA 125 to monitor patients with ovarian epithelial carcinomas
GYNECOLOGIC
ONCOLOGY
33, 193-197 (1989)
Use of CA 125 to Monitor Patients with Ovarian Epithelial Carcinomas’ EDWARD PODCZASKI, M.D., CHARLES WHITNEY, M.D., ALBERTO MANETTA, M.D., JAMES E. LARSON, M.D., KIRK, M.D. ,* CLARK W. STEVENS, M.D., JUDITH LYTER, R.N., AND RODRIGUE MORTEL, M.D. Department
of Obstetrics
and Gynecology,
M. S. Hershey
Medical
Center,
Pennsylvania
State
University,
Hershey,
Pennsylvania
JANET 17033
Received August 31, 1987 The CA 125 radioimmunoassay
hasbeenincreasinglyusedto monitor the courseof patientswith ovarian epithelialcarcinomas. The purposeof this report is to describeour experiencein the useof this assayand to better defineits clinical utility. Fifty-one patients had serum CA 125 follow-up during primary chemotherapy. All 51 patients demonstratedeither a normal CA 125 level at the completionof chemotherapyor a substantialfall in CA 125valueswith treatment. In 48 of 51 patients, the drop in CA 125 levelswas temporally related to the clinical regression or remissionof tumor. Forty of thesepatientsunderwentsecondlook laparotomy; 23 patients(58%) had residualdisease.A total of 45 patients had serumCA 125 determinationsat the time of second-looklaparotomy. Eight patients with microscopicdisease and 11of 18 patientswith grossresidualdiseasehad a “negative” (lessthan 35 U/ml) CA I25 level. The predictive value of an elevatedCA 125 level was 1.00. However, the predictive value of a negative value was only 0.50. Hence, a negative CA 125 level cannot be a substitutefor a second-looklaparotomy. Only 7 of 18 patients (39%)with grossresidualdiseaseat second-look surgery had an elevatedCA 125 level. Patientswith an elevated CA 125and grossresidualtumor at the second-looklaparotomy uniformly demonstratedlarge, bulky disease.Furthermore, the survival of patients with grossresidual diseaseat second-look laparotomycorrelatedwith the preoperativeCA 125value. Serum CA 125 determinationsalso show promisein the follow-up of patientswith a negativesecond-looklaparotomy. The serumCA 125level from patientswith a “negative” second-looklaparotomy can becomeelevatedmonthsbeforerecurrentdisease isappreciated. 0 1989 Academic
F’ress, Inc.
INTRODUCTION
The identification of serum factors to monitor the course of ovarian epithelial carcinomas has been a long-standing pursuit. Using hybridoma technology, Bast et al. de’ Supported in part by BRSG Grant SO7 RRO5680-18 awarded by the Biomedical Research Support Grant Program, Division of Resources, National Institutes of Health. 2 Present address: William Beaumont Hospital, Royal Oak, MI.
veloped a murine, monoclonal antibody that reacted with an ovarian carcinoma tumor-associated antigen, CA 125 [ 11. This antibody was subsequently used to develop a commercially available radioimmunoassay that measured CA 125 in serum [2]. By this assay, only 1 of 888 “healthy” patients and 6% of 143 patients with nonmalignant disease had serum CA 125 levels in excess of 35 U/ml. However, 83 of 101 patients (82%) with surgically demonstrable ovarian carcinomas had elevated levels of the antigen. Rising or falling CA 125 levels also correlated with progression or regression of disease as assessed by clinical criteria. In the past 3 years, there has been increasing clinical experience in the use of CA 125 to monitor patients with ovarian epithelial carcinomas. Serum CA 125 levels have been correlated with initial operative findings [3], response to primary chemotherapy [4], and disease status at secondlook laparotomy [5]. The purpose of this report is to describe our clinical experience in the use of this assay. An effort was made to relate the serum CA 125 levels to clinical landmarks in the natural history of this malignancy. Furthermore, CA 125 levels were analyzed in terms of tumor- and treatment-related variables. MATERIALS
AND METHODS
A total of 119 patients with ovarian epithelial carcinomas were retrospectively studied in terms of serum CA 125 levels and the clinical course of their disease. Five hundred sixty-nine serum determinations were performed during the study. Patients underwent from 1 to 26 CA 125 determinations during the period of follow-up. Operative staging was performed in accordance with FIG0 guidelines [61. Tumor histology was in accordance with the WHO classification [7]. Blood samples were obtained by venipuncture. The serum was separated by coagulation and later low-speed centrifugation at room temperature. Serum specimens
193 0090-8258/89$1 SO Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.
194
PODCZASKI
were frozen and stored at -20°C. At a later date, the specimens were thawed and the serum CA 125 was measured using a batch assay. Each specimen was run in duplicate along with appropriate controls. The serum CA 125 levels were determined by a solid-phase, sandwich radioimmunoassay using a murine, monoclonal antibody. The assay was originally developed by Centocor (Malvern, PA), and later commercially distributed by Amersham. Specimens with a CA 125 in excess of 500 U/ml were diluted and reassayed. The serum CA 125 levels were analyzed in terms of the natural history of ovarian epithelial carcinomas. CA 125 levels of 35 U/ml or less were considered normal; values greater than 35 U/ml were considered as elevated or “positive” [2]. The preoperative CA 125 levels were related to the findings at initial staging/debulking surgery and tumor characteristics. The effect of chemotherapy on subsequent CA 125 levels was also examined. Rising or falling levels were analyzed according to the clinical course of the ovarian malignancy. As previously described by Bast and co-workers [2], a doubling or halving of CA 125 levels was considered as a significant increase (“rise”) or decrease as compared to prior values. Fifty-one patients had CA 125 determinations during primary chemotherapy. Five patients had stage I disease, six had stage II disease, and forty had stage III or IV disease. Three patients with stage I disease and three with stage II disease received melphalan after initial surgery. The remaining patients were treated with combination chemotherapy using a cisplatin-cyclophosphamide regimen. A total of 45 patients had CA 125 determinations at the time of second-look laparotomy. All patients who underwent the procedure had no clinical evidence of tumor prior to surgery. The serum CA 125 level was correlated with the operative findings and the subsequent clinical course of the patient. The sensitivity, specificity, and accuracy of the assay were calculated for the prediction of disease status at second-look laparotomy. Seven additional patients underwent reexploration as a “thirdlook laparotomy” or to rule out persistent disease after treatment of recurrent tumor. All seven patients were in disease remission and had no evidence of tumor by clinical examination. Life tables for patient groups were calculated by KaplanMeier actuarial estimates [8]. Survival functions were compared using the log rank method [9]. Means for patient groups were compared using the analysis of variance. RESULTS Twenty-eight patients had preoperative CA 125 levels before diagnostic laparotomy that demonstrated an ovarian epithelial malignancy. The relationship between preop-
ET AL.
erative CA 125 levels, tumor stage, and tumor histology is illustrated in Fig. 1. In general, patients with advanced disease had marked elevations in CA 125 levels. The majority of patients with stage III or IV disease and tumors of serous histology had CA 125 levels in excess of 1000 U/ml. Patients with tumors of mutinous, endometrioid, or clear cell histology had lower levels of CA 125 than did patients with serous malignancies of similar stage. After staging/debulking surgery, 51 patients demonstrated either a normal CA 125 level at the completion of chemotherapy or a substantial fall in CA 125 values with chemotherapy. In 48 of the 51 patients, the drop in CA 125 levels was temporally related to the clinical regression or remission of tumor. Three patients developed clinically obvious persistence of tumor despite initially falling or low levels of serum CA 125 (cl00 U/ml). Forty of these patients have already undergone second-look laparotomy. Although all 40 patients were in disease remission, second-look laparotomy demonstrated residual disease in 23 of the 40 patients (58%). Sixteen patients undergoing serial CA 125 follow-up demonstrated rising values. A “rise” in CA 125 levels was regarded as at least a doubling of prior values. Fourteen of the sixteen patients (88%) demonstrated clinical or radiologic progression of disease which corresponded temporally to the rising CA 125 values. One patient has demonstrated rising values after a “negative” second-
CA 125
/I
U
histology FIG. 1. Relationship between tumor stage, histology, and CA 125 level at the diagnosis of ovarian epithehal carcinomas. Each cell in the graph contains the CA 125 level from a single patient at initial diagnosis of an ovarian epithelial carcinoma.
CA 125 AND OVARIAN
look laparotomy; recurrent disease has not yet been clinically documented. Preoperative CA 125 levels were available for 45 patients who underwent second-look laparotomy. The preoperative CA 125 determinations and outcomes of second-look laparotomy are summarized in Table 1. Although the specificity was 1.00, the calculated sensitivity of the CA 125 assay for predicting disease status at second-look laparotomy was 0.27. The predictive value of a positive test was 1.00. However, the predictive value of a negative test (~35 U/ml) was only 0.50. The overall accuracy of the assay to anticipate disease status at second-look was 0.58. The data in Table 1 were stratified according to the amount of disease present at second-look laparotomy (Table 2). Patients without gross evidence of tumor at exploration but with residual tumor on pathologic examination of systematic, random biopsies or positive peritoneal cytologies were classified as having “microscopic disease.” Nineteen patients had a negative second-look laparotomy. Preoperative CA 125 levels for this group of patients ranged from 0 to 16 (mean of 7.4) U/ml. Two of the nineteen patients later demonstrated recurrent disease alter a “negative” second-look laparotomy. Patients with a “negative” second-look and later clinical evidence of recurrent tumor had preoperative CA 125 levels that were indistinguishable from those of patients who did not develop recurrence. Eight patients had microscopic disease at second-look surgery (Fig. 2). All patients had advanced (stage III or IV) disease and had tumors of serous histology. The CA 125 levels ranged from 4.7 to 19.6 (mean of 10.2) U/ml. CA 125 levels for this group were not statistically different from those obtained for patients with a negative secondlook. Eighteen patients had gross disease at second-look laparotomy. All patients had stage III or IV disease. Fifteen patients had tumors of serous histology. Preoperative CA 125 levels ranged from 6.0 to 5000 U/ml and formed a bimodal distribution. Eleven patients with gross residual disease had preoperative CA 125 levels that were less than 35 U/ml. Six of the eleven patients with negative CA 125 levels prior to second-look laparotomy have subsequently demonstrated clearly elevated CA 125 deter-
EPITHELIAL
195
CANCERS
TABLE 2 Statusof Diseaseat Second-LookLaparotomy and PreoperativeCA 125 Values in 45 Patients Serum CA 125
No disease
Microscopic disease
Gross disease
Positive (>35 U/ml) Negative (635 U/ml)
0 19
0 8
7 11
Total
19
8
18
minations. Only 7 of the 18 patients (39 percent) with gross residual disease had preoperative CA 125 values which exceeded 35 U/ml. All seven patients had bulky residual disease with tumor nodules that exceeded 3 cm in diameter. The survival of patients with gross residual disease at second-look laparotomy can be correlated with the CA 125 level prior to surgery (P < 0.01). Patients with CA 125 levels in excess of 35 U/ml have an adverse prognosis as compared to patients with normal preoperative values
4800
.
1 1
.
3000
3100
1300
300
.
100
.
2 E 1.5
50
TABLE 1 Status of Diseaseat Second-LookSurgery and Preoperative CA 125 Values in 45 Patients Serum CA 125
Residual disease
Negative
7 19
0 19
DlSfRSf
Positive (>35 U/ml) Negative (~35 U/ml)
STAWS
RT
SfCOND-LOOK
1APAROTOflW
FIG. 2. Level of CA 125 and disease status at second-look laparotomy. The solid line marks 35 U/ml.
196
PODCZASE
(Fig. 3). None of the patients with an elevated CA 125 level has survived in excess of 18 months. In contrast, 4 of the 11 patients (36%) with normal CA 125 levels have already survived beyond 20 months. Twenty-seven patients with a negative second-look laparotomy were subsequently followed with serial CA 125 levels and clinical examinations. Five patients with a “negative” second-look subsequently developed recurrent disease from 5 to 47 months after laparotomy. One patient had an elevated CA 125 value at the time that recurrence was documented. Three patients had normal CA 125 values at the diagnosis of recurrent disease. All three patients had localized, small recurrences and had no evidence of ascites. The CA 125 levels subsequently became elevated in the follow-up of these three patients. The fifth patient had demonstrated elevated CA 125 levels 3 months before recurrent tumor was clinically appreciated. An additional patient has had elevated CA 125 levels after a “negative” second-look laparotomy; however, examinations have not yet demonstrated recurrence. Seven additional patients underwent reexploration as a “third-look laparotomy” or to rule out disease persistence after treatment of recurrent tumor. All seven patients were clinically in disease remission and had no evidence of disease by examination. Six patients had CA 125 levels below 35 U/ml. Four patients had no tumor found at reexploration; two patients had gross residual disease. One patient had a preoperative CA 125 of 57 U/ml; gross residual disease was found at surgery. 1.00
0.80
; 0.60 E 2 g s b 0.40
b
I
b9
-
CA125
> 35Wml
rnONTHS
FIG. 3. Life tables for patients with gross residual disease at secondlook laparatomy as stratified to preoperative CA 125 levels.
;I ET AL.
DISCUSSION The calculated, predictive value of a negative or positive CA 125 value to anticipate disease status at second-look laparotomy is in agreement with the results of Berek and colleagues [5]. All patients with an elevated preoperative CA 125 value had residual disease at second- or thirdlook laparotomy. Furthermore, rising CA 125 values also appear to have considerable clinical significance. Fourteen of sixteen patients with rising levels demonstrated tumor progression that corresponded temporally to the increasing CA 125 values. An excellent correlation was also demonstrated between a falling CA 125 level and patient response to primary chemotherapy as assessed by clinical examination. It has been reported that rising or falling levels of CA 125 correlate with progression or regression of disease in 42 of 45 (93%) instances [2]. However, the strength of this association diminishes when evaluated by more rigorous criteria, such as laparotomy. Forty-five patients in clinical disease remission underwent second-look laparotomy. Although 38 of 45 patients (84%) had a negative CA 125 value prior to surgery, only 19 patients (42%) had a negative second-look laparotomy. Another 2 of 7 patients with a “third-look laparotomy” also had residual tumor, despite negative CA 125 levels. Although the predictive value of an elevated CA 125 level is excellent, a normal value (~35 U/ml) has limited significance. Hence, a negative CA 125 level is no substitute for a second-look laparotomy. The production of a tumor-associated antigen is affected by both tumor heterogeneity and tumor volume. The data demonstrate a correlation between the tumor burden and the resulting CA 125 level. Patients with microscopic or minimal gross residual disease at second-look laparotomy had CA 125 levels that were uniformly less than 35 U/ml. Ail seven patients with gross residual disease and elevated CA 125 levels had bulky tumor implants at laparotomy. Six of the eleven patients with gross residual disease and a normal CA 125 level have gone on to demonstrate elevated CA 125 levels. Hence, the low CA 125 levels prior to second-look laparotomy cannot be attributed to loss of cell clones producing the antigen. The data of Berek and colleagues also suggest a relationship between CA 125 prior to second-look laparotomy and the surgical extent of disease [5]. Patients with microscopic disease had a median CA 125 of 8.21 U/ml. Patients with residual disease from 5 to 14 mm in size had a median CA 125 of 11.3 U/ml. Furthermore, patients with tumor implants greater than or equal to 15 mm in size had a median CA 125 of 79.7 U/ml. Lavin and co-workers have emphasized the CA 125 level 3 months after completion of chemotherapy as a critical predictor of response to therapy [41. However,
CA 125 AND OVARIAN
the CA 125 level at second-look laparotomy may have more significance than previously appreciated. Patients with gross residual tumor at second-look laparotomy and a normal preoperative CA 125 have a statistically significant survival advantage as compared to those with levels in excess of 35 U/ml. This survival advantage may be related to a lesser tumor burden in those patients with a negative CA 125, but with macroscopic residual disease at second-look laparotomy. The survival advantage demonstrated in patients with gross residual disease, yet normal CA 125 levels at secondlook laparotomy, is consistent with the observations of others. Niloff and co-workers have evaluated the time interval to clinical tumor recurrence as a function of CA 125 level at the time of second-look surgery [lo]. An elevated CA 125 level at second-look laparotomy (>35 U/ml) was associated with a 60% chance of clinical recurrence as compared to 5% in patients with normal CA 125 levels. The CA 125 assay also appears to have clinical usefulness in the follow-up of patients after a negative secondlook laparotomy. An elevated CA 125 value can be the first indication of recurrent tumor after a “negative” second-look laparotomy. However, the present data also demonstrate that patients with small, localized recurrences can have a normal serum CA 125. REFERENCES 1. 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. Clin. Invest. 68, 1331-1337 (1981).
EPITHELIAL
CANCERS
197
2. Bast, R. C., Jr., Klug, T. L., St. John, E., Jenison, E., Niloff, J. M., Lazarus, H., Berkowitz, R. S., Leavitt, T., Griffiths, C. T., Parker, L., Zurawski, V. R., and Knapp, R. C. A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer, N. Engl. J. Med. 309, 883-887 (1983). 3. Einhorn, N., Bast, R. C., Jr., Knapp, R. C., Tjernberg, B., and Zurawski, V. R. Preoperative evaluation of serum CA 125 levels in patients with primary epithelial ovarian cancer, Obsrer. Gynecol. 67, 414-416 (1986). 4.
Lavin, P. T., Knapp, R. C., Malkasian, G., Whitney, C. W., Berek, J. C., and Bast, R. C., Jr. CA 125 for the monitoring of ovarian carcinoma during primary therapy, Obstet. Gynecol. 69, 223-227 (1987). Berek, J. S., Knapp, R. C., Malkasian, G. D., Lavin, P. T., Whitney, C., Niloff, J. M., and Bast, R. C., Jr. CA 125 serum levels correlated with second-look operations among ovarian cancer patients, Obstet. Gynecol. 67, 685-689 (1986). American Joint Committee for Cancer Staging and End Results Reporting. Manualfor suing cancer. Lippincott, Chicago (1978). Serov, S. F., Scully, R. E., and Sobin, L. H. International histological classification of tumors, in Histological typing of ovnrian turnours, No. 9. World Health Organization, Geneva (1973). Kaplan, E. L., and Meier, P. Nonparametric complete observations, J. Amer. Star. Assoc.
estimation from in53, 457-481 (1958).
Peto, R., Pike, M. C., Armitage, P., Breslow, N. E., Cox, D. R., Howard, S. V., Mantel, N., McPherson, K., Peto, J., and Smith, P. G. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples, hit. J. Cancer 35, l-39 (1977). 10. Niloff, J. M., Knapp, R. C., Lavin, P. T., Malkasian, G. D., Berek, J. S., Mortel, R., Whitney, C., Zurawski, V. R., and Bast, R. C., Jr. The CA 125 assay as a predictor of clinical recurrence in epithelial ovarian cancer, Amer. J. Obsret. Gynecol. 155, 55-60 (1986).
ONCOLOGY
33, 193-197 (1989)
Use of CA 125 to Monitor Patients with Ovarian Epithelial Carcinomas’ EDWARD PODCZASKI, M.D., CHARLES WHITNEY, M.D., ALBERTO MANETTA, M.D., JAMES E. LARSON, M.D., KIRK, M.D. ,* CLARK W. STEVENS, M.D., JUDITH LYTER, R.N., AND RODRIGUE MORTEL, M.D. Department
of Obstetrics
and Gynecology,
M. S. Hershey
Medical
Center,
Pennsylvania
State
University,
Hershey,
Pennsylvania
JANET 17033
Received August 31, 1987 The CA 125 radioimmunoassay
hasbeenincreasinglyusedto monitor the courseof patientswith ovarian epithelialcarcinomas. The purposeof this report is to describeour experiencein the useof this assayand to better defineits clinical utility. Fifty-one patients had serum CA 125 follow-up during primary chemotherapy. All 51 patients demonstratedeither a normal CA 125 level at the completionof chemotherapyor a substantialfall in CA 125valueswith treatment. In 48 of 51 patients, the drop in CA 125 levelswas temporally related to the clinical regression or remissionof tumor. Forty of thesepatientsunderwentsecondlook laparotomy; 23 patients(58%) had residualdisease.A total of 45 patients had serumCA 125 determinationsat the time of second-looklaparotomy. Eight patients with microscopicdisease and 11of 18 patientswith grossresidualdiseasehad a “negative” (lessthan 35 U/ml) CA I25 level. The predictive value of an elevatedCA 125 level was 1.00. However, the predictive value of a negative value was only 0.50. Hence, a negative CA 125 level cannot be a substitutefor a second-looklaparotomy. Only 7 of 18 patients (39%)with grossresidualdiseaseat second-look surgery had an elevatedCA 125 level. Patientswith an elevated CA 125and grossresidualtumor at the second-looklaparotomy uniformly demonstratedlarge, bulky disease.Furthermore, the survival of patients with grossresidual diseaseat second-look laparotomycorrelatedwith the preoperativeCA 125value. Serum CA 125 determinationsalso show promisein the follow-up of patientswith a negativesecond-looklaparotomy. The serumCA 125level from patientswith a “negative” second-looklaparotomy can becomeelevatedmonthsbeforerecurrentdisease isappreciated. 0 1989 Academic
F’ress, Inc.
INTRODUCTION
The identification of serum factors to monitor the course of ovarian epithelial carcinomas has been a long-standing pursuit. Using hybridoma technology, Bast et al. de’ Supported in part by BRSG Grant SO7 RRO5680-18 awarded by the Biomedical Research Support Grant Program, Division of Resources, National Institutes of Health. 2 Present address: William Beaumont Hospital, Royal Oak, MI.
veloped a murine, monoclonal antibody that reacted with an ovarian carcinoma tumor-associated antigen, CA 125 [ 11. This antibody was subsequently used to develop a commercially available radioimmunoassay that measured CA 125 in serum [2]. By this assay, only 1 of 888 “healthy” patients and 6% of 143 patients with nonmalignant disease had serum CA 125 levels in excess of 35 U/ml. However, 83 of 101 patients (82%) with surgically demonstrable ovarian carcinomas had elevated levels of the antigen. Rising or falling CA 125 levels also correlated with progression or regression of disease as assessed by clinical criteria. In the past 3 years, there has been increasing clinical experience in the use of CA 125 to monitor patients with ovarian epithelial carcinomas. Serum CA 125 levels have been correlated with initial operative findings [3], response to primary chemotherapy [4], and disease status at secondlook laparotomy [5]. The purpose of this report is to describe our clinical experience in the use of this assay. An effort was made to relate the serum CA 125 levels to clinical landmarks in the natural history of this malignancy. Furthermore, CA 125 levels were analyzed in terms of tumor- and treatment-related variables. MATERIALS
AND METHODS
A total of 119 patients with ovarian epithelial carcinomas were retrospectively studied in terms of serum CA 125 levels and the clinical course of their disease. Five hundred sixty-nine serum determinations were performed during the study. Patients underwent from 1 to 26 CA 125 determinations during the period of follow-up. Operative staging was performed in accordance with FIG0 guidelines [61. Tumor histology was in accordance with the WHO classification [7]. Blood samples were obtained by venipuncture. The serum was separated by coagulation and later low-speed centrifugation at room temperature. Serum specimens
193 0090-8258/89$1 SO Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.
194
PODCZASKI
were frozen and stored at -20°C. At a later date, the specimens were thawed and the serum CA 125 was measured using a batch assay. Each specimen was run in duplicate along with appropriate controls. The serum CA 125 levels were determined by a solid-phase, sandwich radioimmunoassay using a murine, monoclonal antibody. The assay was originally developed by Centocor (Malvern, PA), and later commercially distributed by Amersham. Specimens with a CA 125 in excess of 500 U/ml were diluted and reassayed. The serum CA 125 levels were analyzed in terms of the natural history of ovarian epithelial carcinomas. CA 125 levels of 35 U/ml or less were considered normal; values greater than 35 U/ml were considered as elevated or “positive” [2]. The preoperative CA 125 levels were related to the findings at initial staging/debulking surgery and tumor characteristics. The effect of chemotherapy on subsequent CA 125 levels was also examined. Rising or falling levels were analyzed according to the clinical course of the ovarian malignancy. As previously described by Bast and co-workers [2], a doubling or halving of CA 125 levels was considered as a significant increase (“rise”) or decrease as compared to prior values. Fifty-one patients had CA 125 determinations during primary chemotherapy. Five patients had stage I disease, six had stage II disease, and forty had stage III or IV disease. Three patients with stage I disease and three with stage II disease received melphalan after initial surgery. The remaining patients were treated with combination chemotherapy using a cisplatin-cyclophosphamide regimen. A total of 45 patients had CA 125 determinations at the time of second-look laparotomy. All patients who underwent the procedure had no clinical evidence of tumor prior to surgery. The serum CA 125 level was correlated with the operative findings and the subsequent clinical course of the patient. The sensitivity, specificity, and accuracy of the assay were calculated for the prediction of disease status at second-look laparotomy. Seven additional patients underwent reexploration as a “thirdlook laparotomy” or to rule out persistent disease after treatment of recurrent tumor. All seven patients were in disease remission and had no evidence of tumor by clinical examination. Life tables for patient groups were calculated by KaplanMeier actuarial estimates [8]. Survival functions were compared using the log rank method [9]. Means for patient groups were compared using the analysis of variance. RESULTS Twenty-eight patients had preoperative CA 125 levels before diagnostic laparotomy that demonstrated an ovarian epithelial malignancy. The relationship between preop-
ET AL.
erative CA 125 levels, tumor stage, and tumor histology is illustrated in Fig. 1. In general, patients with advanced disease had marked elevations in CA 125 levels. The majority of patients with stage III or IV disease and tumors of serous histology had CA 125 levels in excess of 1000 U/ml. Patients with tumors of mutinous, endometrioid, or clear cell histology had lower levels of CA 125 than did patients with serous malignancies of similar stage. After staging/debulking surgery, 51 patients demonstrated either a normal CA 125 level at the completion of chemotherapy or a substantial fall in CA 125 values with chemotherapy. In 48 of the 51 patients, the drop in CA 125 levels was temporally related to the clinical regression or remission of tumor. Three patients developed clinically obvious persistence of tumor despite initially falling or low levels of serum CA 125 (cl00 U/ml). Forty of these patients have already undergone second-look laparotomy. Although all 40 patients were in disease remission, second-look laparotomy demonstrated residual disease in 23 of the 40 patients (58%). Sixteen patients undergoing serial CA 125 follow-up demonstrated rising values. A “rise” in CA 125 levels was regarded as at least a doubling of prior values. Fourteen of the sixteen patients (88%) demonstrated clinical or radiologic progression of disease which corresponded temporally to the rising CA 125 values. One patient has demonstrated rising values after a “negative” second-
CA 125
/I
U
histology FIG. 1. Relationship between tumor stage, histology, and CA 125 level at the diagnosis of ovarian epithehal carcinomas. Each cell in the graph contains the CA 125 level from a single patient at initial diagnosis of an ovarian epithelial carcinoma.
CA 125 AND OVARIAN
look laparotomy; recurrent disease has not yet been clinically documented. Preoperative CA 125 levels were available for 45 patients who underwent second-look laparotomy. The preoperative CA 125 determinations and outcomes of second-look laparotomy are summarized in Table 1. Although the specificity was 1.00, the calculated sensitivity of the CA 125 assay for predicting disease status at second-look laparotomy was 0.27. The predictive value of a positive test was 1.00. However, the predictive value of a negative test (~35 U/ml) was only 0.50. The overall accuracy of the assay to anticipate disease status at second-look was 0.58. The data in Table 1 were stratified according to the amount of disease present at second-look laparotomy (Table 2). Patients without gross evidence of tumor at exploration but with residual tumor on pathologic examination of systematic, random biopsies or positive peritoneal cytologies were classified as having “microscopic disease.” Nineteen patients had a negative second-look laparotomy. Preoperative CA 125 levels for this group of patients ranged from 0 to 16 (mean of 7.4) U/ml. Two of the nineteen patients later demonstrated recurrent disease alter a “negative” second-look laparotomy. Patients with a “negative” second-look and later clinical evidence of recurrent tumor had preoperative CA 125 levels that were indistinguishable from those of patients who did not develop recurrence. Eight patients had microscopic disease at second-look surgery (Fig. 2). All patients had advanced (stage III or IV) disease and had tumors of serous histology. The CA 125 levels ranged from 4.7 to 19.6 (mean of 10.2) U/ml. CA 125 levels for this group were not statistically different from those obtained for patients with a negative secondlook. Eighteen patients had gross disease at second-look laparotomy. All patients had stage III or IV disease. Fifteen patients had tumors of serous histology. Preoperative CA 125 levels ranged from 6.0 to 5000 U/ml and formed a bimodal distribution. Eleven patients with gross residual disease had preoperative CA 125 levels that were less than 35 U/ml. Six of the eleven patients with negative CA 125 levels prior to second-look laparotomy have subsequently demonstrated clearly elevated CA 125 deter-
EPITHELIAL
195
CANCERS
TABLE 2 Statusof Diseaseat Second-LookLaparotomy and PreoperativeCA 125 Values in 45 Patients Serum CA 125
No disease
Microscopic disease
Gross disease
Positive (>35 U/ml) Negative (635 U/ml)
0 19
0 8
7 11
Total
19
8
18
minations. Only 7 of the 18 patients (39 percent) with gross residual disease had preoperative CA 125 values which exceeded 35 U/ml. All seven patients had bulky residual disease with tumor nodules that exceeded 3 cm in diameter. The survival of patients with gross residual disease at second-look laparotomy can be correlated with the CA 125 level prior to surgery (P < 0.01). Patients with CA 125 levels in excess of 35 U/ml have an adverse prognosis as compared to patients with normal preoperative values
4800
.
1 1
.
3000
3100
1300
300
.
100
.
2 E 1.5
50
TABLE 1 Status of Diseaseat Second-LookSurgery and Preoperative CA 125 Values in 45 Patients Serum CA 125
Residual disease
Negative
7 19
0 19
DlSfRSf
Positive (>35 U/ml) Negative (~35 U/ml)
STAWS
RT
SfCOND-LOOK
1APAROTOflW
FIG. 2. Level of CA 125 and disease status at second-look laparotomy. The solid line marks 35 U/ml.
196
PODCZASE
(Fig. 3). None of the patients with an elevated CA 125 level has survived in excess of 18 months. In contrast, 4 of the 11 patients (36%) with normal CA 125 levels have already survived beyond 20 months. Twenty-seven patients with a negative second-look laparotomy were subsequently followed with serial CA 125 levels and clinical examinations. Five patients with a “negative” second-look subsequently developed recurrent disease from 5 to 47 months after laparotomy. One patient had an elevated CA 125 value at the time that recurrence was documented. Three patients had normal CA 125 values at the diagnosis of recurrent disease. All three patients had localized, small recurrences and had no evidence of ascites. The CA 125 levels subsequently became elevated in the follow-up of these three patients. The fifth patient had demonstrated elevated CA 125 levels 3 months before recurrent tumor was clinically appreciated. An additional patient has had elevated CA 125 levels after a “negative” second-look laparotomy; however, examinations have not yet demonstrated recurrence. Seven additional patients underwent reexploration as a “third-look laparotomy” or to rule out disease persistence after treatment of recurrent tumor. All seven patients were clinically in disease remission and had no evidence of disease by examination. Six patients had CA 125 levels below 35 U/ml. Four patients had no tumor found at reexploration; two patients had gross residual disease. One patient had a preoperative CA 125 of 57 U/ml; gross residual disease was found at surgery. 1.00
0.80
; 0.60 E 2 g s b 0.40
b
I
b9
-
CA125
> 35Wml
rnONTHS
FIG. 3. Life tables for patients with gross residual disease at secondlook laparatomy as stratified to preoperative CA 125 levels.
;I ET AL.
DISCUSSION The calculated, predictive value of a negative or positive CA 125 value to anticipate disease status at second-look laparotomy is in agreement with the results of Berek and colleagues [5]. All patients with an elevated preoperative CA 125 value had residual disease at second- or thirdlook laparotomy. Furthermore, rising CA 125 values also appear to have considerable clinical significance. Fourteen of sixteen patients with rising levels demonstrated tumor progression that corresponded temporally to the increasing CA 125 values. An excellent correlation was also demonstrated between a falling CA 125 level and patient response to primary chemotherapy as assessed by clinical examination. It has been reported that rising or falling levels of CA 125 correlate with progression or regression of disease in 42 of 45 (93%) instances [2]. However, the strength of this association diminishes when evaluated by more rigorous criteria, such as laparotomy. Forty-five patients in clinical disease remission underwent second-look laparotomy. Although 38 of 45 patients (84%) had a negative CA 125 value prior to surgery, only 19 patients (42%) had a negative second-look laparotomy. Another 2 of 7 patients with a “third-look laparotomy” also had residual tumor, despite negative CA 125 levels. Although the predictive value of an elevated CA 125 level is excellent, a normal value (~35 U/ml) has limited significance. Hence, a negative CA 125 level is no substitute for a second-look laparotomy. The production of a tumor-associated antigen is affected by both tumor heterogeneity and tumor volume. The data demonstrate a correlation between the tumor burden and the resulting CA 125 level. Patients with microscopic or minimal gross residual disease at second-look laparotomy had CA 125 levels that were uniformly less than 35 U/ml. Ail seven patients with gross residual disease and elevated CA 125 levels had bulky tumor implants at laparotomy. Six of the eleven patients with gross residual disease and a normal CA 125 level have gone on to demonstrate elevated CA 125 levels. Hence, the low CA 125 levels prior to second-look laparotomy cannot be attributed to loss of cell clones producing the antigen. The data of Berek and colleagues also suggest a relationship between CA 125 prior to second-look laparotomy and the surgical extent of disease [5]. Patients with microscopic disease had a median CA 125 of 8.21 U/ml. Patients with residual disease from 5 to 14 mm in size had a median CA 125 of 11.3 U/ml. Furthermore, patients with tumor implants greater than or equal to 15 mm in size had a median CA 125 of 79.7 U/ml. Lavin and co-workers have emphasized the CA 125 level 3 months after completion of chemotherapy as a critical predictor of response to therapy [41. However,
CA 125 AND OVARIAN
the CA 125 level at second-look laparotomy may have more significance than previously appreciated. Patients with gross residual tumor at second-look laparotomy and a normal preoperative CA 125 have a statistically significant survival advantage as compared to those with levels in excess of 35 U/ml. This survival advantage may be related to a lesser tumor burden in those patients with a negative CA 125, but with macroscopic residual disease at second-look laparotomy. The survival advantage demonstrated in patients with gross residual disease, yet normal CA 125 levels at secondlook laparotomy, is consistent with the observations of others. Niloff and co-workers have evaluated the time interval to clinical tumor recurrence as a function of CA 125 level at the time of second-look surgery [lo]. An elevated CA 125 level at second-look laparotomy (>35 U/ml) was associated with a 60% chance of clinical recurrence as compared to 5% in patients with normal CA 125 levels. The CA 125 assay also appears to have clinical usefulness in the follow-up of patients after a negative secondlook laparotomy. An elevated CA 125 value can be the first indication of recurrent tumor after a “negative” second-look laparotomy. However, the present data also demonstrate that patients with small, localized recurrences can have a normal serum CA 125. REFERENCES 1. 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. Clin. Invest. 68, 1331-1337 (1981).
EPITHELIAL
CANCERS
197
2. Bast, R. C., Jr., Klug, T. L., St. John, E., Jenison, E., Niloff, J. M., Lazarus, H., Berkowitz, R. S., Leavitt, T., Griffiths, C. T., Parker, L., Zurawski, V. R., and Knapp, R. C. A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer, N. Engl. J. Med. 309, 883-887 (1983). 3. Einhorn, N., Bast, R. C., Jr., Knapp, R. C., Tjernberg, B., and Zurawski, V. R. Preoperative evaluation of serum CA 125 levels in patients with primary epithelial ovarian cancer, Obsrer. Gynecol. 67, 414-416 (1986). 4.
Lavin, P. T., Knapp, R. C., Malkasian, G., Whitney, C. W., Berek, J. C., and Bast, R. C., Jr. CA 125 for the monitoring of ovarian carcinoma during primary therapy, Obstet. Gynecol. 69, 223-227 (1987). Berek, J. S., Knapp, R. C., Malkasian, G. D., Lavin, P. T., Whitney, C., Niloff, J. M., and Bast, R. C., Jr. CA 125 serum levels correlated with second-look operations among ovarian cancer patients, Obstet. Gynecol. 67, 685-689 (1986). American Joint Committee for Cancer Staging and End Results Reporting. Manualfor suing cancer. Lippincott, Chicago (1978). Serov, S. F., Scully, R. E., and Sobin, L. H. International histological classification of tumors, in Histological typing of ovnrian turnours, No. 9. World Health Organization, Geneva (1973). Kaplan, E. L., and Meier, P. Nonparametric complete observations, J. Amer. Star. Assoc.
estimation from in53, 457-481 (1958).
Peto, R., Pike, M. C., Armitage, P., Breslow, N. E., Cox, D. R., Howard, S. V., Mantel, N., McPherson, K., Peto, J., and Smith, P. G. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples, hit. J. Cancer 35, l-39 (1977). 10. Niloff, J. M., Knapp, R. C., Lavin, P. T., Malkasian, G. D., Berek, J. S., Mortel, R., Whitney, C., Zurawski, V. R., and Bast, R. C., Jr. The CA 125 assay as a predictor of clinical recurrence in epithelial ovarian cancer, Amer. J. Obsret. Gynecol. 155, 55-60 (1986).