Early Stage Detection Of Ovarian Cancer

Because of the fact that ovarian cancer is relatively rare in the general population, (incidence of 16 cases per 100,000 women per year), any screening method must have

From: Current Clinical Oncology: Molecular Pathology of Gynecologic Cancer Edited by: A. Giordano, A. Bovicelli, and R. Kurman © Humana Press Inc., Totowa, NJ

an extremely high specificity in order to avoid detecting a large number of false-positives. For example, a screening test with a specificity of 99% and a sensitivity of 99% would detect approx 30 false-positives for each real case of ovarian cancer, leading to a positive predictive value of only 3-5%. For these reasons, it is accepted that a good ovarian cancer marker will require a specificity of more than 99.6% to achieve any clinical relevance for screening of the general population (2). On the other hand, in high-risk populations, screening tests with inferior sensitivity and specificity may still be useful. Generally, it is now believed that the ideal test will be multiparametric, involving multiple different markers or detection techniques (3).

2.1. CA 125/Muc16

An antibody that reacted with an ovarian cancer antigen was identified almost 25 years ago (4) and named OC125. The recognized antigen, a glycoprotein initially named CA 125, was found to be shed into culture supernatant and a serum-based radioimmunoassay was quickly developed (5). Cloning the CA 125 gene turned out to be extremely difficult, but the feat was finally accomplished in 2001 and the complementary DNA was found to have striking similarities to mucin molecules. For this reason, the gene encoding CA 125 was named MUC16. Because the CA 125 protein is secreted, it was hypothesized early that a blood test based on this protein may be useful for early detection of ovarian cancer.

The serum CA 125 assay has been evaluated for ovarian cancer screening, as a tool to differentiate benign from malignant ovarian masses, and as an indicator of tumor status during and after chemotherapy (6,7). Using the assay, it was initially shown that 99% of healthy women have less than 35 U/mL CA 125, whereas 82% of sera from women with epithelial ovarian cancer had levels higher than 35 U/mL (5). Interestingly, in 90% of the cases, levels of CA 125 corresponded to tumor volume when studied longitudinally. CA 125 has also been observed elevated in a variety of benign diseases, including endometriosis. Unfortunately, CA 125 is typically not sufficiently specific or sensitive for screening of the general population. For example, only 50% of stage I patients have CA 125 higher than 35 U/mL (8). In addition, benign gynecological conditions can lead to unacceptably high levels of false-positive tests (low specificity) in premenopausal women, although this problem is less significant in postmenopausal women.

It has been suggested that combining CA 125 with other screening methods, such as transvaginal ultrasonography may improve specificity (9), but the results did not reach the levels that would allow for screening of the general population. It has been suggested that monitoring CA 125 levels in patients for a period of time may lead to improved sensitivity for early detection by providing a baseline and clarifying trends (10,11). However, it is believed that up to 20% of ovarian cancers do not express CA 125, suggesting a maximum theoretical sensitivity of 80% for this marker, regardless of the improvement on detection. It is likely that the theoretical sensitivity of CA 125 for early stage ovarian will be smaller, possibly down to 50%. In addition, the relatively low specificity would lead to many false-positive and a large number of unnecessary surgeries. Combining CA 125 with other markers, such as CA19-9, CA72-4, and CA15-3 has been shown to improve sensitivity of detection (12).

The use of CA 125 as a marker might be most useful in monitoring recurrent ovarian cancer (13). It has indeed been observed that CA 125 tracks disease accurately in more than 80% of ovarian cancer patients. However, this aspect of CA 125 as a bio-marker is more relevant to the next section dealing with prognosis factors and will be discussed later.

2.2. He4/Whey Acidic Protein (WAP)-Type Four-Disulfide Core2

From the beginning of the gene-expression profiling era, it was hypothesized that detailed knowledge of gene expression in cancer might lead to the identification of candidate tumor markers. Microarrays, serial analysis of gene expression, and EST analysis have been used to study a variety of human tumors, including ovarian cancer. HE4/WAP-type four-disulfide core (WFDC)2 was one of the first new ovarian cancer candidates identified using complementary DNA arrays (14,15). This finding was confirmed soon afterwards using serial analysis of gene expression (16). HE4/WFDC2 contains two WFDC domains, which are known to function as protease inhibitors, although no protease inhibition activity has yet been identified for HE4. A blinded study with ovarian cancer patients and controls recently demonstrated that HE4 has similar specificity and sensitivity as CA 125, although the HE4/WFDC2 assay appeared to be less likely to yield false-positive in patients with nonmalignant diseases (17). Therefore, HE4/WFDC2, in combination with CA 125 or other markers may represent a promising approach for general screening of the population (18). Screening trials are currently being conducted to investigate this possibility.

2.3. Kallikreins

The kallikreins family of serine proteases includes 15 members, which share significant homology, but whose exact physiological functions remain unknown (19). Many members of the kallikrein family have altered expression in ovarian and other cancers. Interestingly, hK3, also known as prostate specific antigen has been widely used for prostate cancer screening and is probably the most useful tumor marker to date. Recently, it has been hypothesized that certain members of the kallikrein family might also be useful for ovarian cancer screening. Specifically, hK6 (20), hK10 (21), and hK11 (22) have all been shown to be elevated in the serum of a majority of ovarian cancer patients. Interestingly, the sensitivity of hK6 and hK10 for early stage (stages I and II) was found to be approx 25% (at 90% specificity), but this figure could be increased to more than 90% when combined with CA 125 (20,21). However, large and detailed clinical studies with exact sensitivity and specificity figures still need to be done. As suggested earlier, because hKs may have different patterns of expression compared with CA 125, a test combining both types of markers might be useful.

2.4. Proteomics and Serum Patterns

Recently, it was suggested that proteomics patterns in serum may be useful in ovarian cancer detection. A study utilizing mass spectrometry identified protein patterns in the serum of women with ovarian cancer and compared these patterns with those observed in healthy women (23). A specific protein pattern was identified that was capable of recognizing ovarian cancer in a population of women at risk with a sensitivity of 100% and a specificity of 95%, which is probably not sufficient for screening of the general population. It is worth noting that the exact identities of the proteins making up the "pattern" are unknown. Therefore, it is unclear whether the proteins making up the pattern are produced by the tumor itself or by a host reaction to the tumor. In this case, the results may not be specific for ovarian cancer or even for cancer at all. Another weakness is that the validation in the initial paper was performed on cancers from "at risk" women, and it is unclear whether sporadic ovarian cancers will yield the same pattern, because they are known to have quite a different biology. Finally, the specificity of the assay is not sufficient for screening of the general population. Although, in its infancy, this is a promising approach that may eventually be useful in the detection of ovarian or other cancers.

Another proteomics approach identified specific proteins that are altered in the serum of women with ovarian cancer (24). In that study, three proteins were found to be differentially expressed in the serum (ApoAl [decreased]; truncated transthyretin [decreased]; a fragment of a-trypsin inhibitor heavy chain H4 [elevated]). When combined with CA 125, these three biomarkers lead to a significant improvement in sensitivity more than CA 125, demonstrating again that combining biomarkers can indeed be desirable. However, the sensitivity of 74% (at a specificity of 97%) was still insufficient for this approach to be successfully applied to the general population.

2.5. Others Markers and Combinations of Markers

Large-scale gene-expression methods have identified many genes that are overex-pressed in ovarian cancer (14-16,25-45). These data have provided a wealth of candidates that may be useful as potential tumor markers, especially, the genes that are known to encode secreted proteins. The candidates identified independently by multiple studies are particularly promising. Among genes encoding secreted proteins, HE4, osteopontin, SLPI, and SPINT1 were identified by multiple studies as upregulated in ovarian cancer (46). These candidates are currently being evaluated by various groups for their potential as screening markers. In addition, many membrane proteins identified by multiple studies may also be useful in diagnosis and therapy. These candidates include Ep-CAM, MUC1, SPINT2, CD9, CLDN3, CLDN4, and HER3.

Because it appears that no single marker is elevated and secreted in all ovarian cancers, it is likely that a combination of markers will be necessary to detect a majority of ovarian cancers. In addition, biomarker-based test may be combined with other techniques, such as transvaginal sonography to attain yet higher specificities and sensitivities (3). As mentioned earlier, a combination of CA 125, CA19-9, CA72-4, and CA15-3 has been shown to improve sensitivity, but the usefulness of the added markers was restricted to initial diagnosis. Thus, it did not provide any advantage in the follow-up and detection of recurrent disease. In addition, using immunohistochemistry, a recent study demonstrated that a combination of four markers (CLDN3, CA 125, MUC1, and VEGF) could identify all 158 ovarian cancers tested, including eight early stage serous cancers (45). It remains to be seen whether this combination of proteins will also be useful when adapted to a blood test for screening the general population.


Because of the wide variability in the clinical outcome of ovarian cancer patients, it is clear that the identification of molecular markers that could predict overall survival or response to chemotherapy with better accuracy than the classical prognosis factors

(tumor grade, stage, and so on) might be of significant clinical value. For example, the identification of patients with tumors that have a high probability of developing resistance to conventional chemotherapy might make these patients ideal candidates for alternative or novel therapeutic regimens. In this section, the major molecular markers that have been investigated as possible ovarian cancer prognostic factors are reviewed.

3.1. Cell-Cycle Regulators

Cancers are characterized by aberrant proliferation resulting from alterations in cell-cycle regulatory mechanisms. In addition, many oncogenes and tumor suppressor genes have been implicated in pathways regulating the cell cycle, providing a direct mechanistic link between cell-cycle control and tumorigenesis. On the other hand, it is unclear whether any of the cell-cycle components can be used as tumor marker for prognosis. In this section, the current data on the main cell-cycle regulators investigated for their prognostic value in ovarian cancer will be reviewed.

The p53 gene encodes a nuclear phosphoprotein that can bind specific DNA sequences and function as a transcription factor to positively or negatively regulate the transcription of other growth regulatory genes. p53 is involved in cell-cycle control, DNA damage response, stress response, cell senescence, genomic stability, and apopto-sis (47). p53 is mutated in the germline of Li-Fraumeni patients, a cancer predisposition syndrome and it is believed to be one of the most commonly mutated genes in human cancer, as it is found somatically altered in about 50% of all cancers (48-50). Most missense mutations in p53 appear to change the conformation of the protein, leading to increased stability and higher steady-state levels. Indeed, a close correlation between p53 immunoreactivity and p53 mutation has been shown (51).

Soon after the realization that p53 was a major player in human cancer, ovarian cancer was also shown to frequently overexpress the p53 protein and contained mutations in at least 50% of the cases, regardless of stage (52-55). A large number of studies have investigated the prognostic ability of p53 mutations and/or overexpression in ovarian cancer. There is no consensus on the predictive value of p53 in the literature. p53 has been shown to be predictive of overall survival by many groups (56-79). However, when analyzed by multivariate analysis, few of these studies found that p53 was an independent prognostic marker (58,61,64-66,70,71,74,76,77). On the other hand, a large number of reports have found no predictive value at all for p53 (53,68,80-96). It has been shown that, in certain cases, p53 status can be predictive of response to chemotherapy, and could therefore be useful in tailoring treatment to individual patients (60,62,66,67,97-101). p53 may therefore have a role in determining the sensitivity of ovarian tumors to chemotherapy. This would be consistent with the role of p53 in DNA damage. However, as is the case for survival, many studies have showed that p53 status does not appear to affect the response to chemotherapy (56,64,68,74,83,85,89,94,102). Table 1 summarizes these findings.

Overall, because of the lack of reproducibility in various studies, it has been concluded that p53 does not represent a robust marker for prediction of either survival or response to chemotherapy. The lack of reproducibility may be attributed to several factors, such as difference in the patient cohort (sample size, differences in treatment, grade, and so on) and different techniques for p53 alteration detection (mutation vs

Table 1

Studies Evaluating the Prognostic Value of p53

Survival Chemotherapy

Table 1

Studies Evaluating the Prognostic Value of p53



Positive (%)








10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook

Post a comment