In the Western world, gynaecological cancers represent approximately 15% of all cancers in women and are responsible for approximately 10% of all cancer deaths. In terms of frequency, endometrial cancers are the most common, followed by cancers of the ovary and the uterine cervix. However, ovarian cancer has the highest mortality rates with cervical and endometrial ranked second and third, respectively. The aim of this publication is to review the role of tumour markers in these three gynaecological malignancies.

CANCER OF THE OVARY

Epidemiology and aetiology

The incidence of ovarian cancer is highest in the Western world with a lifetime risk of developing this malignancy being approximately 1:70. In the United States, ovarian cancer is the fifth main cause of death with an estimated incidence of 50 cases per 100,000 women. Annually in the United States, there are approximately 20,000 new cases with about 12,500 deaths (1,2).

Although the exact aetiology of ovarian cancer is unknown, the most consistent risk factor is ovulation. The risk of ovarian cancer decreases with increasing parity, anovulation and the use of certain oral contraceptives. In 5-10% of ovarian cancers, the presence of a family history may indicate a genetic factor (1).

Classification

Over 90% of ovarian malignancies are epithelial tumours arising form the coelomic epithelium. Most of the remaining types are either germ cell or sex cord stromal cancers. Epithelial cancers can be further classified based on the cell type into serous, mucinous, endometrioid, clear cell, mixed epithelial, undifferentiated, transitional and squamous cell (2). Of these histological sub-types, serous cell is the most common.

Tumour markers

The best available marker for epithelial ovarian cancer is the mucin, CA125. The reference interval most frequently quoted for CA125 is 0-35 kU/L, although 99% of apparently healthy post-menopausal women have levels below 20 kU/L. In apparently healthy pre-menopausal women, levels of 100 kU/L or higher can occur during menses (3,5). Although elevated CA125 levels are found in approximately 80% of all patients with epithelial ovarian cancer, high levels are found in only about 50% of patients with FIGO Stage 1 disease (4,5).

This lack of sensitivity for early disease, and the fact that CA125 can be elevated in multiple benign diseases (Table 1), limits the use of CA125 for the diagnosis of early epithelial ovarian cancer. A further limitation is that CA125 can be elevated in adenocarcinomas other than ovarian cancer. Although CA125 can also be elevated in germ cell tumours of the ovary (4), the markers of choice for this type of ovarian cancer are a -fetoprotein (AFP) and human chorionic gonadotrophin (hCG) and its b -subunit (hCGb ).

Tumour marker applications in epithelial ovarian cancer

Screening

The lack of early symptoms means that approximately 70% of patients with ovarian cancer present with advanced disease. While the overall five-year relative survival rate is of the order of 30%, the survival rate for Stage III and IV disease combined is only 10% (1). In contrast, a five-year survival of 90% may be achieved for patients with early stage disease confined to the ovary (1).

As patient outcome is strongly related to disease stage at diagnosis, it has long been believed that early diagnosis could result in better prognosis. Consequently, in recent years a number of studies have been carried out investigating the use of the serum marker CA125, as well as abdominal ultrasound, transvaginal ultrasound and bi-manual pelvic examination in screening for ovarian cancer (3,5).

As a screening test, the main problems with CA125 are lack of sensitivity for early stage disease (only about 50% of patients with Stage I have elevated levels) and lack of specificity (Tables 1 and 2). Because of these limitations and the absence of prospective randomized trials documenting a benefit for the early diagnosis of ovarian cancer, the use of CA125 cannot be recommended at present for general population screening to detect sporadic forms of this disease. CA125, however, in combination with transvaginal ultrasound may have a role in the early detection of ovarian cancer in women with a hereditary ovarian cancer syndrome (5). Although there are no data showing that screening these high risk women can reduce their mortality from ovarian cancer, a National Institutes of Health (NIH) consensus statement has recommended that these women undergo at least annual recto-vaginal pelvic examination and transvaginal ultrasound and that serum CA125 is measured (5).

Diagnosis

Diagnosis of ovarian cancer is usually carried out by surgery followed by histopathology. However, pre-operative serum levels of CA125, especially in post-menopausal women, may be useful in the differential diagnosis of benign and malignant pelvic masses. In one multi-centre prospective study, 228 post-menopausal women with pelvic masses were evaluated with CA125 determination, transvaginal ultrasound and pelvic examination (6). The accuracy of CA125 (cut-off concentration 35 kU/L) in differentiating between benign and malignant masses was 77%, which was almost identical to accuracy achieved with pelvic examination and ultrasound(76% and 74% respectively). Using logistic regression analysis, CA125 was a more powerful predictor of disease diagnosis than ultrasound but less strong than pelvic examination. Significantly, no cancer was found in any subject in which all three tests were negative (6). Pre-operative levels of CA125 can thus contribute to optimal pre-operative and intra-operative management (6). Patients with suspected ovarian cancer should be referred to a gynaecological surgeon. On the other hand, women likely to have a benign lesion could avoid such referral as well as major surgery (6).

Prognosis

The traditional prognostic factors for ovarian cancer include tumour stage, grade, histological type and size of residual tumour after primary debulking (cytoreductive surgery). However, multiple studies have now shown that CA125 levels after either 1, 2 or 3 courses of chemotherapy is one of the strongest available indicators of disease outcome (7). A prolonged half-life for CA125, or a less than 7-fold decrease during the early months of treatment, has also been shown to predict poor outcome (7). A further study has shown that a CA125 concentration >70 kU/L before the third course of chemotherapy was the single most important factor for predicting disease progression at twelve months (7). In this latter report, the inclusion of residual disease bulk at the end of initial surgery only marginally improved the predictive ability (7).

Monitoring

The most important application of CA125 is in the monitoring of patients with epithelial ovarian cancer. Serial CA125 levels can pre-clinically detect recurrent disease with lead times of 1-17 months (median 3-4 months) (3, 4). Furthermore, longitudinal monitoring with this marker has the potential to detect recurrent disease earlier and more cost-effectively than radiological procedures (8). While early detection of recurrent disease may lead to altered patient management, no study has yet shown that this leads to enhanced survival.

Several groups have investigated whether serial CA125 assays can replace second-look laparotomy in the follow-up of patients with cancer of the ovary. The main conclusion from these studies is that while high levels of CA125 (>35 kU/L) are almost always associated with disease recurrence, a low level does not necessarily indicate the absence of disease. Jacobs and Bast (3) summarised the results from 15 studies on the value of CA125 assays in predicting findings at second-look laparotomy, and concluded that the marker had an accuracy of 67%. Residual disease was detected in 156 of 165 (94.8%) of patients with an elevated CA125 value. However, disease was also present in 180 of 414 women with a normal CA125 level (3). This was most frequently observed in patients with low-volume disease. Thus, CA125 was only raised in 21% of patients described as having microscopic disease, but was elevated in 79% of the patients where recurrent tumour was greater than 1 cm in diameter at laparotomy (3).

Other potential markers for epithelial ovarian cancer

While CA125 is the marker of choice for epithelial ovarian cancers, other markers may complement its use. These complementary markers include CA15.3, OVX1, CA72.4 and cytokeratin 8 and 18 markers (TPA and TPS) (9,10). Further work is required to establish how these markers can contribute to the diagnosis and management of patients with ovarian cancer.

Use of tumour markers in ovarian germ cell tumours

While AFP and hCG are the main markers for ovarian germ cell tumours, the particular marker expressed depends on the state of differentiation of the tumour as well as on histological type (11). Thus, while undifferentiated cancers of primordial germ cell origin (e.g. dysgerminoma) may synthesise neither AFP or hCG, tumours with trophoblastic elements usually produce hCG and those with yolk sac components frequently express AFP. [See this issue, pp XX]

Immunohistochemical determination of both AFP and hCG in tumour tissue may aid pathological diagnosis and classification. High initial serum concentrations of AFP appear to predict poor outcome using univariate analysis but not with multivariate analysis (11). Finally, both AFP and hCG are necessary in the follow up of patients with diagnosed ovarian germ cell cancers.

CANCER OF THE CERVIX

World-wide, cervical malignancy ranks second to breast cancer as the main cause of cancer mortality in women. Risk factors include low socio-economic class, multi-parity, early marriage and multiple sex partners. Infection with human papilloma virus (HPV) (types 16, 18, 31 and 45) also plays a role in the genesis of cervical cancer. In the United States, the overall five-year survival for this malignancy is approximately 70%. However, for patients diagnosed with early stage disease, the five-year survival increases to almost 90%.

Classification

About 90% of the cervical cancers are of the squamous cell type. The other types are mostly adenocarcinomas or adeno-squamous carcinomas. Very rarely, stem cell carcinomas, sarcomas or neuroendocrine cancers can be found.

Tumour markers in cervical cancer

For squamous cell carcinomas, squamous cell carcinoma antigen (SCCA) is the marker of choice. SCCA is a 48 kDa protein with strong homology to a family of protease inhibitors known as serpins. Sensitivity of SCCA for carcinoma of the cervix varies from less than 30% for Stage I disease to over 90% for Stage IV disease (12). SCCA can also, however, be elevated in other squamous cell cancers (e. g. lung, head and neck, oesophagus and vagina), as well as in benign diseases of the skin (e. g. psoriasis, eczema), lung (e.g. sarcoidosis), liver and kidney.

Use of tumour markers in cervical cancer

Screening

Screening asymptomatic women for cervical cancer using the Papanicolau (Pap) stain has been shown to reduce both the incidence and mortality of this malignancy in many Western countries. Using this test, pre-invasive lesions known as carcinoma in situ (CIS) or cervical intraepithelial neoplasia (CIN) can be detected. The peak incidence of these lesions can occur ten to fifteen years before the peak incidence of invasive cancer. Serum tumour markers currently make no contribution to screening for cervical cancer.

Diagnosis

The diagnosis of cervical cancer is based on histopathology. Because of lack of specificity and sensitivity, markers contribute little to the diagnosis of early disease. For SCCA in the squamous cell type, serum levels generally increase with tumour volume, stage and lymph node involvement (13). Pre-treatment levels may be an independent predictor of lymph node metastasis (14).

Prognosis

In multiple studies, high pre-treatment levels of SCCA have been found to predict poor outcome in patients with squamous cell cancer (13,14,15). In some of these studies, SCCA was reported to be an independent prognostic marker (13,14), i.e. independent of histology, and FIGO stage and grade, but in other reports it was not (15). In adenocarcinomas of the cervix, CA125 but

not SCCA was found to be a prognostic indicator (16).

Monitoring

SCCA has a number of potential applications in the follow-up of patients with diagnosed squamous cell carcinoma of the cervix. These include early detection of recurrent disease, monitoring of neo-adjuvant chemotherapy, and monitoring treatment of recurrent disease. In the follow-up of cervical cancer, the overall correlation between clinical course of disease and variation in SCCA levels is usually greater than 80% (13). Serial SCCA levels can detect recurrent disease pre-clinically, with lead-times of two to six months, and this may be of value in selecting patients for salvage radiotherapy or surgery. At present however, a clinical value in monitoring patients with SCCA has not been shown.

Other tumour markers in cervical cancer

Other markers investigated for cervical cancer include CEA, hCG beta core fragment (hCGb cf), and cytokeratins (TPA and CYFRA 21-1) (15). Based on available evidence, use of these markers for cervical cancer cannot be recommended at present.

CANCER OF THE ENDOMETRIUM

Although endometrial cancers comprise about 50% of all female genital cancers, these cancers cause fewer than 15% of all gynaecological cancer deaths, at least in many Western countries. The low death rate appears to be primarily due to early diagnosis. Approximately 80% of all endometrial cancers are diagnosed while confined to the uterus. Survival rates vary from approximately 80% for Stage I disease to 10% for Stage IV disease.

Classification

Sixty to 80% of all endometrial cancers are adenocarcinomas. Less common types include serous, mucinous, papillary, clear cell and squamous carcinomas.

Tumour markers

The marker most frequently elevated in endometrial cancer is CA125. Combining results from five separate studies, Jacobs and Bast (3) calculated that high levels of CA125 (>35 kU/L) were present in 32% of 267 patients with endometrial cancer. In patients with Stage I and II disease combined, elevated concentrations were found in 22% while in patients with Stage III and IV disease, 82% had high levels (3).

Use of markers in endometrial cancer

Screening

No existing marker has the necessary sensitivity and specificity to screen for early endometrial cancer.

Diagnosis

The traditional method of diagnosing endometrial malignancy is histological examination of tissue obtained at curettage. As CA125 is elevated in only a minority of patients with low stage disease, it contributes little to early diagnosis.

Monitoring

For monitoring patients with endometrial cancer, CA125 is the best available marker. Elevated levels are found in approximately 60% of patients with recurrent endometrial malignancy. As with SCCA in cervical cancer, there is no evidence at present that follow-up with serial CA125 levels enhances patient outcome. The role of CA125 in monitoring patients with uterine serous carcinoma is of particularly questionable value (17).

CONCLUSION

The most useful marker in gynaecological cancer is CA125 for epithelial ovarian malignancy. Although multiple studies have evaluated the potential of CA125 to screen for ovarian cancer in asymptomatic women, an insufficient number of women have so far been investigated to show whether or not early detection improves survival. A randomized trial is necessary to determine whether mortality from ovarian cancer can be decreased by screening. Such trials involving CA125, transvaginal sonography and physical examination are presently in progress .

The most useful application of CA125 at present is in the monitoring of ovarian cancer patients being treated by chemotherapy. Patients whose tumours are not responding can be spared the side-effects and costs of ineffective therapy. In addition, CA125 determinations if performed at the appropriate times may be more cost-effective than radiological procedures. CA125 in endometrial adenocarcinomas and SCCA in squamous cell carcinoma of the cervix may provide useful information but the clinical value of this data remains to be determined.

References

1. Austoker J: Screening for ovarian, prostatic and testicular cancer. Br Med J 309: 315-320, 1994.
2. Kristensen GB, Trope C: Epithelial ovarian carcinoma. Lancet 349: 113-117, 1997.
3. Jacobs I, Bast RC: The CA125 tumour-associated antigen: a review of the literature. Human Reprod 4: 1-12, 1989.
4. Bonfrer JMG, Korse CM, Verstraeten RA, van Kamp GJ, Hart AAM, Kenemans P: Clinical evaluation of the Byk LIA-mat CA125 II assay: discussion of a reference value. Clin Chem 43: 491-497, 1997.
5. NIH Consensus Development Panel on Ovarian Cancer: NIH Consensus Conference, Ovarian Cancer: screening, treatment and follow-up. JAMA 273: 491-497, 1995.
6. Schutter EMJ, Kenemans P, Sohn C, Kristen P, Crombach G, Westermann R, et al: Diagnostic value of pelvic examination, ultrasound and serum CA125 in post-menopausal women with a pelvic mass. Cancer 74: 1398-1406, 1994.
7. Rustin GJS: The clinical value of tumour markers in the management of ovarian cancer. Ann Clin Biochem 33: 284-289, 1996.
8. Rustin GJS, Nelstrop A, Stilwell J, Lambert HE: Savings obtained by CA125 measurements during therapy for ovarian carcinoma. Eur J Cancer 28: 79-82, 1992
9. Berek JS, Bast RC: Ovarian cancer screening, the use of serial complementary markers to improve sensitivity and specificity for early detection. Cancer 76: 2092-2096, 1995.
10. Shabana A, Onsrud M: Tissue polypeptide-specific antigen and CA125 as serum tumor markers in ovarian carcinoma. Tumor Biol 15: 361-367, 1994.
11. Bower M: The value of tumor markers in ovarian germ cell tumors. Tumor Marker Update 8: 1-7, 1996.
12. Duk JM, De Bruijn H, Groenier KH, Hollema H, ten Hoor KA, Krans M, et al.: Cancer of the uterine cervix: sensitivity and specificity of serum squamous cell carcinoma antigen determinations. Gynecol Oncol 39: 186-194, 1990.
13. Scambia G, Panici BP, Foti E, Amoroso M, Salerno G, Ferranina G et al. Squamous cell carcinoma antigen: prognostic significance and role in the monitoring of neo-adjuvant chemotherapy response in cervical cancer. J Clin Oncol 12: 2309-2316, 1994.
14. Duk JM, Klaas H, de Bruijn H, Hollema H, ten Hoor KA, van der Zee A, et al: Pre-treatment serum squamous cell carcinoma antigen: a newly identified prognostic factor in early stage cervical cancer. J Clin Oncol 14: 111-118, 1996.
15. Gaarenstroom KN, Bonfrer JMG, Kenter GG, Korse CM, Hart AAM, et al.: Clinical value of pre-treatment serum CYFRA 21-1, tissue polypeptide antigen and squamous cell carcinoma antigen levels in patients with cervical cancer. Cancer 76: 807-813, 1995.
16. Duk JM, de Bruijn H, Klaas KH, Fleuren GJ, Aalders JG: Adenocarcinoma of the uterine cervix: prognostic significance of pre-treatment serum CA125, squamous cell carcinoma antigen and carcinoembryonic antigen levels in relation to clinical and histopathologic tumor characteristics. Cancer 65: 1830-1837, 1990.
17. Price FV, Chambers SK, Carcangiu ML, Kohorn EI, Schwartz PE, Chambers JT: CA125 may not reflect disease status in patients with uterine serous carcinoma. Cancer 82: 1720-1725, 1998.

TABLE 1. Approximate proportion of patients with various benign diseases giving rise to elevated levels of CA125. Please note that menstruation and the first term of pregnancy may also result in high levels of this marker. [Data summarised from Reference 3].

 

Disorder	Approximate % of patients with serum CA125 >35 kU/L
Endometriosis	24
Benign ovarian tumours	10
Acute salpingitis	40
Chronic salpingitis	8
Uterine myoma	10
Cirrhosis	67
Cirrhosis with ascites	100
Chronic active hepatitis	10
Acute pancreatitis	32
Chronic pancreatitis	2
Renal failure	15

TABLE 2. Approximate proportion of patients with non-ovarian malignancies giving rise to elevated levels of CA125. [Data summarised from Reference 3].

Malignancy	Approximate % of patients with serum CA125 >35 kU/L
Endometrial	31-33
Cervical	25
Breast	17-20
Colorectal	15
Pancreas	52-58
Lung	29-37
Liver	50-78
Gastric	30
Biliary tract	46

Back