0013-7227/02/$15.00/0
Printed in U.S.A.
The Journal of Clinical Endocrinology & Metabolism 87(7):3074 –3077
Copyright © 2002 by The Endocrine Society
CLINICAL CASE SEMINAR
Intraoperative Testosterone Assay for Virilizing Ovarian
Tumor Topographic Assessment: Report of a Leydig Cell
Tumor of the Ovary in a Premenopausal Woman with an
Adrenal Incidentaloma
C. REGNIER, A. BENNET, D. MALET, T. GUEZ, M. PLANTAVID, P. ROCHAIX, X. MONROZIES,
J.-P. LOUVET, AND P. CARON
Department of Endocrinology (C.R., A.B., D.M., T.G., J.-P.L., P.C.), Centre Hospitalier Universitaire (CHU) Rangueil;
Departments of Clinical Biochemistry (M.P.) and Gynecology (X.M.), CHU La Grave; and Department of Histopathology
(P.R.), Institut Claudius Régaud, Toulouse, France
Ovarian virilizing tumors are rare and can lead to assessment
difficulties because of their small size. A 41-yr-old female was
referred for evaluation of hirsutism that had increased within
the previous 3 yr. Menstrual cycle length was normal. Plasma
testosterone was 3.9 ng/ml (normal range, 0.2– 0.8 ng/ml), was
not suppressible by 2 mg dexamethasone (4.3 ng/ml), and was
increased (6.3 ng/ml) after three daily injections of hCG (5000
IU). Abdominal computed tomography scan showed an adrenal nodule (13 ⴛ 6 mm) that remained unchanged after 3
months. Ultrasound examination of the pelvis was normal.
Ovarian and adrenal venous catheterization did not yield additional information. Topographic assessment was made by
V
IRILIZING TUMORS are rare; they develop from the
ovary or adrenal cortex and are observed more frequently in postmenopausal women. The diagnosis of ovarian virilizing tumors can be difficult in women of reproductive age, because the size of such tumors is often too
small to allow detection by ultrasound or radiological
examination (1).
We report the case of a virilizing ovarian tumor in a premenopausal woman. The diagnosis of the tumor location
was made using an intraoperative measurement of testosterone levels in blood samples taken directly from the ovarian veins, thereby leading to conservative surgery.
Case Report
A 41-yr-old female, without significant past medical history, was
referred to our Department of Endocrinology for evaluation of hirsutism. Her hirsutism had developed within the last 3 yr and had become
more severe within the last 6 months. The patient did not report any
alteration regarding her libido or length of her menstrual cycles. The
patient was taking no medication.
On physical examination the patient was found to have hirsutism on
the chin, breasts, and lower abdomen. Mild clitoral enlargement was
observed. No other clinical abnormality was found: the body mass index
was within the normal range (22 kg/m2), and blood pressure was 110/80
mm Hg. The patient had no specific symptoms of Cushing’s syndrome.
Abbreviations: CT, Computed tomography; DHEAS, dehydroepiandrosterone sulfate; MRI, magnetic resonance imaging.
intraoperative measurement of testosterone in the samples
taken from each ovarian vein (competitive chemiluminescent
immunoassay ADVIA Centaur; right ovarian vein, 105 ng/ml;
left ovarian vein, 5 ng/ml; peripheral blood, 7 ng/ml). Right
annexectomy resulted in normalization of testosterone levels
(0.22 ng/ml). Histopathological examination found a Leydig
cell tumor of hilar type (1.5 cm). This observation illustrates
the usefulness of intraoperative measurement of testosterone
by a rapid automated technique for topographic assessment of
ovarian virilizing tumor in premenopausal women. (J Clin
Endocrinol Metab 87: 3074 –3077, 2002)
The abdominal palpation and gynecological examination did not provide additional information.
The patient had high serum testosterone (3.9 ng/ml), which implied
the diagnosis of a virilizing tumor (1). Serum dehydroepiandrosterone
sulfate (DHEAS) was normal (92 ␮g/dl), which argued against an adrenal source of testosterone. The increase in testosterone levels during
hCG administration (5000 IU/d for 3 d) (2) was consistent with the
hypothesis of an ovarian virilizing tumor. Other hormonal data comprised normal levels of ⌬4-androstenedione, slightly elevated LH and
FSH levels (LH, 10 IU/liter; FSH, 9 IU/liter), SHBG levels at the lower
limit of the normal range (3.0 mg/liter; normal range in premenopausal
women, 3– 6 mg/liter), normal urinary 17-cetocorticosteroids (7.5 mg/d)
and 17-hydroxycorticosteroids (4.8 mg/d), normal urinary free cortisol
(38.5 ␮g/d), and normal suppression of urinary free cortisol by dexamethasone (2 mg/d for 2 d), ruling out Cushing’s syndrome (Table 1).
Radiological examinations comprised a computed tomography (CT)
scan, a magnetic resonance imaging (MRI) scan, and two transvaginal
ultrasound examinations of the ovaries, performed by two different
radiologists. The ultrasound examinations did not find any abnormality.
The abdominal CT scan revealed an adrenal mass of 13 ⫻ 6 mm. The
characteristics of the adrenal mass on CT scan [round shape, hypodensity (less than 12 Houndsfeld units), increased density after contrast
agent injection (30 Houndsfeld units)] suggested the diagnosis of adrenocortical adenoma. The aspect of the adrenal nodule was unchanged
on a CT scan performed 3 months later. The characteristics of the nodule
on MRI scan were also those of a benign adenoma; chemical shift sequences showed an isointense signal in phase and an hypointense signal
out of phase compared with the renal cortex. Iodomethylnorcholesterol
scanning performed during daily administration of 2 mg dexamethasone did not detect any abnormality. Further hormonal evaluation was
performed, which ruled out primary hypermineralocorticism and
pheochromocytoma.
3074
Regnier et al. • Clinical Case Seminar
J Clin Endocrinol Metab, July 2002, 87(7):3074 –3077 3075
Results
TABLE 1. Hormonal results during combined dexamethasonehCG test
1
2
3
4
5
An unequivocal gradient of testosterone concentrations
was found, localizing the source of testosterone secretion in
the right ovary (Table 2). The patient underwent right annexectomy. The ovary contained several cysts at various
maturation steps; the biggest cysts were 9 mm large. Macroscopic examination detected a 1.5-cm nodule in the ovarian
hilus. Microscopic examination showed a hilar, well circumscribed tumor composed of steroid-type cells with eosinophilic cytoplasm sometimes containing crystals of Reinke.
The cells were diffusely arranged, but frequently separated
in nests or cords by anuclear eosinophilic areas. Fibrinoid
replacement of the walls of moderate-sized vessels was frequent. Thus, the histopathological diagnosis was a Leydig
cell tumor of hilar type (Fig. 1).
Three days after right annexectomy, serum testosterone
was normal (0.22 ng/ml), and hirsutism progressively subsided after surgery. Thirty months after surgery, hyperandrogenism has not recurred, and the patient has not become
pregnant.
7
Discussion
5
6
7
Dexamethasone 2 mg/d
Day
hCG 5000 IU/d
Testosterone (ng/ml)
⌬-4-Androstene-dione (ng/ml)
DHEAS (␮g/dl)
Urinary free cortisol (␮g/d)
4.0
1.8
92
39
39
4.3
0.6
29
⬍7
6.0
0.8
21
⬍7
6.3
0.8
18
⬍7
6.0
1.0
17
⬍7
TABLE 2. Testosterone levels in venous samples taken during
preoperative catheterization of adrenal and ovarian veins (under
basal- and hCG-stimulated conditions) and during surgery
(intraoperative)
Veins
Right ovarian
Right adrenal
Left ovarian
Left adrenal
Peripheral
Testosterone (ng/ml)
Basal
hCG
Intraoperative
3.04
2.63
3.02
2.89
3.41
5.7
5.3
5.9
4.8
105
Selective catheterization of the ovarian and adrenal veins by an experienced radiologist did not reveal any gradient of testosterone concentration in the samples taken from ovarian and adrenal veins. Catheterization was performed under basal (unstimulated) conditions and
after hCG administration (3) (Table 2). However, no correct catheterization of the right ovarian vein was possible during any of the two
examinations. Because no technical difficulty occurred during catheterization of the adrenal veins, it was concluded that the location of the
virilizing tumor was probably in one of the ovaries.
The patient wished to preserve her reproductive function. Intraoperative examination by an experienced surgeon found micropolycystic
ovaries with a small functional ovarian cyst on the left ovary; no other
macroscopic abnormality was found. Therefore, the decision was made
to perform intraoperative measurement of testosterone levels in blood
samples collected by the surgeon from the ovarian veins during laparotomy to perform a unilateral annexectomy.
Virilizing tumors of the ovaries constitute less than 5% of
all ovarian neoplasms (4). Many histological tumor types
have been described, and different classifications have been
proposed. The most frequent type is the Sertoli-Leydig cell
Materials and Methods
Blood samples were taken by the surgeon from the right and left
ovarian veins and from an antecubital vein. Serum testosterone levels
were measured intraoperatively using an automated direct chemiluminescent immunoassay (Advia Centaur, provided by Bayer Corp., Puteaux, France). The intraassay coefficient of variation is less than 6.2%.
The percentage of cross-reaction is 5.4% with 5␣-dihydrotestosterone,
0.94% with ⌬4-androstenedione, and less than 0.1% with other steroids.
Serum testosterone concentrations can be measured approximately 18
min after centrifugation of the sample. No manual aliquoting of the
centrifuged sample is required, and primary centrifuged samples can be
loaded directly onto the instrument. The ADVIA Centaur method enables the measurement of testosterone concentrations up to 15 ng/ml
and preprograming of dilutions before running the assay; measurement
of testosterone in the primary samples and in 1:10 diluted samples was
made for each sample taken from the ovarian veins. The sample volume
is 15 ␮l, but because of the dead volume, a minimum 50-␮l sample is
necessary. As with the more recent automated methods, it is possible to
load an additional sample once the automaton has been started, which
allows an emergency test to be performed. Thus, the testosterone concentration result is available less than 30 min after the sample has been
received in the laboratory, including all of the steps required (centrifugation, loading of the tube, and measurement by the automated
method).
FIG. 1. Microscopic aspect of the tumor. A, The tumor (star) was well
circumscribed and in a hilar position, as confirmed by the presence of
the rete ovarii (arrow). B, Some tumor cells contained crystals of
Reinke (arrow). C, The tumor cells were frequently separated in nest
or cords by anuclear eosinophilic areas with fibrinoid replacement of
the walls of some moderate-sized vessels (arrow). Initial magnifications were: A, ⫻100; B, ⫻400; C, ⫻200.
3076
J Clin Endocrinol Metab, July 2002, 87(7):3074 –3077
tumor (arrhenoblastoma), which is observed in premenopausal women. In our patient the virilizing tumor is a preferentially hilar cell tumor, which is one of the steroid cell
tumors, representing 0.1% of all ovarian tumors (5). Eighty
percent of such tumors result in hyperandrogenism (5). They
are small tumors (⬍4 cm), unilateral in more than 95% of
cases, with excellent prognosis because almost all of them are
benign. Hilar Leydig cell tumors are more often observed in
postmenopausal women and are gonadotropin dependent.
The rare hilar non-Leydig cell tumors that are diagnosed in
younger women belong to another histological type classified as nonspecific tumors. Only 50% of them are associated
with hyperandrogenism, and 20 – 40% are malignant (6).
However, when located in the hilus, they are more likely to
be benign, and a small size is a criterion for better prognosis.
In our premenopausal patient, the histopathological findings
are those of a Leydig cell tumor, because the tumor contains
large or medium-size cells with an abundant eosinophilic
cytoplasm, crystals of Reinke, small centered nuclei and
unique nucleolus, and rare mitoses.
In our patient localization of the virilizing tumor was
provided only by intraoperative measurement of testosterone, whereas hormonal data, radiological investigations,
scintigraphy imaging, selective venous catheterization, and
even inspection of the ovaries during laparotomy failed to
localize the tumor. First we had to determine that the virilizing tumor was not the left adrenal nodule; such a nodule
is actually a silent adrenal adenoma, as detected by CT scan
in 2% of subjects (incidentaloma) (7). Typical virilizing adrenal tumors are large malignant masses with DHEAS levels
and are unresponsive to hCG. However, small virilizing adrenal tumors may be observed, and it is known that DHEAS
levels are normal in as many as 20% of adrenal virilizing
tumors, and that several virilizing adrenal tumors express
hCG receptors (7). Finally, the lack of gradient between the
two adrenal veins, which had been catheterized without
technical problems, provided evidence against an adrenal
source of testosterone.
The next step was to show that the virilizing tumor was
actually in the right ovary, because correct catheterization
of the right ovarian vein had not been possible. None of
the radiological investigations showed the tumor. Ovarian
virilizing tumors of the adult are often too small to allow
detection by ultrasound examination, and adult premenopausal patients with virilizing ovarian tumors may have
micropolycystic ovaries (8). Tumor size should be at least
0.5 or 2 cm to allow correct detection by transvaginal (9)
or pelvic (10) sonography. Gray scale transvaginal sonography combined with color Doppler studies can more easily allow the diagnosis of small steroid cell tumors (5). MRI
scan could give more contrasted imaging and better detection of ovarian tumors. Recently, an MRI technique
comprising transvaginal examination has appeared promising (11). Iodomethylnorcholesterol scanning can also be
used (12). Selective catheterization of ovarian veins can be
very informative (13–15), but is associated with technical
problems, especially for catheterization of the right ovarian vein for anatomical reasons (12, 16). Finally, exploration by laparoscopy or laparotomy does not always lead
to correct determination of the location of small ovarian
Regnier et al. • Clinical Case Seminar
tumors (1). Thus, intraoperative measurement of testosterone in the ovarian veins is valuable, because it can
localize the source of testosterone secretion, probably better than intraoperative direct sonography of the ovaries,
which provides only morphological information.
In our patient intraoperative measurement of testosterone levels in samples taken by the surgeon from each
ovarian vein revealed a gradient of testosterone levels and
led to the correct localization and conservative surgery
of the ovarian virilizing tumor. To date, several direct
immunoassays for testosterone are available. Eight nonradioactive direct testosterone immunoassays [ACS 180
(Ciba Corning Diagnostics, Norwood, MA), Architect
i2000 (Abbot Diagnostics, Maidenhead, UK), AutoDelfia
(Pharmacia Wallac, Milton Keynes, UK), Elecsys 2010
(Roche Diagnostics, East Sussex, UK), Immulite 2000
(Diagnostic Products Ltd., Gwynedd, Wales, UK), Immuno 1 (Bayer, Tarrytown, NY), Vidas (BioMerieux,
Marcy L’Etoile, France), Vitros Eci (Ortho-Clinical Diagnostics, Cardiff, Wales, UK)] and two RIAs [Immunotech
(Beckman Coulter, Westbrook, ME) and DPC (Diagnostic
Products, Los Angeles, CA)] have been recently evaluated
by Taieb et al. (17), whereas Wheeler (18) reviewed the
performances of several automated immunoassay analyzers, including a few methods for measurement of testosterone levels (ACS:180 and Advia-Centaur, provided by
Bayer Corp., and Immulite and Immulite 2000, provided
by Diagnostic Products). The study by Taieb et al. (17)
showed that all direct assays had good interassay reproducibility and provided accurate evaluation of plasma
testosterone for concentrations within the male normal
range (⬃5 ng/ml), i.e. the same range of values as that
observed in most female patients with virilizing tumors.
A strong correlation was also observed by Fitzgerald et al.
(19) between the results given by ACS 180 and those of
isotopic dilution gas chromatography-mass spectrometry
for this range of testosterone concentrations. ACS 180 is
the first fully automated nonradioactive testosterone immunoassay approved by the U.S. FDA (20). The ADVIACentaur method, which we used for the present study,
employs the same reagents as those of ACS 180; the latter
method requires 15 min for testosterone evaluation, according to Wheeler et al. (20). In the ADVA-Centaur
method, only the automation employed is different. More
recently, a new modular chemiluminescence immunoassay analyzer, compared with gas chromatography-mass
spectrometry, showed better reproducibility than ACS 180
(21). All of the more recent automated direct nonradioactive testosterone immunoassays allow the insertion of additional emergency samples once the instrument has been
started, the loading of centrifuged primary samples without manual aliquoting, and the preprograming of the
preparation of diluted samples. All such direct testosterone assays make possible rapid and accurate evaluation of
testosterone levels within the range observed in patients
with virilizing tumors. Thus, the short time required for
the assay makes such direct assays especially suitable for
intraoperative localization of testosterone-secreting
tumors.
Regnier et al. • Clinical Case Seminar
J Clin Endocrinol Metab, July 2002, 87(7):3074 –3077 3077
Conclusion
In women with recent-onset or rapidly progressive hirsutism, with or without normal menstrual cycles, measurement of testosterone levels is mandatory to screen for virilizing tumors. The finding of an adrenal nodule on CT scan
does not rule out an ovarian source of the tumoral secretion
of androgens.
Intraoperative measurement of testosterone in samples
taken directly from the ovarian veins can be useful to localize
the source of testosterone secretion when selective catheterization of adrenal and ovarian veins fails to provide full
diagnostic evidence for the localization of the tumor. This
probably represents one major advantage of the recent automated methods of testosterone measurements.
6.
7.
8.
9.
10.
11.
12.
13.
Acknowledgments
14.
Received November 15, 2001. Accepted April 11, 2002.
Address all correspondence and requests for reprints to: P. Caron,
M.D., Service d’Endocrinologie, Hôpital Rangueil, avenue Jean Poulhès,
31403 Toulouse Cedex, France. E-mail: [email protected].
References
1. Pugeat M, Mirakian P, Dutrieux-Berger N, Forest MG, Tourniaire J 1997
Androgen-secreting ovarian neoplasm. In: Azziz R, Nestler JE, Dewailly D,
eds. Androgen excess disorders in women. Philadelphia, New York: Lippincott-Raven; 555–568
2. Caron Ph, Cogne M, Rumeau JL, Hoff J 1993 Androgenic granulosa cell tumor
of the ovary: in vivo hormonal studies. J Endocrinol Invest 16:545–548
3. Cohen I, Cuperman S, Altaras MM, Ben-Nun I, Goldberg E, Beyth Y 1992
Combined ovarian vein catheterization with ovarian stimulation in the diagnosis of androgen overproduction. Acta Obstet Gynecol Scand 71:245–248
4. Faraj G, Di Gregorio S, Misiunas A, Faure AN, Villabrile P, Stringa I, Petroff
N, Bur G 1998 Virilizing ovarian tumor of cell type not otherwise specified:
a case report. Gynecol Endocrinol 12:347–352
5. Monteagudo A, Heller D, Husami N, Levine RU, McCaffrey R, Timor-
15.
16.
17.
18.
19.
20.
21.
Tritsch IE 1997 Ovarian steroid cell tumors: sonographic characteristics. Ultrasound Obstet Gynecol 10:282–288
Young RH 1991 Ovarian tumors other than those of surface epithelial-stromal
type. Hum Pathol 22:763–775
Derksen J 1997 Androgen secreting adrenal neoplasms. In: Azziz R, Nestler
JE, Dewailly D, eds. Androgen excess disorders in women. Philadelphia, New
York: Lippincott-Raven; 545–553
Wong IL, Lobo RA 1996 Ovarian androgen-producing tumors. In: Adashi EY,
Rock JA, Rosenwaks Z, eds. Reproductive endocrinology, surgery, and technology. Philadelphia, New York: Lippincott-Raven; 1571–1598
Lobo RA 1991 Ovarian hyperandrogenism and androgen-producing tumor.
Endocrinol Metab Clin North Am 20:773– 805
Walsh JW, Taylor KJW, McWasson JF, Schwartz PE, Rosenfiel AR 1979 Gray
scale ultrasound in 204 proved gynaecologic masses: accuracy and specific
diagnosed criteria. Radiology 130:391–397
Occhipinti KA, Frankel SD, Hricak H 1993 The ovary. Computed tomography
and magnetic resonance imaging. Radiol Clin North Am 31:1115–1132
Taylor L, Ayers JW, Gross MD, Peterson EP, Menon KM 1986 Diagnostic
considerations in virilization: iodomethylnorcholesterol scanning in the localization of androgen secreting tumors. Fertil Steril 46:1005–1010
Friedman CI, Schmidt GE, Kim MH, Powel J 1985 Serum testosterone concentrations in the evaluation of androgen-producing tumors. Am J Obstet
Gynecol 153:44 – 49
Bricaire C, Raynaud A, Benotmane A, Clair F, Mowszowicz I, Wright F,
Moreau JF, Kuttenn F, Mauvais-Jarvis P 1991 Selective venous catheterization
in the evaluation of hyperandrogenism. J Endocrinol Invest 14:949 –956
Moltz L 1993 Diagnosis of endocrinally active ovarian tumors. Int J Gynecol
Pathol 12:170 –172
Sorensen R, Moltz L, Schwartz U 1986 Technical difficulties of selective blood
sampling in the differential diagnosis of female hyperandrogenism. Cardiovasc Intervent Radiol 9:75– 82
Taieb J, Mathian B, Boudou P, Millot F, Badonnel JY, Lacroix I, Mathieu E,
Queyrel N, Somma-Delpero C, Patricot MC 2001 Evaluation analytique de dix
trousses de dosage direct de la testostérone: comparaison à la CPG/SM. Immunoanal Biol Spec 16:338 –344
Wheeler MJ 2001 Automated immunoassay analysers. Ann Clin Biochem
38:217–229
Fitzgerald RL, Herold DA 1997 Serum total testosterone: immunoassay compared with negative chemical ionization gas chromatography-mass spectrometry. Clin Chem 43:749 –755
Wheeler MJ, D’Souza A, Matadeen J, Croos P 1996 Ciba Corning ACS:180
testosterone assay evaluated. Clin Chem 42:1445–1449
Ognibene A, Drake CJ, Jeng KY, Pascucci TE, Hsu S, Luceri F, Messeri G 2000
A new modular chemiluminescence immunoassay analyser evaluated. Clin
Chem Lab Med 38:251–260