| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Chair of Obstetrics and Gynecology, Department of Pediatrics, Obstetrics, and Reproductive Medicine (A.T., S.L., F.P.), and Unit of Surgical Oncology, Department of Human Pathology and Oncology (D.M., F.R.), University of Siena, 53100 Siena, Italy; and Department of Obstetrics and Gynecology (F.M.R., A.L.S.-F., S.A.T.), University of Minas Gerais, 30170-120 Belo Horizonte, Brazil
Address all correspondence and requests for reprints to: Felice Petraglia, M.D., Department of Pediatrics, Obstetrics, and Reproductive Medicine, University of Siena, Policlinico, S. Maria alle Scotte Viale Bracci, 53100 Siena, Italy. E-mail: petraglia{at}unisi.it.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The aim of the present study was to evaluate whether the serum concentration of total inhibin has the sensitivity/specificity characteristics to become a diagnostic test for epithelial ovarian cancer in postmenopausal women.
Design: This was a controlled, cross-sectional study.
Setting: The study was conducted at the University of Siena.
Patients: Blood specimens were collected from postmenopausal women with: 1) epithelial ovarian cancer, stage II-III (n = 89); 2) benign ovarian tumors (n = 25); 3) breast (n = 10), colon (n = 10), and stomach (n = 10) cancers; and 4) controls (n = 95). In the group of women with epithelial ovarian cancer, blood specimens were also collected after surgical removal of the tumor. In four cases of women with stage IIC mucinous tumor, blood specimens were collected during the follow-up time.
Intervention: Total inhibin was measured by a new double-antibody ELISA.
Results: Women with epithelial ovarian cancers showed serum total inhibin levels significantly higher than those with benign tumor or with nonovarian tumors or controls (P < 0.001). Patients with serous (n = 40) or mucinous tumors (n = 17) showed the highest total inhibin levels (P < 0.001). At 95% specificity, the total inhibin assay detected 37 of 40 (93%) serous tumors and 16 of 17 (94%) mucinous tumors. When total inhibin was combined with CA-125, all cases of serous and mucinous tumors were detected, and the overall sensitivity for epithelial ovarian cancers was 99% at 95% specificity. A significant decrease of total inhibin levels was shown in women with serous and mucinous carcinoma as result of surgery (P < 0.001). In the four women who were followed up, recurrence was associated to an increase of total inhibin levels.
Conclusions: The present data show that total inhibin is a sensitive and specific marker of epithelial ovarian cancers in postmenopausal women. Total inhibin may therefore be combined with CA-125 for noninvasive diagnosis of epithelial ovarian cancer and may also be a useful serum marker to monitor disease-free intervals.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Human ovary releases two forms of inhibins, namely inhibin A and inhibin B. They are glycoproteins: inhibin A consists of an 
-subunit disulfide linked to a ßA-subunit [ + ßA], and inhibin B consists of an -subunit and ßB-subunit [ + ßB] (3). Because ß-subunits are produced essentially by granulosa cells and -subunit is also expressed by the theca interna cells in addition to the granulosa cells, the ovarian expression of the -subunit mRNA is higher than that of ß-subunit mRNA (4, 5). The -subunit is unique to all inhibin molecules. The human -subunit is synthesized as a 53-kDa, 366-amino acids (aa) precursor containing an 18-aa signal sequence, a 43-aa proregion also called pro-, and a 305-aa mature form. The mature form is divided into a 171-aa N-terminal segment and a 133- to 134-aa C-terminal segment (C).
After menopause, the depletion of ovarian follicles is associated with a decrease of serum inhibin A and inhibin B levels to nondetectable or near nondetectable levels (6, 7). Some ovarian tumors, however, produce and release inhibins into circulation, even in the postmenopause. Whereas granulosa cell tumors can be identified by the markedly elevated levels of inhibin B (8), epithelial ovarian cancers are poorly detected with inhibin A, inhibin B, and inhibin pro- C assays due to the large intersubject variation in the proportion of inhibin forms released by the same type of tumor (9, 10, 11, 12, 13). Thus, epithelial ovarian cancers are preferably diagnosed by nonspecific assays, which detect simultaneously all inhibin forms in circulation (10, 11).
Pioneer methods of measuring circulating inhibins in a nonspecific way include the Monash RIA (9) and C immunofluorometric assay (12), but these assays are not suitable for wide use in clinical routine (13). The recent development of a new total inhibin immunoenzymatic assay (ELISA), which detects the free inhibin -subunit and the inhibin dimers, prompted us to evaluate its value as tumor marker in women with epithelial ovarian cancers.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Serum samples were collected and analyzed from postmenopausal women with: 1) epithelial ovarian cancers (stage II-III) (n = 89) including serous (n = 40), mucinous (n = 17), endometrioid (n = 10), and clear cell adenocarcinoma (n = 22); 2) benign epithelial ovarian tumors (n = 25); and 3) breast (n = 10), colon (n = 10), or stomach (n = 10) cancers; and 4) normal controls (n = 95). The control subjects were postmenopausal women attending our outpatient clinic for routine gynecologic check-up, in whom thorough clinical and pelvic examination and transvaginal ultrasound did not show any evidence of ovarian tumors or any type of cancer. Patients and controls had similar ages (median, 68 yr and 65 yr, respectively; range, 55–85 yr) and at least 2 yr elapsed since menopause.
In the group of patients with epithelial ovarian cancer, serum samples were also collected after surgery. In addition, four patients with stage IIC mucinous tumor were monitored for relapse over a period of 5 yr and had longitudinal blood samples analyzed for total inhibin concentration. Histopathology reports were reviewed to confirm diagnosis for each patient with epithelial ovarian cancer and the stage of the disease was determined according to the International Federation of Gynecology and Obstetrics. The research ethics committees of all participating hospitals approved the study protocol.
Specimens were collected from a peripheral vein and centrifuged at 3000 rpm for 10 min, and the serum was stored at –20 C until assayed.
Assay for total inhibin
Total inhibin was measured according to the description by the manufacturer Diagnostic System Laboratories/Oxford Bio Innovation (Oxford, UK). This new inhibin assay detects both dimeric and free -subunits and is more sensitive and rapid than previous methods. It uses two monoclonal antibodies: one detection antibody (PO#23) that recognizes aa 79–98 of the inhibin C-subunit, and one capture antibody (R1) directed to aa 1–32 of the inhibin C-subunit. The characteristics of these antibodies, including the chromatographic profile of the molecular forms of inhibin detected, have been described elsewhere (13, 14). The assay detection limit is 2.5 pg/ml, and the coefficients of variation within and between plates were 7.9 and 8.6%, respectively.
Assay for CA-125
CA-125 analysis was performed by Cobas Core CA-125 enzyme-immunoassay analysis kit (Roche, Basel, Switzerland). The sensitivity of the assay was less than 1 U/liter, and the intra- and interassay variations were less then 5.6 and 7.8%, respectively. A calibrator provided by the manufacturer stated that the mean serum concentration in 172 healthy individuals was 12.3 U/liter, with an upper limit of 30.4 U/liter. For routine use, the manufacturer recommends CA-125 concentrations exceeding 35 IU/liter to be considered increased.
Statistical analysis
Because total inhibin and CA-125 concentrations departed significantly from normal distribution, data are presented as medians and interquartile ranges. Differences between groups were tested by Kruskal-Wallis ANOVA followed by Dunn’s test for multiple comparisons, whereas within-subject differences were analyzed by the Wilcoxon nonparametric test for paired observations. All differences were considered significant if P < 0.05. Receiver operator characteristics (ROC) curves were obtained with their respective areas under the curves and 95% confidence intervals, using the Analyze-it software package to Microsoft Excel and GraphPad Prism (version 3.0, San Diego, CA). The combination of total inhibin and CA-125 was tested by MultiROC analysis (15), defining positive test as CA-125 higher than the 95% specificity cutoff (i.e. the 95th percentile of control group) or total inhibin higher than variable cutoff values. Because MultiROC curves do not reach the y-axis (see Fig. 2A
), partial areas under the ROC curves were calculated starting at the beginning of the MultiROC curve, and then rescaled to a 1 x 1 square (16).
|
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). Patients with serous or mucinous tumors showed the highest concentrations of total inhibin, and patients with endometrioid cancer also showed levels higher than in controls (Fig. 1). In all groups of women with epithelial ovarian cancers, serum CA-125 levels were significantly higher than in controls (Fig. 1).
|
|
, the percentage of cancers detected at 95% specificity varied according to the histological subtype but was always improved by the combination of total inhibin and CA-125. In detail, the detection rate of all tumors raised from 84–87% with single markers to 99% with combined markers (P < 0.05; Table 2). The addition of total inhibin increased the CA-125 detection rate for mucinous tumors from 14 of 17 (82%) to 17 of 17 cases (100%) without loosing specificity (95%). Remarkably, the detection rate of clear cell adenocarcinomas increased from 59–68% with single markers to 96% with combined markers (P < 0.05; Table 2).
|
, the cutoffs pointed by ROC curve analysis as those yielding true negative results in 95% of the control subjects (95% specificity) were 10 pg/ml for total inhibin and 32 U/liter for CA-125. For combined markers, the definition of positive test that corresponded to 95% specificity was total inhibin higher than 11 pg/ml or CA-125 higher than 32 U/liter (Table 2).
The diagnostic performance of total inhibin, compared with CA-125 and that of both markers combined, are summarized in Table 3. The area under the ROC curve (depicted in Fig. 2) was similar for both markers isolated but increased significantly (P < 0.05) when total inhibin was combined with CA-125 for the detection of epithelial ovarian cancers. Whereas the positive likelihood ratio was maintained due to the adoption of cutoffs yielding 95% specificity, the negative likelihood ratio decreased remarkably when both markers were associated to improve the test sensitivity (Table 3).
|
). A significant decrease as result of surgery was observed in all four histological subtypes, markedly in the serous (P < 0.0001) and mucinous (P < 0.001) tumor subgroups. CA-125 levels decreased significantly after surgery in all groups (data not shown). There was a weak correlation between total inhibin and CA-125 levels before surgery (Spearman’s r = 0.218) but not after surgery (r = 0.008).
|
). Before surgery, total inhibin levels in all cases were elevated, with significant decrease after surgery and rebounds coinciding with disease recurrence (Fig. 4).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In common with its predecessor nonspecific inhibin assays (9, 10, 11, 12), the current method was able to detect mucinous tumors with greater sensitivity than other epithelial ovarian cancers. However, the detection rate of serous tumors was high in the present study, reaching 93% with the cutoff point 11 pg/ml. This is intriguing because a previous study using the same pair of monoclonal antibodies has shown a much lower detection rate for serous tumors (13), which makes less probable that the current assay has higher sensitivity for this type of cancer. Most likely, this discrepancy might reflect the genetic variability of these tumors along with differences in tumor size because most of our cases were in advanced stage. No other cancer (breast, colon, stomach) or benign epithelial ovarian tumors showed any increase of total inhibin levels, supporting the specificity of ovarian cancer in secreting those precursors, subunits, and mature forms of inhibin.
Although the present assay is named total inhibin, it is unlikely that all inhibin forms are equally immunoreactive; therefore, the assay results are not simply the arithmetical sum of dimeric inhibins, free -subunits, and precursor forms because these molecules may react with the assay antibodies to various extents (13). The absolute numbers of total inhibin concentration reflect the unique characteristics of the pair of antibodies used in the current assay. Thus, we cannot assume that other assays also claiming to measure total inhibin will necessarily give the same numbers for each sample or have the same performance in detecting ovarian cancer.
The cellular origin of -subunit secretion from epithelial ovarian tumors includes both epithelial and stromal cells, as indicated by immunostaining and in situ hybridization studies, although the main source remains disputed (19). A possible role of the -inhibin subunit in ovarian cancer emerged from studies with knockout mice in which the -inhibin subunit gene was inactivated. The absence of the -subunit gene abolishes both inhibin A and inhibin B and also monomeric and precursor forms of the -subunit. With this genotype, male and female mice develop mixed or incompletely differentiated gonadal stromal tumors (20). Therefore, inhibin- was proposed as a tumor suppressor gene. The mechanisms of tumor suppression by inhibin have not been fully characterized but may include inhibition of p27 protein synthesis, similarly to TGF-ßs, and increase of cyclin D2, which is expressed at high levels in some ovarian tumors and is also slightly increased in gonadal tumors expressing -inhibin (21, 22, 23).
The present evidence that serum total inhibin levels decreased after surgery in women with epithelial ovarian cancer and increased in all patients when tumors resurged also support a specific secretion of total inhibin from ovarian cancer. The simple concept that all relapses of ovarian cancers are recurrences of the same disease is contradicted by findings that only 23% of all late recurrences (>2 yr) have the same molecular genetic fingerprint. Thus, they may be considered as second primary tumors (24). Nevertheless, the inhibin molecules detected by the new ELISA are produced by such a high percentage of epithelial ovarian cancers that it is tempting to propose this test for postoperative surveillance. In the case series presented here, the new total inhibin assay detected all relapses, regardless of the length of disease-free intervals. Thus, further studies with statistical validation should assess the feasibility and cost-effectiveness of adopting this inhibin assay in the postoperative monitoring of patients treated for epithelial ovarian cancers.
Overall, several questions on ovarian cancer screening, follow-up, and after-care remain unsolved. Due to the relatively low incidence of epithelial ovarian cancers in the general population, these questions will require multicenter collaborations with a large number of subjects to be answered. Previous observations showing that CA-125 is effective for detecting the majority of epithelial cancers, whereas inhibin tests primarily detect mucinous carcinomas and granulosa cell tumors led to the suggestion that the combination of CA-125 and inhibin assays may provide a means of monitoring all ovarian cancers (12). The combination of total inhibin ELISA and CA-125, having shown optimal figures of sensitivity and specificity and being practical, fast, and commercially available methods, may perhaps deserve such an effort to be validated as markers for the screening of asymptomatic women.
In conclusion, the present study showed that: 1) serum concentrations of total inhibin in postmenopausal women with epithelial ovarian cancers are significantly higher than those of healthy postmenopausal women; 2) concentrations of total inhibin in women with epithelial ovarian cancers decrease significantly after surgery; 3) the combination of CA-125 and total inhibin yields sensitivity and specificity indexes of clinical relevance; and 4) total inhibin concentrations increase with the recurrence of mucinous tumors, thus providing a new candidate marker to monitor disease-free intervals.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose, and they have no financial connection with total inhibin manufacturers. The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, and approval of the manuscript.
First Published Online May 1, 2007
Abbreviations: aa, Amino acid; C, 133- to 134-aa C-terminal segment; ROC, receiver operator characteristic.
Received January 31, 2007.
Accepted April 20, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-subunit ELISA in serum from postmenopausal women with ovarian cancer. J Clin Endocrinol Metab 87:816–824 is a tumour suppressor gene with gonadal specificity in mice. Nature 360:313–319[CrossRef][Medline]Read all eLetters
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
