- - - - - - - - - - - - - - SYMPOSIUM ARTICLE - -
Inhibin as a marker for granulosa cell tumor RICHARD
E.
LApPOHN, HENRY
G.
BURGERI, JOKE BoUMA, MOHAN BANGAH\ AND MINDERT KRANS
From the Department of Obstetrics and Gynecology, Groningen State University Hospital, Groningen, The Netherlands, and the 'Medical Research Centre, Prince Henry's Hospital, Melbourne, Australia
Acta Obstet Gynecol Scand 1992; 71 Suppl 155: 61-65
In order to determine whether serum-immunoreactive inhibin could constitute a biochemical marker for the presence and progression of ovarian granulosa cell tumors and their metastases. we measured immunoreactive inhibin concentrations in series of serum samples obtained from 8 patients with granulosa cell tumor. Six series were tested in retrospect. From these. three came from patients who had been treated with an abdominal hysterectomy and bilateral salpingo-oophorectomy. In the 2 patients with residual or recurrent disease. inhibin was elevated. 4 and 20 months respectively before clinical manifestations of recurrence became evident; it reflected the effects of secondary therapy. Inhibin remained undetectable in one patient who was free of disease during 1I years of follow-up. Inhibin concentrations were also inappropriately increased in 2 of 3 women with amenorrhea and infertility resulting from small granulosa cell tumors. After removal. inhibin concentrations became normal and fertility resumed. Fertility also returned in the third patient. There was a significant negative correlation between the serum inhibin and FSH concentrations. consistent with autonomous production of inhibin by granulosa cell tumors. It is concluded that granulosa cell tumors have the capacity to produce inhibin. In retrospect. inhibin proved to be a marker for both primary and also recurrent and residual disease.
Granulosa cell tumors (GCf) are found in 5-10% of ovarian cancer cases. They originate from specialized ovarian mesenchyme. the stroma. and often betray their presence by synthesizing estrogens. Though generally not especially aggressive. the poor prognosis for GCf patients is correlated to the presence of extra-ovarian spread. age over 40, bilateral tumor. large volume, and histological characteristics such as solid growth and a high mitotic index. Primary treatment is surgical. Since tumor spread is predominantly in continuity to the intraperitoneal organs. complete surgical removal can be very difficult. Approximately 80% of patients with extraovarian spread die of recurrences (I, 2, 3). In one report the actuarial survival at 20 years was only 34% (4). A long-term follow-up is necessary in view of the protracted nature of the disease. With the recent advent of successful combination
chemotherapy for patients with recurrent or metastatic GCf (5, 6), it became important to have a circulating marker as an early indicator of recurrent disease. Circulating estradiol, classically a specific product of the granulosa cells, is unsuitable as a marker, since at least a third of the primary tumors and most extra-ovarian recurrences do not produce this hormone. Granulosa cells (and Sertoli cells in the male) produce a group of peptide hormones, the inhibins, which selectively inhibit the release of Follicle Stimulating Hormone (FSH) from the pituitary gland (7. 8). The production of these hormones is governed by FSH (9). It was of interest to investigate the possibility that immunoreactive in hibin might be produced by GCf and might constitute an early marker of recurrent disease.
Acta Obstet Gyn ecol Scand Suppl J55
62
R. E. Lappohn et af.
Materials and methods Patients By the end of 1988, the oncological serum bank of Groningen State University's Department of Obstetrics and Gynecology contained sera from 7 patients with a GCf. One case was represented by a single sample and was excluded from the investigation. From 3 patients, who had been-seen in consultation and remained under the care of their local gynecologist, only pre- and postoperative samples were available. These 3 patients had secondary amenorrhea and infertility. They had been referred because of resistance to various kinds of treatment to induce ovulation. The other 3 patients had been referred after primary removal of large GCf at other Departments. In the first patient, a then 55-year-old woman whose menopause had occurred at age 40, laparotomy after admission in shock in April, 1977 had revealed a ruptured (stage lIB) GCf, which was removed with the other adnexa and the uterus. Postoperatively, the abdomen was irradiated with 50.5 Gy. She subsequently received four courses of polychemotherapy (cyclophosphamide, actinomycin D, and cisplatin (CAP». During 12 years of follow-up. no evidence of recurrent disease was found. The second patient was born in 1917 and had her menopause at age 44. She was admitted with a first vaginal recurrence 13 months after primary removal of a large (500 g) GCf of the left ovary (stage LA) in 1977. In the course of the 8 years after primary surgery, three vaginal and three intra-abdominal recurrences were treated surgically; after detection of the first abdominal recurrence, radiotherapy was applied without effect and CAP courses were not tolerated. The patient died 15 months after the detection of liver metastases. The third patient, born in 1965, was admitted elsewhere in August, 1983, because of abdominal distension, pain, anorexia and breakthrough bleeding during the use of oral contraceptives. Immediately after the removal of a 2,800 g tumor of the right ovary she was transferred to our Department. A juvenile grade III-IV GCf was diagnosed and in January, 1984 the uterus and right adnexa were removed. Since there was no visible tumor, the original process was classified as stage I. Estrogen was prescribed. In October, 1986, the patient complained of pain in the upper abdomen. A large retroperitoneal recurrence was partially removed. The patient received radiotherapy (22 and 20 Gy locally) and at present she has no complaints. After discovery of small solid ovarian tumors (in 2 of them despite high estradiol and very low FSH Acta Obstet Gynecol Scand Suppl/55
concentrations), laparotomy in the 3 patients with infertility revealed the presence of unilateral small GCf which were treated by simple adnexectomy and inspection by bisection of the contralateral ovary. Since 1989, 2 more patients with advanced GCf have been referred to our Department. Patient no. 7, aged 29 at admission in December, 1989, was known to have a juvenile GCf diagnosed at age 26, which had been treated with an abdominal hysterectomy and bilateral adnexectomy. At laparotomy she was found to have extensive tumor growth throughout the abdomen, necessitating colonic diversion. Debulking was tried, but could not be achieved. The patient died before chemotherapy could be started. Patient no. 8, aged 49, had had an abdominal hysterectomy and right adnexectomy at age 33. She underwent an uretero-ileo-cutaneostomy after an unsuccessful urinary incontinence operation in 1979. In November, 1990, she complained of abdominal pain, severe weight loss and loss of appetite. A solid left adnexal tumor was felt and at laparotomy in December, 1990 a 797 g GCf of the left ovary was removed. The resection margins were not free of tumor. In February, 1991, polychemotherapy was started with Carboplatin, 5 FU, and actO. The three courses were completed in April, 1991.
Laboratory methods All samples were measured in duplicate. Inhibin, Luteinizing Hormone (LH). FSH and estradiol in samples from the bank were measured in the same assay; prospective samples were determined separately. Inhibin RIAs. One of the first descriptions of a reliable radio-immunoassay for inhibin came from Melbourne (10), and the inhibin determinations in sera from the serum bank were performed there. Serum inhibin was measured in 200 IJ.I samples by a RIA, described previously (10, 11). In the absence of ovaries and after the menopause, inhibin cannot be detected; in the early follicular phase of the menstrual cycle (day -12 from the LH peak), the mean serum inhibin concentration (± SEM) (n = 20) is 314 ± 35 U/I; normal mid-follicular phase (day -7) values are 379 ± 45.1 (n = 24) and on the day of the LH peak the mean ± SEM in 31 women is 772 ± 37.9. Higher values are found in the luteal phase (day +7): 1666 ± 90.9 (n = 31). For the prospective use of inhibin, a commercially available two-site immunoassay developed by Medgenix, Fleurus, Belgium, was used. This assay uses a goat polyclonal antibody for capture and a mouse monoclonal directed a two sites of the a-subunit of
Inhibin, marker for granulosa cell tumor inhibin conjugated to peroxidase as a tracer. The assay is specific, but measures higher molecular weight immunoreactive-forms of inhibin. Betweenassay coefficients of variation range between 4.1 and 8.2%. The assay sensitivity is reported to be 0.4 VlmI. Normal follicular phase values vary between 0.44 and 1.02 U/ml; periovulatory and mid-luteal values are up to 2.21 and 2.42 Ulml, respectively (II).
Gonadotropin assays. Serum LH and FSH levels were determined by IRMA, using the Serono MAIA® clone kits tor LH and FSH. Values are expressed in terms of lRP 68/40 for LH and the 2nd IRP 78/549 FSH immunoassay standards. The detection limit for both gonadotropins is 0.3 UII. The within-assay coefficients of variation are 1.7-4.7% (for high and low concentrations of LH) and 1.5 -9.9% for FSH, respectively. Normal values for follicular phase serum LH values are 3-12 UII; for FSH, 3-10 UII. In the first 10 years after the menopause, serum FSH is over 50 UII. Estradiol RIA. Plasma Estradiol was measured by RIA as previously described (12), with minor modifications. The assay sensitivity is 0.02 nmolll. Precision for estradiol values up to 0.37 nmolll is 12%, for higher values, 8%. Cross-reactivity with other estrogens is < 2%.
Results During 11 years of follow-up, patient no. 1 always had normal postmenopausal gonadotropin values. Inhibin was undetectable in all samples. The first sample in the serum bank from patient 2 was taken before her second admission. Her inhibin value was inappropriately elevated (740 UII); FSH and estradiol were low, with 21 UII and 0.07 nmolll; 5 months after the operation, postmenopausal values were found. Serum inhibin was now below the detection limit. Sera obtained during the following 2 years, when the patient had no complaints and findings at physical examination were unremarkable, showed normal FSH and estradiol values; inhibin was not detectable. In June, 1980, 3 months before an abdominal recurrence was found, inhibin was 900 VII. Inhibin levels did not react to irradiation with 55 Gy and in October, 1980a second recurrence had to be excised. In the subsequent serum, FSH was in the postmenopausal range but inhibin was 130 UII. After a third recurrence (May 1981) three CAP courses were given, but the therapy was not completed because of leukopenia. Seven months later a 7 x 10 em tumor was partially removed from behind the colon
63
ascendens. Again, FSH rose to postmenopausal levels but inhibin remained high at 620 1AI1. Thirteen months later (March 1983) another vaginal recurrence was surgically ablated. From that time onward, plasma estradiol concentrations started to increase and inhibin concentrations rose to over 2,000 UII. Her condition deteriorated and in December, 1983 liver metastases were found. In February, 1984 an ileocolonic resection was performed. The patient died in March, 1985. The first serum bank sample from patient no. 3 was drawn on the 5th postoperative day, when she was using oral contraceptives again. Inhibin was undetectable and FSH was low, probably as a result of the steroid medication. After removal of the uterus and right adnexa, the original process was classified as stage I. Although there was no clinical evidence of disease for 32 months, inappropriately elevated inhibin concentrations were found in the sera samples taken at one year (1,800 UII) and at 23 months (1,340 UII) after the final operation. FSH and estradiol were low. At 33 months, the patient complained of pain in the upper abdomen. On admission for relaparotomy the inhibin concentration was 2,860 UII. A large retroperitoneal recurrence was partially removed, but the serum inhibin concentration fell only slightly. She then received two courses of radiotherapy, 22 and 20 Gy locally, since when the inhibin levels have remained undetectable. At present the patient has no complaints. Two of the 3 patients with infertility had inappropriately increased inhibin concentrations in the preoperative samples. In the third patient, the GCT was the size of a normal corpus luteum. Normal inhibin values were obtained in the postoperative phase and all 3 conceived within a few menstrual cycles. During the follow-up of 4-12 years there have been no signs of recurrence. In patient no. 7, who died of her recurrences, slightly increased inhibin concentrations were found in the Melbourne assay: 161 and 201 UII. Inhibin was just detectable (0.24 and 0.29 U/ml) in the Medgenix assay. However, the tumor was strongly positive after immunostaining with inhibin antibody. After the failed debulking operation, inhibin fell slightly in the Melbourne assay (117 U/ml), but remained unchanged in the other 0.31 U/ml). The histology of the tumor removed from patient no. 8 was uncertain, but after immunostaining with inhibin antibody the diagnosis of GCT became clear. The initial hormone values were also typical: the gonadotropins were completely suppressed and estradiol was high, at 1.22 nrnol/l. Testosterone and 17-hydroxyprogesterone concentrations, both measured by ria, were high: 6.4 and 3.7 nmolll, respectively. Preoperative inhibin concentrations were ACID Obstet Gynecol Sctuul Supp/ 155
64
R. E. Lappohn et al.
22.2 and 24.5 Vlml. After the operation the gonadotropins rose to postmenopausal levels, the steroids disappeared and inhibin fell to 0.14 Vlml. Just before chemotherapy, inhibin had risen again to 1.69 Vlml; it remained detectable after the third course.
Discussion Despite the relative rarity of GCf (the incidence in Sweden has been reported as 1.7 per 100,000 women per year (13), the grave prognosis in patients with extra-ovarian spread (14) and especially in those with recurrences (15) warrants a vigilant follow-up in order to treat residual and recurrent tumor at the earliest possible moment. This is even more true now that advanced and recurrent disease appears to be responsive to combined treatment with cisplatin, vinblastine, and bleomycin (EORTC protocol 5504-3) (16). Although the use of plasma estradiol to monitor the response to polychemotherapy has been described (17), estradiol is low in too many cases to be generally useful as a marker. Three of our patients with recurrences after bilateral ovariectomy and one with primary tumor have demonstrated the insufficiency of estradiol as a marker for GCf. In contrast, the inhibin concentration was inappropriately increased in all but one (with a very small GCf) of our patients, many months before clinical signs appeared and action was taken. Inhibin is a gonadal protein belonging to a family of substances that modulate FSH secretion by the pituitary. Inhibin A and Bare FSH suppressors; they are heterodimers of a 26 kD a-sub-unit and 14 kD 13 (A and B) sub-units. The inhibins are specific products of the granulosa cell, but their presence was also described in an ovarian Sertoli-Leydig tumor (18). They are closely related in different animal species, which indicates a high degree of conservation (8). This suggests that tumor cells may easily retain the capacity to produce inhibin. Inhibin is undetectable in otherwise normal ovariectomized women; the presence of inhibin and relatively low FSH levels as demonstrated in all our samples from ovariectomized women with a growing granulosa cell tumor confirms the latter's capacity to secrete inhibin. Inhibin was undetectable and FSH was at postmenopausal levels in the sera from one ovariectomized woman who was apparently free of disease during 11 years of follow-up. Inhibin values above the normal range and low FSH levels were found in preoperative sera from 2 of the 3 women with small stage I tumors; however, normal levels were encountered in the patient with a Acta Obstet Gynecol Scand Suppf 155
very small tumor. Thus, inhibin levels are apparently correlated to tumor mass and, in the presence of physiological production, inhibin results are more difficult to interpret. Values below the detection limit after the first surgical excisions in patients 2 and 3 probably imply a near-complete removal. Inhibin reappeared in the serum of these 2 patients long before there was clinical suspicion of tumor growth, giving a lead time of at least 5 and 22 months. Furthermore, in patient no. 2, the high inhibin levels closely reflected the failure of radiotherapy, the inability to surgically remove recurrences and the incompleteness of polychemotherapy. Likewise, the persistent presence of a low concentration of inhibin in patient no. 8, who received less than optimal therapy because of her restricted renal function, is an ominous sign. With one exception, the highest inhibin concentrations were found in sera with the lowest FSH. There is a significant rank-order correlation (Spearman r = 0.577, P < 0.(01) between the two hormones. Low FSH levels in cases of granulosa cell tumor have been reported by others (14). Although there is firm evidence that under physiological circumstances, during the follicular phase of the menstrual cycle, the production of inhibin from the granulosa cell is governed by FSH, these data are consistent with the autonomous production of inhibin by granulosa cell tumors, with consequent FSH suppression. Our findings provide support for the use of inhibin as a biochemical marker for GCf. Moreover, evidence is presented for the autonomous secretion of this hormone. Inhibin antibodies may be useful for the histological recognition of GCf. The inverse correlation between serum FSH and inhibin concentrations suggests that serial determination of FSH may also serve as a marker. However, FSH concentrations in postmenopausal and castrated women may vary over a wide range and also depend on the presence of physiological sources of estrogen production and the use of hormonal replacement therapy. Therefore, specificity and sensitivity of FSH determinations may be too low. One of the very few successes in ovarian carcinoma therapy has been the development of special markers for rare tumors, e.g. u-fetoprotein for yolk sac tumors of the ovary. The availability of these markers, together with the effective present-day chemotherapy, has greatly improved the outlook for patients with these malignancies. We expect that shortly the same may be said for the somewhat less rare advanced stages of ovarian granulosa cell tumor.
Inhibin, marker for granulosa cell tumor
Acknowledgement H. G. Burger and M. Bangah are supported by the National Health and Medical Research Council of Australia.
References I. Schwartz PE. Smith JP. Treatment of ovarian stromal tumors. Am J Obstet Gynecol 1976; 125: 402-11. 2. Stenwig JT. Hazekamp JTHE. Beecham lB. Granulosa cell tumors of the ovary. A clinico-pathological study of 118 cases with long term follow-up. Gynecol Oncol 1979; 7: 136-52. 3. Evans AT, Gotfey TA. Malkasian GD, Annegers JE A c1inico-pathologic review of 118 granulosa and 882 theca cell tumors. Obstet Gynecol 1980; 55: 231-38. 4. Dempster J. Geirsson RT, Duncan 10. Survival after ovarian granulosa and theca cell tumors. Scott Med J 1987; 32: 38-39. 5. Jacobs AJ, Deppe G. Cohen CJ. Combination chemotherapy of ovarian granulosa cell tumor with cis-platinum and doxorubicin. Gynecol Oncol 1982; 14: 294-97. 6. Colombo N. Sessa C. Landoni F. Sartori E. Pecorelli S. Mangioni C. Cisplatin, vinblastine, and bleomycin combination chemotherapy in metastatic granulosa cell tumor of the ovary. Obstet Gynecol 1986; 67: 265-68. 7. McLachlan RI, Robertson OM, de Kretser OM, Burger HG. Advances in the physiology of inhibin and inhibin-related peptides. Clin Endocrinol 1988; 29: 77-92. 8. Ying SY. Inhibins, activins, and follistatins: Gonadal proteins modulating the secretion of follicle stimulating hormone. Endocr Rev 1988; 9: 267-93. 9. Buckler HM. McLachlan RI, MacLachlan VB, Healy DL. Burger HG. Serum inhibin levels in polycystic ovary syndrome: basal levels and response to luteinizing hormone releasing hormone agonist and exogenous gonadotropin administration. J Clin Endocrin Metab 1988; 66: 798-8113. Ill. McLachlan RI. Robertson OM, Healy DL, Burger HG, De Kretser OM. Circulating immunoreactive in-
II.
12.
13.
14.
15.
16.
17.
18.
65
hibin levels during the normal menstrual cycle. J Clin Endocrinol Metab 1987: 65: 954-61. Ponce let E, et al: Serum levels of immunoreactive inhibin throughout the human menstrual cycle. Abstract II. 2, Inhibin symposium, Brussels, 28 and 29 September, 1990. Jurjens H, Pratt 11, Woldring MG. Estradiol determination without extraction and purification. J Clin Endocrinol Metab 1975; 40: 19-26. Stenwig JT, Hazekamp JT. Beecham JB. Granulosa cell tumors of the ovary: a clinicopathological study of 118 cases with long-term follow-up. Gynecol Oncol 1979; 7: 136-52. Bjtirkholm E, Petterson E Granulosa-cell and thecacell tumors: the clinical picture and long-term outcome for the Radiumhemmet series. Acta Obstet Gynecol Scand 1980: 59: 361-65. Evans AT, Gaffey TA, Malkasian GO, Annegers JE Clinico-pathologic review of 118 granulosa and 82 theca cell tumors. Obstet Gynecol 1980; 55: 231-38. Pecorelli S, Wagener P, Bonazzi C, et al. Cisplatin (C). vinblastine (V), and bleomycin (B) combination chemotherapy in recurrent or advanced granulosa cell tumor of the ovary (GCfO): an EORTC gynecologic cancer cooperative group study. Proc Am Soc Clin Oncol 1988; 7: 147, abstract. Kaye SB, Davies E. Cyclophosphamide, adriamycin, and cis-platinum for the treatment of advanced granulosa cell tumors, using serum estradiol as a tumor marker. Gynecol Oncol 1986; 24: 261-64. Ohashi M, Hasegawa Y. Haji M. et al. Production of immunoreactive inhibin by a virilizing ovarian tumour (Sertoli-Leydig tumour). Clin Endocrinol 1990; 33: 613-18.
Address for correspondence: R. E. Lappohn, M.D. Dept of Obstetrics and Gynecology Academisch Ziekenhuis P.O. Box 30.001 NL-97oo RB Groningen The Netherlands
Acta Obstet Gynecol Scand Suppl/55
Inhibin as a marker for granulosa cell tumor RICHARD
E.
LApPOHN, HENRY
G.
BURGERI, JOKE BoUMA, MOHAN BANGAH\ AND MINDERT KRANS
From the Department of Obstetrics and Gynecology, Groningen State University Hospital, Groningen, The Netherlands, and the 'Medical Research Centre, Prince Henry's Hospital, Melbourne, Australia
Acta Obstet Gynecol Scand 1992; 71 Suppl 155: 61-65
In order to determine whether serum-immunoreactive inhibin could constitute a biochemical marker for the presence and progression of ovarian granulosa cell tumors and their metastases. we measured immunoreactive inhibin concentrations in series of serum samples obtained from 8 patients with granulosa cell tumor. Six series were tested in retrospect. From these. three came from patients who had been treated with an abdominal hysterectomy and bilateral salpingo-oophorectomy. In the 2 patients with residual or recurrent disease. inhibin was elevated. 4 and 20 months respectively before clinical manifestations of recurrence became evident; it reflected the effects of secondary therapy. Inhibin remained undetectable in one patient who was free of disease during 1I years of follow-up. Inhibin concentrations were also inappropriately increased in 2 of 3 women with amenorrhea and infertility resulting from small granulosa cell tumors. After removal. inhibin concentrations became normal and fertility resumed. Fertility also returned in the third patient. There was a significant negative correlation between the serum inhibin and FSH concentrations. consistent with autonomous production of inhibin by granulosa cell tumors. It is concluded that granulosa cell tumors have the capacity to produce inhibin. In retrospect. inhibin proved to be a marker for both primary and also recurrent and residual disease.
Granulosa cell tumors (GCf) are found in 5-10% of ovarian cancer cases. They originate from specialized ovarian mesenchyme. the stroma. and often betray their presence by synthesizing estrogens. Though generally not especially aggressive. the poor prognosis for GCf patients is correlated to the presence of extra-ovarian spread. age over 40, bilateral tumor. large volume, and histological characteristics such as solid growth and a high mitotic index. Primary treatment is surgical. Since tumor spread is predominantly in continuity to the intraperitoneal organs. complete surgical removal can be very difficult. Approximately 80% of patients with extraovarian spread die of recurrences (I, 2, 3). In one report the actuarial survival at 20 years was only 34% (4). A long-term follow-up is necessary in view of the protracted nature of the disease. With the recent advent of successful combination
chemotherapy for patients with recurrent or metastatic GCf (5, 6), it became important to have a circulating marker as an early indicator of recurrent disease. Circulating estradiol, classically a specific product of the granulosa cells, is unsuitable as a marker, since at least a third of the primary tumors and most extra-ovarian recurrences do not produce this hormone. Granulosa cells (and Sertoli cells in the male) produce a group of peptide hormones, the inhibins, which selectively inhibit the release of Follicle Stimulating Hormone (FSH) from the pituitary gland (7. 8). The production of these hormones is governed by FSH (9). It was of interest to investigate the possibility that immunoreactive in hibin might be produced by GCf and might constitute an early marker of recurrent disease.
Acta Obstet Gyn ecol Scand Suppl J55
62
R. E. Lappohn et af.
Materials and methods Patients By the end of 1988, the oncological serum bank of Groningen State University's Department of Obstetrics and Gynecology contained sera from 7 patients with a GCf. One case was represented by a single sample and was excluded from the investigation. From 3 patients, who had been-seen in consultation and remained under the care of their local gynecologist, only pre- and postoperative samples were available. These 3 patients had secondary amenorrhea and infertility. They had been referred because of resistance to various kinds of treatment to induce ovulation. The other 3 patients had been referred after primary removal of large GCf at other Departments. In the first patient, a then 55-year-old woman whose menopause had occurred at age 40, laparotomy after admission in shock in April, 1977 had revealed a ruptured (stage lIB) GCf, which was removed with the other adnexa and the uterus. Postoperatively, the abdomen was irradiated with 50.5 Gy. She subsequently received four courses of polychemotherapy (cyclophosphamide, actinomycin D, and cisplatin (CAP». During 12 years of follow-up. no evidence of recurrent disease was found. The second patient was born in 1917 and had her menopause at age 44. She was admitted with a first vaginal recurrence 13 months after primary removal of a large (500 g) GCf of the left ovary (stage LA) in 1977. In the course of the 8 years after primary surgery, three vaginal and three intra-abdominal recurrences were treated surgically; after detection of the first abdominal recurrence, radiotherapy was applied without effect and CAP courses were not tolerated. The patient died 15 months after the detection of liver metastases. The third patient, born in 1965, was admitted elsewhere in August, 1983, because of abdominal distension, pain, anorexia and breakthrough bleeding during the use of oral contraceptives. Immediately after the removal of a 2,800 g tumor of the right ovary she was transferred to our Department. A juvenile grade III-IV GCf was diagnosed and in January, 1984 the uterus and right adnexa were removed. Since there was no visible tumor, the original process was classified as stage I. Estrogen was prescribed. In October, 1986, the patient complained of pain in the upper abdomen. A large retroperitoneal recurrence was partially removed. The patient received radiotherapy (22 and 20 Gy locally) and at present she has no complaints. After discovery of small solid ovarian tumors (in 2 of them despite high estradiol and very low FSH Acta Obstet Gynecol Scand Suppl/55
concentrations), laparotomy in the 3 patients with infertility revealed the presence of unilateral small GCf which were treated by simple adnexectomy and inspection by bisection of the contralateral ovary. Since 1989, 2 more patients with advanced GCf have been referred to our Department. Patient no. 7, aged 29 at admission in December, 1989, was known to have a juvenile GCf diagnosed at age 26, which had been treated with an abdominal hysterectomy and bilateral adnexectomy. At laparotomy she was found to have extensive tumor growth throughout the abdomen, necessitating colonic diversion. Debulking was tried, but could not be achieved. The patient died before chemotherapy could be started. Patient no. 8, aged 49, had had an abdominal hysterectomy and right adnexectomy at age 33. She underwent an uretero-ileo-cutaneostomy after an unsuccessful urinary incontinence operation in 1979. In November, 1990, she complained of abdominal pain, severe weight loss and loss of appetite. A solid left adnexal tumor was felt and at laparotomy in December, 1990 a 797 g GCf of the left ovary was removed. The resection margins were not free of tumor. In February, 1991, polychemotherapy was started with Carboplatin, 5 FU, and actO. The three courses were completed in April, 1991.
Laboratory methods All samples were measured in duplicate. Inhibin, Luteinizing Hormone (LH). FSH and estradiol in samples from the bank were measured in the same assay; prospective samples were determined separately. Inhibin RIAs. One of the first descriptions of a reliable radio-immunoassay for inhibin came from Melbourne (10), and the inhibin determinations in sera from the serum bank were performed there. Serum inhibin was measured in 200 IJ.I samples by a RIA, described previously (10, 11). In the absence of ovaries and after the menopause, inhibin cannot be detected; in the early follicular phase of the menstrual cycle (day -12 from the LH peak), the mean serum inhibin concentration (± SEM) (n = 20) is 314 ± 35 U/I; normal mid-follicular phase (day -7) values are 379 ± 45.1 (n = 24) and on the day of the LH peak the mean ± SEM in 31 women is 772 ± 37.9. Higher values are found in the luteal phase (day +7): 1666 ± 90.9 (n = 31). For the prospective use of inhibin, a commercially available two-site immunoassay developed by Medgenix, Fleurus, Belgium, was used. This assay uses a goat polyclonal antibody for capture and a mouse monoclonal directed a two sites of the a-subunit of
Inhibin, marker for granulosa cell tumor inhibin conjugated to peroxidase as a tracer. The assay is specific, but measures higher molecular weight immunoreactive-forms of inhibin. Betweenassay coefficients of variation range between 4.1 and 8.2%. The assay sensitivity is reported to be 0.4 VlmI. Normal follicular phase values vary between 0.44 and 1.02 U/ml; periovulatory and mid-luteal values are up to 2.21 and 2.42 Ulml, respectively (II).
Gonadotropin assays. Serum LH and FSH levels were determined by IRMA, using the Serono MAIA® clone kits tor LH and FSH. Values are expressed in terms of lRP 68/40 for LH and the 2nd IRP 78/549 FSH immunoassay standards. The detection limit for both gonadotropins is 0.3 UII. The within-assay coefficients of variation are 1.7-4.7% (for high and low concentrations of LH) and 1.5 -9.9% for FSH, respectively. Normal values for follicular phase serum LH values are 3-12 UII; for FSH, 3-10 UII. In the first 10 years after the menopause, serum FSH is over 50 UII. Estradiol RIA. Plasma Estradiol was measured by RIA as previously described (12), with minor modifications. The assay sensitivity is 0.02 nmolll. Precision for estradiol values up to 0.37 nmolll is 12%, for higher values, 8%. Cross-reactivity with other estrogens is < 2%.
Results During 11 years of follow-up, patient no. 1 always had normal postmenopausal gonadotropin values. Inhibin was undetectable in all samples. The first sample in the serum bank from patient 2 was taken before her second admission. Her inhibin value was inappropriately elevated (740 UII); FSH and estradiol were low, with 21 UII and 0.07 nmolll; 5 months after the operation, postmenopausal values were found. Serum inhibin was now below the detection limit. Sera obtained during the following 2 years, when the patient had no complaints and findings at physical examination were unremarkable, showed normal FSH and estradiol values; inhibin was not detectable. In June, 1980, 3 months before an abdominal recurrence was found, inhibin was 900 VII. Inhibin levels did not react to irradiation with 55 Gy and in October, 1980a second recurrence had to be excised. In the subsequent serum, FSH was in the postmenopausal range but inhibin was 130 UII. After a third recurrence (May 1981) three CAP courses were given, but the therapy was not completed because of leukopenia. Seven months later a 7 x 10 em tumor was partially removed from behind the colon
63
ascendens. Again, FSH rose to postmenopausal levels but inhibin remained high at 620 1AI1. Thirteen months later (March 1983) another vaginal recurrence was surgically ablated. From that time onward, plasma estradiol concentrations started to increase and inhibin concentrations rose to over 2,000 UII. Her condition deteriorated and in December, 1983 liver metastases were found. In February, 1984 an ileocolonic resection was performed. The patient died in March, 1985. The first serum bank sample from patient no. 3 was drawn on the 5th postoperative day, when she was using oral contraceptives again. Inhibin was undetectable and FSH was low, probably as a result of the steroid medication. After removal of the uterus and right adnexa, the original process was classified as stage I. Although there was no clinical evidence of disease for 32 months, inappropriately elevated inhibin concentrations were found in the sera samples taken at one year (1,800 UII) and at 23 months (1,340 UII) after the final operation. FSH and estradiol were low. At 33 months, the patient complained of pain in the upper abdomen. On admission for relaparotomy the inhibin concentration was 2,860 UII. A large retroperitoneal recurrence was partially removed, but the serum inhibin concentration fell only slightly. She then received two courses of radiotherapy, 22 and 20 Gy locally, since when the inhibin levels have remained undetectable. At present the patient has no complaints. Two of the 3 patients with infertility had inappropriately increased inhibin concentrations in the preoperative samples. In the third patient, the GCT was the size of a normal corpus luteum. Normal inhibin values were obtained in the postoperative phase and all 3 conceived within a few menstrual cycles. During the follow-up of 4-12 years there have been no signs of recurrence. In patient no. 7, who died of her recurrences, slightly increased inhibin concentrations were found in the Melbourne assay: 161 and 201 UII. Inhibin was just detectable (0.24 and 0.29 U/ml) in the Medgenix assay. However, the tumor was strongly positive after immunostaining with inhibin antibody. After the failed debulking operation, inhibin fell slightly in the Melbourne assay (117 U/ml), but remained unchanged in the other 0.31 U/ml). The histology of the tumor removed from patient no. 8 was uncertain, but after immunostaining with inhibin antibody the diagnosis of GCT became clear. The initial hormone values were also typical: the gonadotropins were completely suppressed and estradiol was high, at 1.22 nrnol/l. Testosterone and 17-hydroxyprogesterone concentrations, both measured by ria, were high: 6.4 and 3.7 nmolll, respectively. Preoperative inhibin concentrations were ACID Obstet Gynecol Sctuul Supp/ 155
64
R. E. Lappohn et al.
22.2 and 24.5 Vlml. After the operation the gonadotropins rose to postmenopausal levels, the steroids disappeared and inhibin fell to 0.14 Vlml. Just before chemotherapy, inhibin had risen again to 1.69 Vlml; it remained detectable after the third course.
Discussion Despite the relative rarity of GCf (the incidence in Sweden has been reported as 1.7 per 100,000 women per year (13), the grave prognosis in patients with extra-ovarian spread (14) and especially in those with recurrences (15) warrants a vigilant follow-up in order to treat residual and recurrent tumor at the earliest possible moment. This is even more true now that advanced and recurrent disease appears to be responsive to combined treatment with cisplatin, vinblastine, and bleomycin (EORTC protocol 5504-3) (16). Although the use of plasma estradiol to monitor the response to polychemotherapy has been described (17), estradiol is low in too many cases to be generally useful as a marker. Three of our patients with recurrences after bilateral ovariectomy and one with primary tumor have demonstrated the insufficiency of estradiol as a marker for GCf. In contrast, the inhibin concentration was inappropriately increased in all but one (with a very small GCf) of our patients, many months before clinical signs appeared and action was taken. Inhibin is a gonadal protein belonging to a family of substances that modulate FSH secretion by the pituitary. Inhibin A and Bare FSH suppressors; they are heterodimers of a 26 kD a-sub-unit and 14 kD 13 (A and B) sub-units. The inhibins are specific products of the granulosa cell, but their presence was also described in an ovarian Sertoli-Leydig tumor (18). They are closely related in different animal species, which indicates a high degree of conservation (8). This suggests that tumor cells may easily retain the capacity to produce inhibin. Inhibin is undetectable in otherwise normal ovariectomized women; the presence of inhibin and relatively low FSH levels as demonstrated in all our samples from ovariectomized women with a growing granulosa cell tumor confirms the latter's capacity to secrete inhibin. Inhibin was undetectable and FSH was at postmenopausal levels in the sera from one ovariectomized woman who was apparently free of disease during 11 years of follow-up. Inhibin values above the normal range and low FSH levels were found in preoperative sera from 2 of the 3 women with small stage I tumors; however, normal levels were encountered in the patient with a Acta Obstet Gynecol Scand Suppf 155
very small tumor. Thus, inhibin levels are apparently correlated to tumor mass and, in the presence of physiological production, inhibin results are more difficult to interpret. Values below the detection limit after the first surgical excisions in patients 2 and 3 probably imply a near-complete removal. Inhibin reappeared in the serum of these 2 patients long before there was clinical suspicion of tumor growth, giving a lead time of at least 5 and 22 months. Furthermore, in patient no. 2, the high inhibin levels closely reflected the failure of radiotherapy, the inability to surgically remove recurrences and the incompleteness of polychemotherapy. Likewise, the persistent presence of a low concentration of inhibin in patient no. 8, who received less than optimal therapy because of her restricted renal function, is an ominous sign. With one exception, the highest inhibin concentrations were found in sera with the lowest FSH. There is a significant rank-order correlation (Spearman r = 0.577, P < 0.(01) between the two hormones. Low FSH levels in cases of granulosa cell tumor have been reported by others (14). Although there is firm evidence that under physiological circumstances, during the follicular phase of the menstrual cycle, the production of inhibin from the granulosa cell is governed by FSH, these data are consistent with the autonomous production of inhibin by granulosa cell tumors, with consequent FSH suppression. Our findings provide support for the use of inhibin as a biochemical marker for GCf. Moreover, evidence is presented for the autonomous secretion of this hormone. Inhibin antibodies may be useful for the histological recognition of GCf. The inverse correlation between serum FSH and inhibin concentrations suggests that serial determination of FSH may also serve as a marker. However, FSH concentrations in postmenopausal and castrated women may vary over a wide range and also depend on the presence of physiological sources of estrogen production and the use of hormonal replacement therapy. Therefore, specificity and sensitivity of FSH determinations may be too low. One of the very few successes in ovarian carcinoma therapy has been the development of special markers for rare tumors, e.g. u-fetoprotein for yolk sac tumors of the ovary. The availability of these markers, together with the effective present-day chemotherapy, has greatly improved the outlook for patients with these malignancies. We expect that shortly the same may be said for the somewhat less rare advanced stages of ovarian granulosa cell tumor.
Inhibin, marker for granulosa cell tumor
Acknowledgement H. G. Burger and M. Bangah are supported by the National Health and Medical Research Council of Australia.
References I. Schwartz PE. Smith JP. Treatment of ovarian stromal tumors. Am J Obstet Gynecol 1976; 125: 402-11. 2. Stenwig JT. Hazekamp JTHE. Beecham lB. Granulosa cell tumors of the ovary. A clinico-pathological study of 118 cases with long term follow-up. Gynecol Oncol 1979; 7: 136-52. 3. Evans AT, Gotfey TA. Malkasian GD, Annegers JE A c1inico-pathologic review of 118 granulosa and 882 theca cell tumors. Obstet Gynecol 1980; 55: 231-38. 4. Dempster J. Geirsson RT, Duncan 10. Survival after ovarian granulosa and theca cell tumors. Scott Med J 1987; 32: 38-39. 5. Jacobs AJ, Deppe G. Cohen CJ. Combination chemotherapy of ovarian granulosa cell tumor with cis-platinum and doxorubicin. Gynecol Oncol 1982; 14: 294-97. 6. Colombo N. Sessa C. Landoni F. Sartori E. Pecorelli S. Mangioni C. Cisplatin, vinblastine, and bleomycin combination chemotherapy in metastatic granulosa cell tumor of the ovary. Obstet Gynecol 1986; 67: 265-68. 7. McLachlan RI, Robertson OM, de Kretser OM, Burger HG. Advances in the physiology of inhibin and inhibin-related peptides. Clin Endocrinol 1988; 29: 77-92. 8. Ying SY. Inhibins, activins, and follistatins: Gonadal proteins modulating the secretion of follicle stimulating hormone. Endocr Rev 1988; 9: 267-93. 9. Buckler HM. McLachlan RI, MacLachlan VB, Healy DL. Burger HG. Serum inhibin levels in polycystic ovary syndrome: basal levels and response to luteinizing hormone releasing hormone agonist and exogenous gonadotropin administration. J Clin Endocrin Metab 1988; 66: 798-8113. Ill. McLachlan RI. Robertson OM, Healy DL, Burger HG, De Kretser OM. Circulating immunoreactive in-
II.
12.
13.
14.
15.
16.
17.
18.
65
hibin levels during the normal menstrual cycle. J Clin Endocrinol Metab 1987: 65: 954-61. Ponce let E, et al: Serum levels of immunoreactive inhibin throughout the human menstrual cycle. Abstract II. 2, Inhibin symposium, Brussels, 28 and 29 September, 1990. Jurjens H, Pratt 11, Woldring MG. Estradiol determination without extraction and purification. J Clin Endocrinol Metab 1975; 40: 19-26. Stenwig JT, Hazekamp JT. Beecham JB. Granulosa cell tumors of the ovary: a clinicopathological study of 118 cases with long-term follow-up. Gynecol Oncol 1979; 7: 136-52. Bjtirkholm E, Petterson E Granulosa-cell and thecacell tumors: the clinical picture and long-term outcome for the Radiumhemmet series. Acta Obstet Gynecol Scand 1980: 59: 361-65. Evans AT, Gaffey TA, Malkasian GO, Annegers JE Clinico-pathologic review of 118 granulosa and 82 theca cell tumors. Obstet Gynecol 1980; 55: 231-38. Pecorelli S, Wagener P, Bonazzi C, et al. Cisplatin (C). vinblastine (V), and bleomycin (B) combination chemotherapy in recurrent or advanced granulosa cell tumor of the ovary (GCfO): an EORTC gynecologic cancer cooperative group study. Proc Am Soc Clin Oncol 1988; 7: 147, abstract. Kaye SB, Davies E. Cyclophosphamide, adriamycin, and cis-platinum for the treatment of advanced granulosa cell tumors, using serum estradiol as a tumor marker. Gynecol Oncol 1986; 24: 261-64. Ohashi M, Hasegawa Y. Haji M. et al. Production of immunoreactive inhibin by a virilizing ovarian tumour (Sertoli-Leydig tumour). Clin Endocrinol 1990; 33: 613-18.
Address for correspondence: R. E. Lappohn, M.D. Dept of Obstetrics and Gynecology Academisch Ziekenhuis P.O. Box 30.001 NL-97oo RB Groningen The Netherlands
Acta Obstet Gynecol Scand Suppl/55
