CA 125: fundamental and clinical aspects

seminars in C A N C E R B I OLOG Y, Vol 9, 1999: pp. 117]124 Article No. scbi.1998.0114, available online at http:rrwww.idealibrary.com on

CA 125: fundamental and clinical aspects R.H.M. VerheijenU , S. von Mensdorff-PouillyU , G.J. van Kamp† and P. KenemansU Since the OC 125 monoclonal antibody (Mab) was generated, other Mabs to the CA 125 glycoprotein have been produced and classified into two families associated with two major epitope regions on the CA 125 molecule. New generation assays, combining Mabs to two distinct regions of the molecule, compare favorably with that of the original assays as demonstrated by ROC curves. The original CA 125 assay suffered from interference of HAMA, an important drawback considering the increasing use of murine antibodies for immunodiagnosis and treatment of ovarian cancer. This problem has been solved for the majority of currently available tests. The sensitivity of the assays for early ovarian cancer remains low, precluding its indiscriminate use for screening and diagnosis of ovarian cancer. Its use in screening for early cancer, combined with ultrasonography, is limited to high risk populations, such as women from families with mutations in the BRCA1 or 2 gene. Although CA 125 assessment may play a limited role in the (early) detection of ovarian cancer, its role in the follow-up during and after therapy is well established. The major contribution of CA 125 is in the monitoring of tumor response to chemotherapy, where it is valuable in detecting those patients with an inadequate response to the chosen treatment. The role of CA 125 in early detection of recurrences remains to be established and is currently the subject of two large clinical trials.

Ideally, these tumor markers should be selective and reliable indicators of the presence of tumors. These criteria for tumor markers are rarely met, human choriogonadotropin being one of the rare exceptions. The introduction by Kohler and Milstein1 of a ¨ technique to produce monoclonal antibodies made it possible to generate antibodies directed to specific epitopes on an antigen. Markers like CA 125 are well established in clinical practice, despite their limitations of selectivity and specificity. Bast et al 2 raised a monoclonal antibody, OC 125, against a tumor-associated glycoprotein, denoted CA 125. CA 125 now belongs to the most successful markers and has been applied in diagnosis, follow-up as well as treatment of ovarian cancer.

CA 125: structure and function The structure of the CA 125 antigen has not yet been elucidated and its biological function is largely unknown. The gene that codes for the peptide portion of the glycosylated molecule has not yet been cloned and the knowledge of its glycosylated structure is very limited. It has been suggested that the locus for CA 125 is closely linked to the 17p locus of the BRCA1 gene, mutations of which predispose for breast and ovarian cancer. Interestingly, polyclonal antibodies recognizing CA 125 have been instrumental in identifying a new NBR1-gene, situated}as the name suggests Ž Next to BRCA1 gene. }in the same genomic region as the BRCA1 gene.3 It has even been suggested that the regulation of various genes in this region may be coordinated including the one that may be coding for CA 125.4 The CA 125 antigen is derived from the coelomic epithelium and was primarily defined by the OC 125 monoclonal antibody. Bast et al 2 raised the OC 125 monoclonal antibody by immunizing BALBrc mice with the OVCA 433 cell line derived from ascitic fluid from a patient with serous papillary cystadenocarcinoma of the ovary. Immunohistochemically the anti-

Key words: CA 125 r ovarian cancer r tumor markers Q1999 Academic Press

Introduction TUMOR-SPECIFIC ANTIGENS could potentially be used to monitor tumor behavior and response to therapy. U

From the Department of Obstetrics and Gynaecology and †Department of Clinical Chemistry, Academic Hospital Vrije Universiteit, Amsterdam, The Netherlands Q1999 Academic Press 1044-579Xr 99r 020117q 08 $30.00r 0

117

R. H. M. Verheijen et al

Assays to detect CA 125 in body fluids

body reacts with various structures arising from the coelomic epithelium, such as pleura, pericardium, peritoneum and Mullerian epithelia. Specifically ar¨ eas with inflammation are positive.5,6 Normal ovarian epithelium does not express the antigen, whereas serous papillary tumors and to a lesser extent mucinous papillary tumors exhibit the antigen more or less abundantly. Phosphorylation of the O-glycosylated CA 125 glycoprotein seems to be an essential step for secretion of the antigen into the circulation. Phosphorylation is effected by epidermal growth factor ŽEGF. through EGF receptors and via serine]thymine]kinase. As soon as the more than 200-kDa antigen is shed autoproteases initiate dephosphorylation which results in approximately 40-kDa fragments that contain the CA 125 and M11 epitopes.7 The OC 125 antibody recognizes the protein portion with a variable carbohydrate content; whether or not these carbohydrates are involved in the recognition by OC 125 is still unclear.8

The original CA 125 assay was a homologous radioimmunoassay, employing the murine OC 125 monoclonal antibody both as catcher and as 125 I-labelled detector antibody.2 Repetition of the antibodydefined epitopes on the same molecule is mandatory for binding and detection. Recently, a second generation of assays has emerged, incorporating the M11 antibody as catcher.9 Since the OC 125 antibody was generated other CA 125 antigen-detecting antibodies have been identified and classified into two families, associated with at least two distinct major epitope regions Žsee Figure 1.. A concerted action launched by the International Society of Oncodevelopmental Biology and Medicine ŽISOBM. TD-1 Workshop investigated 26 monoclonal antibodies against the CA 125 antigen and could classify these as binding to antigenic domains.10 These regions bind OC 125 and M11 antibodies, respectively, as well as their equivalents. A third epitope has

Figure 1. Distribution of epitopes on the CA 125 antigen. The two main epitope regions are characterized by the binding of OC 125 Ž-like . and M11 Ž-like . antibodies respectively. For each region the tested antibodies are listed.

118

CA 125: fundamental and clinical aspects

been found to be detected by only one single antibody, OV 197. As a result of these findings, a new generation of CA 125 assays has been developed, combining an M11-like antibody with the original OC 125 or an antibody from the same group, with various non-isotope labels. These new labels allow application of faster analytical techniques.11 Assays combining monoclonal antibodies recognizing epitopes on two distinct regions are potentially more sensitive since the catcher and detector antibody do not need to compete for the same epitope. This may mean an improvement in cost-effectiveness, due to increased precision and reproducibility.12 The test performance of these new, second generation assays compares favorably with that of the original assays, as was demonstrated by receiver operating characteristics ŽROC. curves that describe the way in which sensitivity and specificity of an assay change with respect to each other as the cut-off criterion is varied. Specifically, the new generation of tests provides better discrimination between early ovarian cancer and benign pelvic tumors Žsee Figure 2, data partly published..13,14 Assessment of the various new assays also led us to conclude that optimal cut-off values calculated on the basis of ROC curves ranged from 30 to 60 Urml,15 instead of the 35 Urml cut-off introduced by Bast et al.2 As expected all tests show a relatively low specificity and sensitivity in the detection of early ovarian cancer, although specificity has been improved by the introduction of the new heterologous double-determinant assays.14 Since tumor marker tests use mouse monoclonal antibodies they may be influenced by the presence of human anti-mouse antibodies ŽHAMA. in the serum of patients. Such anti-mouse immunoglobulin response may occur after contact with rodents or after injection with mouse monoclonal antibodies for scintigraphy or therapy.16 HAMA will interconnect catcher and tracer antibody, thus giving a spurious rise in CA 125 levels. Such interference can be prevented by, e.g. the addition of mouse serum to the assay. The original murine CA 125 test also suffered from interference of HAMA, as did some of the new generation assays Žsee Figure 3, own unpublished data.. This problem has been solved for the majority of currently available tests,17 which is important in view of new diagnostic and therapeutic modalities employing murine monoclonal antibodies to target tumor in the patient through intravenous or intracavitary injection. Otherwise such treatment would preclude follow-up of the patient using a CA 125 assay.

Figure 2. Actual ROC curves for CA 125 and CA 125 II assays: especially for the discrimination of early stage ŽI]II. ovarian tumors versus benign tumors the second generation assays perform better.

CA 125 serum levels in physiologic and benign conditions CA 125 is synthesized within the cell and accumulates at the luminal cell side, where it is actively secreted into the lumen. In benign tumors excessive antigen shedding into the cystic fluid may occur. It has been postulated that in invasive tumors tissue architecture may be disrupted sufficiently to allow antigen release into the bloodstream. The same may occur in benign conditions that cause architectural changes.18 In general, conditions that cause peritoneal or mesothelial changes can give rise to elevation of CA 125 serum levels.19 As a matter of fact, peritoneal mesothelial cells in culture shed five times more CA 125 than ovarian cancer cells.20 It is therefore not surprising that around the time of the menstrual period slightly elevated CA 125 levels are found within an otherwise normal range.21 During the first trimester of pregnancy and post-partum, CA 125 levels are also above the established cut-off value of 35 Urml.22,23 Apart from these physiological rises, benign conditions such as intra-abdominal bleeding, indeed any operation, infection, and in particular uterine fibroids and pelvic endometriosis may cause 119

R. H. M. Verheijen et al

Figure 3. Interference of HAMA at various concentrations added to blanks tested with various second generation assays. Some assays still give false high CA 125 levels, specifically in the presence of an abundance of HAMA ŽCentocor II, BM Enzymun., other tests are not influenced by the presence of HAMA ŽACS OV, Abbott..

CA 125 levels to elevate well above 35 Urml. CA 125 has been successfully used to detect and monitor endometriosis.24 Both menopausal status and hysterectomy influence CA 125 levels significantly. After menopause andror hysterectomy CA 125 levels are significantly lowered.25

Cystic fluid CA 125 levels behave in a similar way to the extent that the majority of both benign and malignant cysts have a high CA 125 content,28 although malignant cystic fluid contains higher levels than benign cystic fluid.29 There are indications that both cystic walls and the peritoneum constitute barriers for the high molecular weight glycoprotein released into cavities, which hence have a relatively high antigen content.28 Thus the concentration ratio between benign cystic fluid and serum has been reported to be 60:1, whereas this ratio was 3:1 for malignant cysts and corresponding serum ŽDietrich, personal communication.. These ratios are a reflection of the amount of CA 125 shed into the serum of patients with malignant tumors and to a much lesser extent of the cystic fluid content itself. Assessment of cystic fluid alone, obtained, e.g. at trans vaginal ultrasound guided puncture, cannot be used to differentiate between benign and malignant ovarian tumor. In vitro experiments have revived interest in the role of not only the inward lining of ovarian cysts but also of the ovarian surface epithelium in the shedding of CA 125. In their search for morphological changes among cell cultures from ovarian surface epithelium, Auersperg et al 30 looked at the in vitro production of CA 125. It appeared that in compar-

Peritoneal and cystic fluid Since CA 125 is shed from the mesothelium one might expect peritoneal fluid to exhibit relatively high levels of CA 125. Normal CA 125 levels in the peritoneal fluid have been calculated by Buller et al 26 to be 152 Urml and by Vergote et al 27 to be 250 Urml. Surprisingly, when assessing peritoneal fluid or washings from patients undergoing a second look, operation levels were not significantly different between those who had tumor and those who did not.26 In a study of 70 ovarian carcinoma patients, Vergote et al 27 also found consistently higher peritoneal CA 125 values than serum values. They established an inverse correlation between peritoneal CA 125 level and prognosis. Clearly, peritoneal fluid cannot be used as an indicator of tumor, but it may be used as a prognosticator. 120

CA 125: fundamental and clinical aspects

Differential diagnosis

ison with ovaries from women without a family history of ovarian cancer, ovaries from women with a familial predisposition showed a higher proportion of cells expressing CA 125, and also that these cells retained this capacity throughout multiple passages. Thus CA 125 expression might be one of the premalignant changes that play a role in carcinogenesis.

Establishing the exact nature of a pelvic mass found at pelvic or ultrasound examination may be extremely difficult as neither method is sufficiently discriminative. Despite the limitations of CA 125 to distinguish between benign and malignant cysts, it is still one of the best serum tests available.38 Many authors have advocated the use of a combination of markers, especially CA 125 and MUC1-assays, to increase specificity.39 ] 42 Some of the proposed new additional markers, e.g. OVX1, are not commercially available and in many cases the marginal gain obtained by using a panel of markers does not outweigh the disadvantages and costs involved. Rather, the combination of CA 125 and ultrasound assessment is a reliable method to assess the benign or malignant nature of a cyst. The use of a logistic modelling technique including pelvic examination, ultrasound, and CA 125 levels gave a reliability of 81% in predicting malignancy,43 which result could only marginally be improved to 87% by the addition of the CA 72-4 assay.44 Specific markers may aid in distinguishing ovarian malignancy from a malignant tumor originating elsewhere. Examples are CEA for gastro-intestinal cancer, CA 19.9 for pancreatic carcinoma, and CA 15.3 for metastases of breast cancer. Yedema et al 45 have shown that if the ratio of CA 125rCEA is less than 20:1, the tumor is most likely of gastro-intestinal origin. Pretreatment assessment in this way is of great importance for the right choice of treatment as well as for prognosis.

Early detection of ovarian cancer More than half of patients with early stage ovarian cancer do not exhibit elevated CA 125 levels.31,32 Conversely, CA 125 values may be increased in some benign and healthy conditions leading to an unacceptably high number of false positive tests.33 Borderline tumors of the ovary behave like malignant tumors in respect to tumor markers and also show a CA 125 increase above normal serum levels in only 40% of early stage ŽFIGO stage 1. tumors but in almost 100% of later stage tumors.34 The second generation CA 125 assays have a better performance of the test especially in the lower stage tumors as indicated by receiver operating characteristics ŽROC. curves when we compared various tests against the original CA 125 test Žsee Figure 2; data partly published..41 Another attempt to improve the performance of CA 125 assays in screening has been the introduction of an algorithm to calculate the risk of ovarian cancer. Using a second generation assay, Skates et al 35 combined the slope and intercept of serial CA 125 measurements and could thus identify patients at risk for early recurrence. Applied in a screening program this algorithm could potentially carry the sensitivity and specificity of this test to acceptable levels.36 The use of complementary markers has not improved the rather poor performance of CA 125 in screening for ovarian cancer.37 Several large studies in asymptomatic women have successfully incorporated ultrasonography as a primary screen next to or after CA 125 testing Žfor an overview see Jacobs36 .. The relatively low sensitivity for early disease indicates that CA 125 is not suitable for the detection of early cancer in an asymptomatic apparently healthy low risk population where the incidence of ovarian cancer is low. However, it can be expected to be effective in high risk populations such as women from families with hereditary breast ovarian carcinoma ŽHBOC. syndrome, caused by mutations in the BRCA1 or 2 genes.36

Follow-up during and after treatment Although CA 125 assessment may play a limited role Table 1. Survival according to CA 125 response after three courses of chemotherapy Early prediction of outcome

CA 125 outcome

Median survival time

After three courses

F 35 Urml ) 35 Urml

27 months 14 months Ž P s 0.003.

t2

F 20 days ) 20 days

28 months 19 months Ž P s 0.003.

Response is given as absolute levels using the standard cut-off level of 35 Urml and as half-life Ž t 2 . with a cut-off of 20 days.50

121

R. H. M. Verheijen et al

in the Žearly. detection of ovarian cancer, its role in the follow-up during and after therapy is well established.46 Soon after the introduction of the CA 125 test it became apparent that both the decline as well as the absolute value of CA 125 assessed after the first courses of chemotherapy were predictors of the final clinical response at the end of chemotherapy.47 CA 125 assessment should not start within the first 3 weeks after surgery, as the surgery itself causes levels to elevate.48 It has been shown by several groups using several chemotherapeutic regimens that inadequate regression of CA 125 after three cycles of chemotherapy is an indication that the patient will not respond to therapy. Buller et al 49 calculated an ideal CA 125 regression rate at 7.6 days, which incidently compares well with the biological half-life of 4.5 days. They even suggested stopping conventional chemotherapeutic treatment and to switch to phase IrII drugs if the regression rates were inadequate. We have shown that either the absolute level of CA 125 or its half-life after three courses of chemotherapy is a strong indicator of prognosis Žsee Table 1..50 Recently, more or less complicated and computerised calculations to predict tumor response on the basis of CA 125 values early in treatment have been published.51,52 A word of caution should perhaps be expressed for the CA 125 response to treatment with paclitaxel, which is now a component of standard first line chemotherapy together with cisplatin.53 In a study relatin g C A 125 levels to resp o n se to cisplatinrpaclitaxel no clear correlation could be found although CA 125 changes did actually correlate with survival.54 This lack of correlation between CA 125 regression and clinical response has been confirmed by others,55 whereas in vitro studies show a good correlation between CA 125 levels in the culture medium and the number of cells surviving paclitaxel treatment.56 But in vitro testing may not be valid in this respect since it has been shown that CA 125 excreting tumor cell lines react with an increased excretion both to EGF and to taxanes.20,57 Clinically, the discrepancy between CA 125 levels and response to paclitaxel treatment may also be explained by fluid retention.58 This example illustrates the various mechanisms that may influence our interpretation of variation in test results. In ovarian cancer, the ultimate benefit of follow-up after clinical remission remains a matter of debate. Specifically the use of CA 125, which will often detect preclinical recurrence, is questioned in view of the lack of effective second-line treatment. The progno-

sis of the 80% of patients that have an elevated CA 125 value at the time of relapse is significantly worse than that of the 20% without CA 125 elevation.59 The question whether early treatment at the time of CA 125 rise will result in a better prognosis than treatment at the time of clinically detectable and evaluable disease is currently the subject of two large trials initiated by the UK Medical Research Council ŽMRC. and the European Organization for Research and Treatment of Cancer ŽEORTC., respectively.

Conclusions Despite the fact that the exact structure of CA 125 has not yet been elucidated and despite the limited sensitivity, and}to a lesser extent}specificity of the available assays this antigen is well established as a tumor marker for ovarian cancer. Its use in screening for early cancer is limited to high risk populations, where it should be combined with ultrasonographic assessment. The test may aid in differential diagnosis, but its major contribution is in the monitoring of tumor response to chemotherapy. Here it is valuable in detecting those patients that do not adequately respond to the chosen chemotherapy and who can therefore discontinue ineffective but otherwise harmful treatment. Its value in the early detection of recurrence has still to be established.

References 1. Kohler G, Milstein C Ž1975. Continuous culture of fused cells ¨ secreting antibody of predefined specificity. Nature 256:495]496 2. Bast RC, Klug TL, St John E, Jenison E, Niloff JM, Lazarus H, Berkowitz RS, Leavitt T, Griffiths CT, Parker L, Zurawski VR, Knapp RC Ž1983. A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. N Engl J Med 309:169]171 3. Chambers JA, Solomon E Ž1996. Isolation of the murine NBR1 gene adjacent to the murine BRCA1 gene. Genomics 38:305]313 4. Xu C-F, Brown MA, Nicolai H, Chambers JA, Griffiths BL, Solomon E Ž1997. Isolation and characterisation of the NBR2 gene which lies head to head with the human BRCA1 gene. Human Mol Genet 6:1057]1062 5. Kabawat SE, Bast RC, Bhan AK, Welch WR, Knapp RC, Clovin RB Ž1983. Tissue distribution of a coelomic epithelium related antigen recognized by the monoclonal antibody OC 125. Int J Gynecol Pathol 2:275]285 6. Kabawat SE, Bast RC, Welch WR, Knapp RC, Colvin RB Ž1983. Immunopathologic characterization of a monoclonal antibody that recognizes common surface antigens of human ovarian tumors of serous, endometrioid, and clear cell types. An J Clin Pathol 79:89]104 7. Fendrick JL, Konishi I, Geary SM, Parmley TH, Quirk JG,

122

CA 125: fundamental and clinical aspects

8.

9.

10.

11.

12. 13. 14. 15.

16.

17. 18. 19.

20.

21. 22. 23.

24.

O’Brien TJ Ž1997. CA 125 phosphorylation is associated with its secretion from the wish human amnion cell line. Tumor Biol 18:278]289 Lloyd KO, Yin BWT, Kudryashov V Ž1997. Isolation and characterization of ovarian cancer antigen CA 125 using a new monoclonal antibody Žvk-8 .: identification as a mucintype molecule. Int J Cancer 71:842]850 O’Brien TJ, Raymond LM, Bannon GA, Ford DH, Hardardottir H, Miller FC, Quirk JG Jr Ž1991. New monoclonal antibodies identify the glycoprotein carrying the CA 125 epitope. Am J Obstet Gynecol 165:1857]1864 Nustad K, Bast RC, O’Brien TJ, Nilsson O, Seguin, Suresh MR, Saga T, Nozawa S, Bormer OP, Bruijn HWA de, Nap M, Vitali A, Gadnell M, Clark J, Shigemasa K, Karlsson B, Kreutz FT, Jette D, Sakahara H, Endo K, Paus E, Warren D, Hammarstrom ¨ S, Kenemans P, Hilgers J Ž1996. Specificity and affinity of 26 monoclonal antibodies against the CA 125 antigen: first report from the ISOBM TD-1 Workshop. Tumor Biol 17:196]219 Kenemans P, van Kamp GJ, Oehr P, Verstraeten RA Ž1993. Heterologous double-determinant immunoradiometric assay CA 125 II: reliable second-generation immunoassay for determining CA 125 in serum. Clin Chem 39:2509]2513 Bonfrer ` JMG, Baan AW, Jansen E, Lentfer D, Kenemans P Ž1994. Technical evaluation of three second generation CA 125 assays. Eur J Clin Biochem 32:201]207 Kenemans P, Verstraeten AA, van Kamp GJ, von Mensdorff-Pouilly S Ž1995. The second generation CA 125 assays. Ann Med 27:107]113 Davelaar EM, van Kamp GJ, Verstraeten RA, Kenemans P Ž1998. Comparison of seven immunoassays for the quantification of CA 125 antigen in serum. Clin Chem 44:1417]1422 Bon GG, Kenemans P, Verstraeten AA, van Kamp GJ, Hilgers J Ž1996. Serum tumor marker immunoassays in gynecologic oncology: establishment of reference values. Am J Obstet Gynecol 174:107]114 Muto MG, Lepisto EM, van den Abbeele AD, Knapp RC, Kassis AI Ž1989. Influence of human antimurine antibody on CA 125 levels in patients with ovarian cancer undergoing radioimmunotherapy or immunoscintigraphy with murine monoclonal antibody OC 125. Am J Obstet Gynecol 161:1206]1212 Roth HJ, Zahn I Ž1997. ACS OV: an alternative to well established tests also measuring CA 125? Clin Lab 43:515]526 Kenemans P, Yedema CA, Bon GG, von Mensdorff-Pouilly S Ž1993. CA 125 in gynaecologic pathology}a review. Eur J Obstet Gynecol Repr Biol 49:115]124 Zeimet AG, Offner FA, Mullerholzner E, Widschwendter M, Abendstein B, Fuith LC, Daxenbichler G, Marth C Ž1998. Peritoneum and tissues of the female reproductive tract as physiological sources of CA-125. Tumor Biol 19:275]282 Zeimet AG, Offner FA, Marth C, Heim K, Feichtinger H, Daxenbichler G, Dapunt O Ž1997. Modulation of CA-125 release by inflammatory cytokines in human peritoneal mesothelial and ovarian cancer cells. Anticancer Res 17:3129]3131 Grover SM, Koh H, Weideman P, Quinn MA Ž1992. The effect of the menstrual cycle on serum CA 125 levels: a population study. Am J Obstet Gynecol 167:1379]1381 Seki K, Kikuchi Y, Uesato T, Kato K Ž1986. Increased serum CA 125 levels during the first trimester of pregnancy. Acta Obstet Gynecol Scand 65:583]585 Kobayashi F, Sagawa N, Nakamura K, Nonogaki M, Ban C, Fujii S, Mori T Ž1989. Mechanism and clinical significance of elevated CA 125 levels in the sera of pregnant women. Am J Obstet Gynecol 160:563]566 Hompes PGA, Koninckx PR, Kennedy S, van Kamp GJ, Verstraeten AA, Cornillie F Ž1996. Serum CA 125 concentrations during midfollicular phase, a clinically useful and repro-

25. 26. 27. 28.

29. 30.

31.

32.

33.

34. 35.

36. 37. 38. 39.

40.

41.

42. 43.

123

ducible marker in diagnosis of advanced endometriosis. Clin Chem 42:1871]1874 Grover S, Quinn MA, Weideman P, Koh H Ž1992. Factors influencing serum CA 125 levels in normal women. Obstet Gynecol 79:511]514 Buller RE, Manetta A, Bloss J, DiSaia PJ, Berman ML Ž1991. Does intraperitoneal CA-125 reflect disease status? Gynecol Oncol 40:66]69 Vergote IB, Onsrud M, Bormer OP, Sert MB, Moen M Ž1992. CA 125 in peritoneal fluid of ovarian cancer patients. Gynecol Oncol 44:161]165 Fleuren GJ, Nap M, Aalders JG, Baptist Trimbos J, de Bruijn HW Ž1987. Explanations for the limited correlation between tumor CA 125 content and serum CA 125 antigen levels in patients with ovarian tumors. Cancer 60:2437]2442 Menczer J, Ben Baruch G, Moran O, Lipitz S Ž1993. Cyst fluid CA 125 levels in ovarian epithelial neoplasms. Isr Obstet Gynecol 81:25]28 Auersperg N, Maines-Bandiera SL, Dyck HG Ž1996. Phenotypic plasticity of ovarian surface epithelium: possible implications for ovarian carcinogenesis, in Ovarian Cancer 4 ŽSharp F, Blackett T, Leake R, Berek J, eds. pp 3]17. Chapman and Hall, London Zurawski VR, Knapp RC, Einhorn N, Kenemans P, Mortel R, Ohmi K, Bast RC, Rits RE, Malkasian G Ž1988. An initial analysis of preoperative serum CA 125 levels in patients with early stage ovarian carcinoma. Gynecol Oncol 30:7]14 Nagele F, Petru E, Medl M, Kainz C, Graf AH, Sevelda P Ž1995. Preoperative CA 125: an independent prognostic factor in patients with stage I epithelial ovarian cancer. Obstet Gynecol 86:259]264 Vuento MH, Stenman U-H, Pirhonen JP, Makinen JI, Laip¨ pala PJ, Salmi TA Ž1997. Significance of a single CA 125 assay combined with ultrasound in the early detection of ovarian and endometrial cancer. Gynecol Oncol 64:141]146 Rice LW, Lage JM, Berkowitz RS, Goodman A, Muto MG, Knapp RC, Bell DA Ž1992. Preoperative serum CA-125 levels in borderline tumors of the ovary. Gynecol Oncol 46:216]229 Skates SJ, Xu F-J, Yu Y-H, Sjovall, Einhorn N, Chang Y, Bast ¨ RC, Knapp RC Ž1995. Towards an optimal algorithm for ovarian cancer screening with longitudinal tumor markers. Cancer 76: 2004]2010 Jacobs I Ž1995. Screening for sporadic ovarian cancer, in The Biology of Gynaecological Cancer ŽLeake R, Gore M, Humphry Ward R, eds. pp 231]244. RCOG Press, London Berek JS, Bast RC Ž1995. Ovarian cancer screening. The use of serial complementary tumor markers to improve sensitivity and specificity for early detection. Cancer 76:2092]2096 Finkler NJ Ž1993. Clinical utility of CA 125 in preoperative diagnosis of patients with pelvic masses. Eur J Obstet Gynecol Repr Biol 49:105]107 Yedema C, Massuger L, Hilgers J, Servaas J, Poels L, Thomas C, Kenemans P Ž1988. Preoperative discrimination between benign and malignant ovarian tumors using a combination of CA125 and CA15.3 serum assays. Int J Cancer Žsuppl.. 3:61]67 Gadducci A, Ferdeghini M, Prontera C, Moretti L, Mariani G, Bianchi R, Fioretti P Ž1992. The concomitant determination of different tumor markers in patients with epithelial ovarian cancer and benign ovarian masses: relevance for differential diagnosis. Gynecol Oncol 44:147]154 Woolas RP, Xu F-J, Jacobs IJ, Yu Y-H, Daly L, Berchuck A, Soper JT, Clarke-Pearson DL, Oram DH, Bast RC Ž1993. Elevation of multiple serum markers in patients with stage I ovarian cancer. J Natl Cancer Inst 85:1748]1751 Jacobs I, Rivera H, Oram DH, Bast RC Ž1993. Differential diagnosis of ovarian cancer with tumour markers CA 125. CA 15-3 and TAG 72.3. Br J Obstet Gynaecol 100:1120]1124 Schutter EMJ, Kenemans P, Sohn C, Kristen P, Crombach G, Westermann R, Mobus V, Kaufmann M, Caffier H, Schmidt¨

R. H. M. Verheijen et al

44.

45.

46.

47.

48.

49. 50.

Rhode P, Kreienberg R, Verstraeten RA, Cornillie F Ž1994. Diagnostic value of pelvic examination, ultrasound, and serum CA 125 in postmenopausal women with pelvic mass. An international multicenter study. Cancer 74:1398]1406 Schutter EMJ, Sohn C, Kristen P, Mobus V, Crombach G, ¨ Kaufmann M, Caffier H, Kreienberg R, Verstraeten AA, Kenemans P Ž1998. Estimation of probability of malignancy using a logistic model combining physical examination, ultrasound, serum CA 125, and serum CA 72-4 in postmenopausal women with a pelvic mass: an international multicenter study. Gynecol Oncol 69:56]63 Yedema CA, Kenemans P, Wobbes Th, Thomas CMG, Bon CG, Mulder C, Voorhorst FJ, Verstraeten AA, van Kamp GJ, Hilgers Ž1992. Use of serum tumor markers in the differential diagnosis between ovarian and colorectal adenocarcinomas. Tumor Biol 13:18]26 Bast RC, Klug TL, SchaetzI E, Lavin P, Niloff JM, Greber TF, Zurawski VR, Knapp RC Ž1984. Monitoring human ovarian carcinoma with a combination of CA 125, CA 19.9, and carcinoembryonic antigen. Am J Obstet Gynecol 149:553]559 Burg van der MEL, Lammes FB, Putten van WLJ, Stoter G Ž1988. Ovarian cancer: the prognostic value of the serum half-life of CA 125 during induction chemotherapy. Gynecol Oncol 30:307]312 Yedema CA, Kenemans P, Thomas CM, Massuger LF, Wobbes Th, Verstraeten A, van Kamp GJ, Hilgers J Ž1993. CA 125 serum levels in the early post-operative period do not reflect tumor reduction obtained by cytoreductive surgery. Eur J Cancer 29a: 966]997 Buller RE, Vasilev S, DiSaia PJ. CA 125 kinetics: a cost-effective clinical tool to evaluate clinical trial outcomes in the 1990s. Am J Obstet Gynecol 174:1241]1254 Yedema CA, Kenemans P, Voorhorst F, Bon G, Schijf C, Beex L, Verstraeten A, Hilgers J, Vermorken J Ž1993. CA 125 half-life in ovarian cancer: a multivariate survival analysis. Br J Cancer 67:1361]1367

51. Rustin GJ, Nelstrop AE, McLean P, Brady MF, McGuire WP, Hoskins WJ, Mitchell H, Lambert HE Ž1996. Defining response of ovarian carcinoma to initial chemotherapy according to CA 125. J Clin Oncol 14:1545]1551 52. Munstedt K, Krisch M, Sachsse S, Vahrson H Ž1997. Serum ¨ CA 125 levels, survival in advanced ovarian cancer. Arch Gynecol Obstet 259:117]123 53. McGuire WP, Hoskins WJ, Brady MF, Kucera PR, Partridge EE, Look KY, Clarke-Pearson DL, Davidson M Ž1996. Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med 334:1]6 54. Davelaar EM, Bonfrer ` JMG, Verstraeten AA, ten Bokkel Huinink WW, Kenemans P Ž1996. CA 125: a valid marker in ovarian carcinoma patients treated with paclitaxel? Cancer 78:118]127 55. Pearl ML, Yashar CM, Johnston CM, Reynolds KR, Roberts JA Ž1994. Exponential regression of CA 125 during salvage treatment of ovarian cancer with Taxol. Gynecol Oncol 53:339]343 56. Bonfrer ` JMG, Linders TC, Hageman PC, Hilkens JGW, Sparreboom A, Molthoff CFM Ž1995. Effect of paclitaxel ŽTaxol. on growth, CA 125 expression and release by ovarian cancer cell lines, in Tumor Markers in Gynecologic Cancer: Basic and Clinical Research Žthesis. ŽBonfrer ` JMG, ed.. 57. Marth C, Zeimet AG, Widschwendter M, Ludescher C, Kaern J, Trope ´ C, Gasl G, Daxenbichler G, Dapunt O Ž1997. Paclitaxel- and docetaxel-dependent activation of CA-125 expression in human ovarian carcinoma cells. Cancer Res 57:3818]3822 58. Schrijvers D, van Dam P, van Oosterom A Ž1996. CA 125: a misleading tumor marker? Int J Obstet Gynecol 55:179]180 59. Makar AP. Kristensen G, Børmer OP, Trope ´ CG Ž1993. Is serum CA 125 at the time of relapse a prognostic indicator for further survival prognosis in patients with ovarian cancer? Gynecol Oncol 49:3]7

124