- Email: [email protected]
Diagnosis of pituitary tumors by hormone assays and computerized tomography
CLINICAL STUDIES
Diagnosis of Pituitary Tumors by Hormone Assays and Computerized Tomography
LUBOMIR J. VALENTA, M.D., Ph.D. ROBERT D. SOSTRIN, M.D. HARVEY EISENBERG, M.D. JAMES A. TAMKIN, M.D. ALAN N. ELIAS, M.D. Irvine and Redondo Beach, California
From the Department of Medicine, University of California, Irvine, and the Medical Imaging Center, South Say Hospital, Redondo Beach, California. This work was supported in part by TL Enterprises, Agoura, California. Reprint requests should be addressed to Dr. Lubomir J. Valenta. Diabetes and Endccrine Clinic of North County, 2131 El Camino Real, Oceanside, California 92054. Manuscript accepted December 1, 198 1.
One hundred and seventy patients with endocrine abnormalities, 29 males and 141 females, were studied by computerized tomography (CT) scanning of the pituitary and by pituitary hormone assays including basal and stimulated hormone concenbations in the Mood. Pituitary macroadenoma was observed in 40 of the 170 patients, microadenoma in 49, primary empty seiia in 46, secondary empty seiia in 25, including 20 of the 40 patients with macroadenoma after surgical removal or radiation therapy of the tumor. Supraseiiar tumor was present in four patients. in 23 women with menstrual abnormalities and 3 men with impotence and hyperproiactinemia, the results of CT scanning were normal. 01 the 40 patients with macroadenoma, 24 presented with a hyperfuncttoning syndrome (acromegaiy or hyperproiactinemia). Among the 49 patfents with microadenomas, 46 had the amenonheagaiactorrhea spndrome, one male presented with acromegaiy and one with Cushkg’s disease, and one female also had Cushfng’s disease. Prtmary empty seiia was associated with the amenorrhea-gaiactorrhea syndrome in females and impotence in males. Only four female patients demonstrating an empty seiia were asymptomatk. Distinctive biochemkai findings were identified in certain groups of patients with different pituitary pathologic features. CT scanning was found very useful in the assessment of the size of the tumor, its supraseiiar or paraseiiar extension and response to therapy, and in differential diagnosis of tumor and empty seiia. Disappointing results were obtained in patients with endocrine abnormalities and normal resufts of CT scanning, and in patients foliowing surgical treatment for a pftuftary tumor, when the identification of recurrent tumor was impossibie from a single examination. it is concluded that CT scanning is the most advanced technique in the diagnosis of anatomic pituitary abnormalities, and it should replace the traditional invasive diagnostic procedures. Until recently, preoperative diagnosis of pituitary tumors depended on invasive procedures such as pneumoencephaiography or cerebral angiography. These procedures helped to assess the extent of the tumor, especially in regard to its supraseliar extension, and also to differentiate solid tumors from nonsurgical lesions such as the “empty” sella. They also were being used to differentiate intrasellar and suprasellar abnormalities. Since the advent of high-resolution computerized tomography (CT), this technique is rapidly replacing the traditional invasive diagnostic procedures [l-l 11. The present work reports our experience with the use of high-resolution CT in the differential diagnosis of pituitary tumors and compares the anatomic ab-
June 1982
The American Journal of Medicine
Volume 72
961
CT SCANNING FOR PITUITARY
TABLE I
TUMORS-VALENTA
ET AL.
Hormone Radioimmunoassay Characterization Interassay Coeliicient of Variation
Assay Sensitivity
Charcoal Second AB Second AB Second AB
10.3 7.5 13.0 6.4
10 pg/ml 0.75 rig/ml 0.5 ng/ml 1 @J/ml
Second AB
7.5
1 mlU/ml
Second AB
9.3
0.5 mlU/ml
Second AB
9.2
PEG Second Second Second Second
7.6
1 pg/dl 0.5 pgldl 12.5 ng/dl 10 ngldl 15 pg/ml 20 ng/dl
Bound/Free Separalion
Hormone Assay ACTH Growth hormone Prolactin Thyroid-stimulating hormone Luteinizing hormone Follicle-stimulating hormone
Cortisol (A.M.) T4 T3
Testosterone Estradiol Progesterone AB = antibody; PEG = polyethylene
glycol.
normalities
pattern
with the hormonal
obtained
in the
same patients. PATIENTS AND METHODS One hundred and seventy patients were studied, 29 males and 141 females.
Pftuitary Reserve Test. Pituitary reserve test was performed after an overnight fast in a patient who rested in a quiet room for at least 30 minutes following placement of a heparinized catheter into the cubital vein. Fasting (Time 0) sample was then obtained and analyzed for serum glucose and, by specific radioimmunoassays, for plasma ACTH, serum cortisol, growth hormone, prolactin. thyroid-stimulating hormone, luteinizing hormone, follicle-stimulating hormone, testosterone and free testosterone in males, estradiol and progesterone in females, T4 (thyroxine) and Ts (triiodothyronine). Sex hormone-binding globulin and thyroxine-binding globulin were also measured for total binding capacity. Subsequently, the patient received a mixture of 500 pg of thyrotropin-releasing hormone, 100 pg of gonadotropin releasing hormone (Factret@) and 0.2 U/kg of body weight of regular insulin as an intravenous bolus. Blood was then collected 20, 30 and 60 minutes later for plasma ACTH, serum cortisol, growth hormone, prolactin, thyroid-stimulating hormone, luteinizing hormone, follicle-stimulating hormone and glucose determinations. Similar protocols have been described previously by others [12-171.
Analytical Techniques.
Serum glucose was measured by
the glucose oxidase assay. Hormones were measured by specific radioimmunoassays. Details of the techniques are summarized in Table I. CT Scanning. CT scanning was performed using the General Electric (GE) CT/T 8800 model. All scans were obtained following intravenous infusion of 150 ml of Conray-60. The patient was placed in a prone position with the head hyperextended within the gantry. A computer-generated digital
862
4.8
AB i- PEG AB AB AB
June 1982 The American Journal of Medicine Volume72
9.0 6.7 7.3
scout radiograph (Scout View@) was taken to determine appropriate gantry angle and position. Contiguous 1.5 mm thick sections were preselected and direct coronal scans obtained. Similarly, with the patient in a supine position, axial sections through the sella turcica were taken. Sagittal reconstructions were made from the axial images. When cisternography was indicated, approximately 5 to 6 ml of the contrast agent metrizamide in a concentration of 170 mg/lOO ml was injected intrathecally into the lumbar cistern. The patient was placed in the Trendelenburg position for approximately 60 seconds and then returned to a recumbent position. Immediately after that, CT scanning was performed as just described.
RESULTS Normal Results of CT Scanning of the Pituitary. Normal results of CT scanning of the pituitary are illustrated in Figure 1. On direct coronal views, normal pituitary gland was seen filling the sella turcica as an area of increased density, which was higher than the density of the brain tissue. The upper surface of the gland was slightly concave. Maximal height was between 5 and 7 mm. Laterally, the pituitary could be separated from cavernous sinuses that were of even higher density. A pituitary stalk was visualized frequently, surrounded by a low-density area corresponding to the suprasellar cistern. The contrast material (Conray60) in the suprasellar blood vessels allowed for visualization of the internal carotid atieries and their branches, namely arteria cerebri anterior and media, analogous to cerebral angiography (Figure 1). On sagittal reconstructions, the bony structure of the sella turcica was similarly outlined as on the traditional
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
Figure 7. A, CT scan of normal pituitary, coronal projection; B, schematic representation of structures seen in A. The white areas represent the bone structures, i.e., the anterior clinoid processes and the sphenoid bone with the sphenoid sinus, forming the floor of the sella turcica. Stippled area corresponds to the pituitary interconnected with the hypothalamus by the pituitary stalk. The third ventricle is seen as a low-density diamond-shaped area above the pituitary stalk. The following vasculatufe can be identified: the “siphon” of the internal carotid artery next to the pituitary gland; continuation of the carotid artery as the supfaclinoid segment, with the take off of the posterior communicating artery; finally division of the carotid artery into the MI segment of the middle cerebral and the AI segment of the anterior cerebralartery. The cavernous sinuses are lateraland caudal to the siphon of the carotid artery.
lateral skull films. It was filled with the electrondense pituitary gland separated from the hypothalamus by the suprasellar cistern of low electron density. Axial sections were less reliable than the coronal and sagittal views for the evaluation of intrasellar contents. Suprasellar tomograms should show an area of low density corresponding to the suprasellar cistern. Intrasellar tomograms visualize pituitary tissue of a high density. Normal Results of Pituitary Reserve Testing. Normal
TABLE II
basal and stimulated values of pituitary hormones as obtained in a large series of healthy subjects are summarized in Table II. Pituitary Macroadenoma. The diagnosis of a pituitary macroadenoma was made whenever a solid pituitary lesion was present that exceeded 1 cm in diameter. This was observed in 40 patients, 26 females and 14 males. The age of the patients at the time of diagnosis varied between 17 and 82 years for the females and between 40 and 83 years for the males.
Pituitary Reserve Test: Normal Values*
Baseline (Time 0) lest Glucose Cortisol ACTH Growth hormone Prolactin Females Males Thyroid-stimulating hormone Luteinizing hormone Follicle-stimulating hormone l
Units mg/dl pg/dt og/mt ng/ml nglml ng/ml pU/ml mlU/ml mlU/ml
Mean f SD
Range
87.5 f 10.7 13.4 f 4.0
70-103 -
37.8 f 32.4 2.0 f 1.3 10.0 10.9 2.9 9.2 9.5
f f f f f
6.2 5.8 1.4 4.5 5.4
MaximalResponse Mean f SD Range
27.5 23.5 159.5 21.0
l-26 5-23 l-6 3-24 3-20
68.5 38.2 13.2 66.3 24.2
f 7.1 f 4.6 f 75.5 f 13.3
17-36 17-41 51-430 6-53
f f f f f
22-130 17-97 6-23 34-150 8-64
32.2 17.5 4.8 24.3 16.2
The data were obtained in at least 30 healthy, age- and sex-matched subjects.
June 1982
The American Journal of Medicine
Volume 72
963
CT SCANNING FOR PITUITARY TUMORS-VALENTA ET AL.
CT scan of a macroadenoma with suprasellar extension. The tumor presents as an enhancing lesion (arrows) that is demonstrated in three different projections: A, direct coronal; 6, direct axial, suprasellar; C, sagittal reconstruction.
Flgure 2.
Eight patients, five males and three females, had acromegaly. Their fasting serum growth hormone level was between 14 and 68 ng/ml. In four cases, the growth hormone level did not increase significantly during the pituitary reserve test. In the remaining four patients, three- to six-fold increase in growth hormone was observed within an hour following thyrotropin-releasing hormone administration. Hyperprolactinemia was diagnosed in 21 patients with macroadenoma, five males and 16 females. All females presented with the syndrome of amenorrheagalactorrhea. All males but one complained of impotence but none of them demonstrated galactorrhea or gynecomastia. Two of the males also had acromegaly, and their serum growth hormone levels were elevated to 55 and 68 ng/ml, respectively. Serum prolactin levels
064
June lQ82
The
American Journal of Medlclne
Volume 72
in the remaining patients with macroadenoma varied between 236 and 800 ng/ml. None of the patients responded to thyrotropin-releasing hormone stimulation by an increase in serum prolactin. In eleven patients, four males and seven females, no hypersecretion of any of the pituitary hormones was demonstrated. However, all but two patients with macroadenoma demonstrated some deficiencies of the pituitary functions. This consisted of nonmeasurable basal growth hormone and absent or blunted growth hormone response to stimulation in 26 of 40 patients, and low basal and stimulated levels of luteinizing hormone and follicle-stimulating hormone in all but two. The basal and stimulated levels of cortisol were low in two patients only. Five patients showed borderline low serum T4 and T3 levels and inappropriately low thy-
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
CT scan of a pituitary macroadenomaof homogeneomlyincreased density extending into the left parasellar area (arrows). F/gufe 3.
roid-stimulating hormone levels, but only two of them were clinically hypothyroid. By CT scanning, as well as by conventional skull x-ray, the sella turcica was enlarged in all patients with macroadenoma. The CT scan showed a tumor of density higher than the brain tissue, more or less homogenous (Figures 2 and 3). The lesion was intrasellar in 18 patients In 14, a supraseiiar extension was observed with an irregular mass deforming the suprasellar cistern (Figure 2). Ten of these patients had defects of the visual fields that regressed after treatment. Finally, the eight remaining patients with macroadenoma demonstrated lateral extension of the tumor (Figure 3). In one patient with long-standing hypothyroidism, a pituitary macroadenoma was demonstrated by CT scanning (Flgure 4A). Substantial regression of the tumor occurred following eight months of levothyroxine replacement therapy (Figure 4B). A history suggestive of pituitary apoplexy was obtained in one patient, a 62 year old female with enlargement of the sella turcica. Her CT scan is illustrated in Ftgure 5. The lowdensity mass lesion inside the sella, with peripheral enhancement, is demonstrated. These findings are similar to those in previously published cases of pituitary apoplexy [9, lo]. Attachment of the pituitary stalk to the periphery of the lesion differentiates such conditions from the empty sella, in which the stalk can be seen inside the lowdensity space [ 151. Three of the patients with macroadenoma received radiation treatment only. Twenty-six patients have been thus far treated by surgery, which confirmed the presence of a solid lesion of a diameter greater than 1 cm.
Pituitary Microadenoma. Pituitary microadenoma was diagnosed when the diameter of the demonstrable pituitary mass lesion was less than 1 cm. The smallest lesion identified was 4 mm in diameter. Microadenoma was observed in 49 of 170 patients, 47 females and two males. The females were between 17 and 46 years old, average 28. The males were 32 and 65 years old. Thus, the patients with microadenoma were mostly younger than the patients with macroadenoma. One of the male patients presented with acromegalic features, and his fasting growth hormone level was elevated. In one male, age 65, and one female, age 21, the diagnosis of Cushing’s disease was established. All remaining females complained of menstrual abnormalities and/or galactorrhea. Their serum prolactin levels varied between 36 and 220 ng/ml, and in all 20 who were tested, an increment of at least 30 ng/ml occurred following stimulation. This was in contrast with the findings in macroadenoma producing excess prolaatin, in which all the basal values were higher than 230 ng/ml and the stimulated values were not markedly increased over the baseline. In 12 of 20 females with micrcadenomas, serum estradiol and progesterone as well as luteinizing hormone and follicle-stimulating hormone levels were low, and the response to stimulation was blwnted. In five, the response of the gonadotropins was normal, and in three, it was exaggerated. No other endocrine abnormalities were present. The diagnosis of microadenoma has been confirmed thus far in 14 patients who underwent surgery. In 20 patients with microadenoma, the lesion was, on CT scanning, of a lower density than the remaining pituitary tissue (Figure 6A). In 17, it was heterogenous,
1982
The American
Journal
of Medlcine
Volume
72
865
CT SCANNING FOR PITUITARY
TUMORS-VALENTA
ET AL
Figure 4. Pituitary macroadenoma with a suprasellar extension in a patient suffering from a long standing hypothyroidism. A, before, and 6, eight months after correction of hypothyroidism by levothyroxine.
666
June 1992
The American Journal of Medicine
Volume 72
(.T SCANNING FOR PITUITARY
F&UIS 6.
TUMORS-VALENTA
ET AL.
A, low density microadenoma
(arrow). The tumor is separated from the suprasellar cistern by a dense zone that may represent normel overlying pituitaty tissue. On surgery, 4 mm solid microa& enema was found in the location of the lesion. B, CT scan of a microadenoma with central enhancing zone and lowdensity peripheral zone (arrow). A solid encapsulated microahwma was found on surgery, of a d&meter of 6 mm.
of a high and low density (Figure 6s). In 12 patients, the lesion was isodense with the pituitary tissue (Figure 7). The tumor was mostly clearly separated from the rest of the pituitary tissue by a zone of a high or low density. However, in eight patients with an isodense lesion, exact separation from normal pituitary was impossible (Figure 7). Signs of the tumor in these cases were: increased height of the pituitary shadow, with vertical dimension exceeding 7 mm: upward convex rather than concave shape; focal thinning of the floor of the sella at the site of the bulging: displacement of the pituitary stalk to the side of the sella. Of the eight patients with such isodense microadenoma, presence of the lesion has been confbmed thus far in far who had surgery. One patient with a microadenoma and an elevated
serum prolactin level had multiple endocrine neoplasia type I. She also demonstrated an incidental finding of two meningiomas. Primary (Idiopathic) “Empty” Sella. P&nary empty sella was diagnosed by the CT scan in 46 of the 170 patients. Seven of the patients were males between ages 33 and 62, and 39 were females, ages 16 through 59, mean age 3 1. In 28 of the 46 patients, conventional x-ray films of the sella turcica and the polytomograms were interpreted as revealing no abnormalities: in the remaining 18 patients, the sella turcica was found to be abnormal, enlarged or eroded. The seven male patients complained of impotence, and their serum testosterone levels were decreased. Their serum luteinizing hormone and follicle-stimulating
June 1982
The American Journal of Medicine
Volume 72
867
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
F&N
7.
lso&n.se microadsnoma. The
tumor is not clearly separated from the rest of the pituitary. Its presence is evidenced by an increased vertical height of the pituitary, its upward convex shape and focal thinning of the floor of the sella (arrow). On surgery, a solid tumor 8 mm in diameter was present.
hormone levels were also low and showed blunted response to stimulation. Their serum prolactin levels were moderately elevated to values between 26 and 50 ng/ml, and there was a blunted response during the pituitary reserve test. Of the 39 females, two had primary amenorrhea due to ovarian dysgenesis. Their serum prolactin levels were mildly elevated, fluctuating between 30 and 40 ng/ml,
and their luteinizing hormone and follicle-stimulating hormone levels, both basal and stimulated, were normal but inappropriately low for the agonadal state (Valenta et al., to be published). Of the remaining female patients, 35 had amenorrhea and/or galactorrhea, and their basal serum prolactin levels varied between 8 and 460 ng/ml, being moderately elevated to less than 50 ng/ml in most of them. Complete pituitary reserve test was performed
Figure 8. Primary (idiopathic) empty sella. The sella turcica is slightly enlarged ofl both sag&i reconstructions (A) of axial views (B), andon direct coronal views (C). A low-density space is present inside the sella, in continuity with the suprasellar cistern. The pituitary presents as a narrow crescent-like zone of high density in the posterior-inferior part of the se/la (A, &tack arrow; C, small whtte arrow). In the axial view (B) the contents of the sella turcica are of a low density where they should show pituitary tissue of high density (B, arrow). Pituitary stalk is visualized running through the low-density space to the pituitary. This has been described as “infundibuium sign” 1151 (see C, large curved white arrow). a = anterior.
BBB
June 1982
The Amerlcen Journal of Medicine
Volume 72
CT SCANNING FOR PITUITARY
in 22 female patients. Variable prolactin responses to stimulation were seen, from blunted to exaggerated, without obvious correlation with the degree of hyperprolactinemia. One patient was biochemically hypothyroid (T4 level of 4 pg/dl and Ts level of 68 ng/dl), with the basal and stimulated thyroid-stimulating hormone levels within the normal range, which was clearly abnormal compared with her serum thyroid hormone levels. In 14 female patients, both basal and stimulated luteinizing hormone and/or follicle-stimulating hormone levels were decreased, and in four patients, both basal and stimulated luteinizing hormone levels were increased. In nine females, partial empty sella was seen in association with microadenoma. These patients are included in the group with microadenoma. Six patients of 22 tested demonstrated subnormal response of serum growth hormone to hypoglycemia, with maximal stimulated values between 4 and 8 ng/ ml. Typical finding on CT scanning of the empty sella was the extension of a low-density space inside the sella turcica. The best views for identification of the intrasellar low-density lesions were the sagittal reconstruction of the axial views (Figure 8A). However, the information could also be readily obtained from the coronal projections (Figure 8C), or from low axial views, which ordinarily show the shadow of the pituitary gland (Figure 86). Another diagnostic sign of the empty sella was the infundibulum sign [ 151, demonstrated by the course of the pituitary stalk through a space of low
TUMORS-VALENTA
ET AL.
density inside the sella to the rudimentary pituitary tissue (Figure 8C). In nine of our patients, the diagnosis of empty sella was subsequently confirmed by a metrizamide study that demonstrated penetration of the intrathecally applied contrast material inside the sella turcica. In one female, the arachnoid outpouohing extended inside the sphenoidal sinus through a defect in the floor of the sella. This was classified as a transition form between the empty sella and basal encephalocele
[=I. Secondary Empty Sella. Eighteen of the 40 patients with pituitary macroadenoma just described were reexamined after pituitary surgery, and two after supravoltage radiation therapy. Five additional patients who received surgical or, in one case, radiation treatment elsewhere were also studied. All these patients showed grossly abnormal sella turcica, enlarged and sometimes eroded, and in all, low-density space was seen inside the sella (Figure 9). However, dense, irregularly shaped material was also present, and it was impossible to decide when it corresponded to normal pituitary tissue, to residual or recurrent tumor, or to tissue reactive to surgery. When applicable, i.e., in the hyperfunctioning syndromes, hormonal studies such as serum growth hormone and somatomedin C, or prolactin levels were more reliable than CT scanning in assessment of the residual tumor. Suprasellar Lesions. Four patients with suprasellar lesions not related to the pituitary gland were examined.
Ffgure 9. CT scan of the secondary empty sella in a patient treated by transsphenoidal surgery for a pituitary macroadenoma. Sagittal reconstructions of direct axial views were obtainedat the level indicated by an interrupted line, in the mi&ie of the sella turcica. A-anterior. The sella turcica shows a focal defect in the anterior-inferior part of the floor due to previous trans-sphenoidal surgery (arrow). The contents of the sella are inhomogenous, partly of a low density in continuity with the suprasellar cistern (secondary emp& sella), partly of a high density in the bottom of the se/la and posteriorly. The high density can represent normal pituitary tissue, residual tumor or tissue reactive to surgery.
June 1982
The American Journal of Mediclne
Volume 72
889
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
Figure 10. CT scan of a suprasellar lesion, an optic chiasm-hypothaiamic glioma. On sag&al reconstruction (A) of direct axial views (6) the tumor is shown to protrude into the supraselbr cistern (arrow). Metrizamide cisternography (C) demonstrates the tumor (tu) as a solid mass deforming the cistemal system. At this level, the pituitary can be seen as a narrow rim of tissue (arrow) at the floor of the sella turcica.
level was normal but her follicle-stimulating hormone level was low, and the response to stimulation was blunted. Craniotomy confirmed the presence of glioma in the area of optic chiasm. Preoperative CT scan in this patient is illustrated in Figure 10. Both enhanced contrast and metrizamide study showed a well-delineated suprasellar mass. Two more patients, one with a suprasellar meningioma and one with a neuroglioma, were examined in whom the results of basal endocrine studies were normal. Normal Results of CT Scanning in Patients with Endocrine Disease. In 26 patients, three males with
One of them presented with oligomenorrhea, mild elevation of serum prolactin and low serum gonadotropin levels. She was shown to have a meningioma of the diaphragma sellae with suprasellar extension. The tumor was removed surgically. The second patient had neurofibromatosis, and two years prior to admission, amenorrhea and galactorrhea had developed. She was shown to have normal basal and stimulated ACTH, cortisol, thyroid-stimulating hormone and growth hormone levels. Her serum prolactin level was slightly elevated to 27 ng/ml and increased to 88 ng/ml following stimulation, which is normal. Her luteinizing hormone
TABLE III
Pltultary Lesions In 170 Patients: Correlations of the CT Scan wlth Clinical Findinas
Totals Cl Scan Macroadenoma Primary empty seila Secondary empty &la Supraseiiar lesions Normal Totals
Acromegaiy F Y
Patients
M
F
40 49
14 2
26 47
5 1
4 23 141
7
4 26 170
3 29
Hyperpoiactinemia’ M F 5
3 -
-
3
1 13
16 46
2 23 137 .--
Gushing’s Disease F M -
1
Hypop tiuitarism F M 12
1
_t
23 127
---
3 -
1
2 1
_.
74
-
_
SA
Hyperproiactinemia with the syndrome of amenorrhea-gaiactorrhea in females or impotence in males. t Biochemicalevidenceof hypogonadotropic hypogonadism is included, but it might be a consequence of hyperprolactinemia. which was present in ail of these patients. * Figures in parentheses indicate the number of patients included under macroadenoma in whom a diagnosis of secondary empty seiia was made after pituitary surgery. l
070
June 1962
The American Journal of Medicine Volume 72
CT SCANNING FOR PITUITARY
impotence and 23 females with menstrual disturbances and/or galactorrhea, normal results of CT scanning of the pituitary were obtained. Most of these patients demonstrated variable degrees of hyperprolactinemia, which, however, did not exceed 50 ng/ml. In some of them, normal serum prolactin levels were present. Correlations of the Clinical and Anatomic Diagnosis. Correlations of the results of CT scanning and clinical findings are summarized in Table III and IV. Acromegaly was diagnosed in 10 patients, eight of whom demonstrated macroadenoma on the CT scan, and one microadenoma. In the remaining case, the CT scan was compatible with secondary empty sella following previous treatment elsewhere. The syndrome of amenorrhea-galactorrhea in females or impotence in males was present in 130 of the 170 patients. The underlying pituitary pathology was pituitary macroadenoma in 16, a microadenoma in 46, primary empty sella in 42, secondary empty sella in five and hypothalamic lesion in two. In 24 patients demonstrating hyperprolactinemia, the results of CT scanning were normal. One male and one female had Cushing’s disease and their CT scans were compatible with the presence of a pituitary microadenoma. Some evidence of hypopituitarism was present in almost all patients with macroadenoma. It consisted mostly of growth hormone, luteinizing hormone and follicle-stimulating hormone deficiency. In hyperprolactinemic states associated with microadenoma, low basal and stimulated luteinizing hormone and folliclestimulating hormone levels were also observed, but were never associated with growth hormone deficiency. In the empty sella syndrome, hyperprolactinemia was present in 35 of the 39 cases, gonadotropin deficiency was not infrequent, blunted growth hormone response to hypoglycemia was observed in six of 22, and inappropriately low thyroid-stimulating hormone response to thyrotropin-releasing hormone was observed once. COMMENTS
The diagnosis of pituitary tumors previously relied on indirect and/or invasive methods such as x-ray study of the sella turcica, pneumoencephalography and cerebral angiography. Some of the small pituitary tumors not causing enlargement of the sella turcica could be suspected on polytomograms of the sella from focal defects or bulging of the floor of the sella. However, a poor correlation has been noted between such changes and the presence of microadenoma [ 171. Similar difficulties have been encountered in differentiating the empty sella from a solid pituitary lesion [ 181. Compared with the previously employed diagnostic techniques, CT scanning is noninvasive, it visualizes directly the soft tissue, as well as the supra- and para-
TABLE IV
TUMORS-VALENTA
ET AL.
Hormonal Findings in 170 Patients of the Present Series
Acromegaly: Elevated fasting growth hormone, in 50 percent further increase in response to thyrotropin-releasing hormone. Elevated prolactin with blunted response in some. Hypogonadotropic hypogonadism in most cases. Anatomic substrate: macroadenoma. Hyperprolactlnemic slates: Menstrual irregularities and/or galactorrhea in females, impotence in males: l Serum prolactin higher than 200 ng/ml; absent or blunted prolactin response to stimulation; hypopituitarism, especiaily involving growth hormone, luteinizing hormone, follicle-stimulating hormone. Ariaatomic substrate: macroadenoma producing prolactin l Serum prolactin between 50 and 200 ng/ml, some response to stimulation preserved; low or normal luteinizing hormone, follicle-stimulating hormone with variable response to stimulation; absence of other signs of hypopituitarism. Anatomic substrate: microadenoma (prolactinoma) l Serum prolactin normal or moderately elevated (mostly below 50 ng/ml); variable response to stimulation; hypopituitarism (luteinizing hormone, follicle-stimulating hormone, growth hormone) more frequent than with microadenomas. Anatomic substrate: primary empty sella l Serum prolactin slightly elevated or normal, responding to stimulation; signs of hypopituitarism may be present. Anatomic substrate: suprasellar lesions l Findings similar to those found in patients with microadenomas except that prolactin levels are mostly below 50 ng/ml, and normal or exaggerated responses of luteinizing hormone, follicle-stimulating hormone to stimulation are observed more frequently. Anatomic substrate: normal CT scan of the pituitary Hypopiluiiarism: Almost all patients with macroadenoma are affected, whether or not they demonstrate excessive production of growth hormone and/or prolactin. In patients other than acromegalics, the most frequently affected hormones are growth hormone, lukeinizinghormone. and fdlick-stimulating hormone. Hypopituitarism alone, without hyperfunctioning syndrome is present in about 35 to 40 percent of patients with macroadenoma. With microadenoma. low luteinizing hormone and follicle-stimulating hormone may be functional, secondary to hyperprolactinemia [21]. No other signs of hypopituitarism are present. Almost half of the patients with primary empty sella demonstrate partial hypopituitarism.
sellar vascular structures [ 191, and maximal information can be obtained during one session in an ambulatory patient. Thus the cost-effectiveness is markedly increased. However, it should be emphasized that a high-resolution CT technique should be used. The radiologic findings should be correlated with hormonal and surgical data. In the present study, the tumors were identified by CT scanning as lesions of variable density. Most of the macroadenomas, i.e., tumors found on CT scanning and upon surgery to be of a diameter greater than 1 cm, were of a density similar to that of normal pituitary tissue. A focal decrease in tissue density could be related to necrotic and degenerative changes. Central low density and peripheral enhancement have been re-
June 1992
The American Journal of Medicine
Volume 72
971
CT SCANNING FOR PITUITARY
TUMORS-VALENTA
ET AL.
ported as being typical of pituitary apoplexy [ 9, lo], and a similar image was observed in one of our patients with a history suggesting previous pituitary apoplexy. The microadenomas were found to be less than 1 cm in diameter both by CT scanning and at surgery. In 14 patients in whom surgery was performed, it confirmed the presence of a microadenoma in the location indicated by the CT scan. All these microadenomas were solid lesions. Thus low density of some of the tumors was not caused by necrotic changes. The exact reason that some of the small tumors present as low-density lesions is unclear, but possibly it is related to their low vascularity. In contrast to macroadenomas, the microadenomas did not cause enlargement of the sella turcica. They were mostly clearly separated from normal pituitary tissue and the suprasellar cistern by a peripheral zone of lower or higher density than the rest of the tumor. Again, the reasons for such distinct separation are not clear. It may be due to the presence of a capsule or compressed normal pituitary tissue. However, some of the isodense microadenomas were not visualized as clearly circumscribed lesions, and their separation from normal pituitary tissue on the CT scan was impossible. In such cases, increased height of the pituitary (greater than 7 mm in the highest vertical diameter), upward convex rather than concave shape, focal defects of the floor of the sella and displacement of the pituitary stalk were found to be useful diagnostic signs in the present as well as previous studies [ 11. The smallest microadenoma of the present series diagnosed by CT scanning and confirmed surgically was 4 mm in diameter. This seems to be the current limit of the employed technique. Smaller lesions than that would be impossible to distinguish from the heterogeneity of the normal pituitary tissue that is always present to some extent. CT scanning including the coronal and sagittal projections is an excellent technique for differentiating the empty sella from solid lesions. In questionable situations, CT scanning with metrizamide cisternography will decide. An artifactual appearance of empty sella can be created by less sophisticated equipment due to volume-averaging by the computer and the presence of a large volume “empty” space of the sphenoid sinus next to the sella turcica. This was not a problem in the present studies. In all our cases of primary empty sella, some soft tissue density could be identified inside the sella, adjacent to the contour of the bony structure, and frequently a pituitary stalk was visualized connecting this tissue with the hypothalamus. This has been described recently as the “infundibulum sign” [ 151. The character of the patients with primary empty sella was different from the classic description of a middle-aged, obese, multiparous, hypertensive woman
872
June 1982
The American Journal of Medicine
Volume 72
without any endocrine abnormalities [20]. Our patients were mostly young adults, normotensive and not necessarily obese, and almost all of them presented with hyperprolactinemia, galactorrhea and menstrual or sexual dysfunction. The changing image of the patient with empty sella is likely due to the fact that our series included patients referred primarily for endocrine abnormalities, and that the diagnosis of the empty sella is made frequently by CT scanning in the presence of normal size of the sella. Previously, such cases could not be diagnosed by conventional skull films. Disappointing results of CT scanning were obtained in our series in two groups of patients: those after surgical or radiation treatment of pituitary tumors and those with clinical abnormality but normal anatomic study. In the former group, it was often impossible to decide when there was a residual or recurrent tumor. In the latter group, the presence of underlying pathology was impossible to discover. Follow up of the clinical and laboratory findings and repeated CT scanning in both of these groups of patients remain, at the present, the only diagnostic means. The overall prevalence of pituitary abnormalities in our series was almost five times higher in females than in males. Of the newly diagnosed pituitary lesions, more than 60 percent were solid tumors, with microadenoma slightly more frequent than macroadenoma. In the remainder, almost 40 percent, of patients with endocrine abnormalities, the underlying pathology was the primary empty sella. In the females, microadenoma was most frequent, followed by the primary empty sella, and macroadenoma was present only in about 20 percent. In contrast, in males, macroadenoma was identified in about half of the patients. The most frequent endocrine abnormality in our series was the amenorrhea and/or galactorrhea syndrome in females and impotence in males, with associated hyperprolactinemia. It was present in 135 of the 170 patients (Table Ill). Ten of 170 patients had acromegaly. In 11 of 170, no hyperfunctioning syndrome could be identified, but all these patients had hypopituitarism. In addition to the diagnostic value of the hormonal assays, the combination of the hormonal patterns in our series had a differential diagnostic importance, as summarized in Table IV. In summary, both hormonal testing and CT scanning have an important role in the differential diagnosis of pituitary lesions, and they make the use of invasive techniques obsolete. ACKNOWLEDGMENT The word processing expertise of Ms. Susan Burns and Miss Barbara McIntyre is appreciated.
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
REFERENCES 1.
2.
3.
4.
5.
6.
7.
6.
9.
10.
11.
Syvertsen A, Haughton VM, Williams AL, Cusick JF: The computed tomographic appearance of the normal pituitary gland and pituitary microadenomas. Radiology 1979; 133: 365-391. Wolpert SM. Post KD, Biller BJ, Molitch ME: The value of computed tomography in evaluating patients with prolactinomas. Radiology 1979; 131: 117-l 19. Leeds NE, Naidich TP: Computerized tomography in the diagnosis of sellar and parasellar lesions. Semin Roentgen 1977; 12: 121-135. Naidich TP, Pinto RS, Kushner MJ, et al.: Evaluation of sellar and parasellar masses by computed tomography. Radiology 1976; 120: 91-99. Reich NE, Zelch JV, Alfidi RJ, Meaney TF, Duchesneau PM, Weinstein MA: Computed tomography in the detection of juxtasellar lesions. Radiology 1976; 116: 333-335. Drayer BP, Rosenbaum AE, Kennerdell JS, Robinson AG. Bank WO, Deeb ZL: Computed tomographic diagnosis of suprasellar masses by intrathecal enhancement. Radiology 1977; 123: 339-344. Wiggli U, Benz UF: Normal computed tomography anatomy of the suprasellar subarachnoid space. Radiology 1978; 128: 65-70. Fitz CR, Wortzman G, Harwood-Nash DC, Holgate RC, Barry JF, Boktt DW: Computed tomography in craniopharyngiomas. Radiology 1978; 127: 687-69 1. Post MID, David NJ, Glaser JS, Safran A: Pituitary apoplexy: diagnosis by computed tomography. Radiology 1980; 134: 665-670. Fitzpatrick D, Tolis G, M&any EE. Taylor S: Pituitary apoplexy. The importance of skull roentgenograrns and computerized tomography in diagnosis. JAMA 1980; 244: 59-61. Manelfe C, Louvet J-P: Computed tomography in diabetes
12.
13.
14.
15.
16.
17.
18.
19.
20. 21.
June 1982
insipidus. J Comput Assist Tomogr 1979; 3: 309-316. Besser GM, Ratcliffe JG, Kilborn JR, Ormston BJ, Hall R: Interaction between thyrotrophin, corticotrophin and growth hormone secretion in man. J Endocrinol 1971; 51: 699. Harsoulis P, Marshall JC, Kuhn SF, Burke CW, London DR, Fraser TR: Combined test for assessment of anterior pituitary function. Br Med J 1973; 4: 326-329. Voigt KH, Dahlen HG, Fehm HL: Simultaneous stimulation test for the anterior pituitary hormones. Horm Metab Res 1974; 6: 436-437. Haughton V, Rosenbaum AE, Williams AL, Drayer B: Recognizing the empty sella by CT: the infundibulum sign. Am J Roentgen01 1981; 136: 293-295. Valenta CJ, Elias AN: lntrasphenoidal arachnoidocele with amenorrhea-galactorrhea and a pituitary apoplexy-like syndrome. Am J Obstet Gynecol (in press). Burrow GN. Wortzman G, Rewcastle NB, Holgate RC. Kovacs K: Microadenomas of the pituitary and abnormal sella tomograms in an unselected autopsy series. N Engl J Med 1981; 304: 156-158. Haney AF, Kramer RS, Wiebe RH, Hamond CB: Hypothalamic-pituitary function and radiographic evaluation of women with hyperprolactinemia and an “empty” sella turcica. Am J Obstet Gynecol 1979; 134: 9 17-924. Hayman LN, Evans RA, Hinck VC: Rapid high dose (RHD) contrast computerized tomography of perisellar vessels. Radiology 1979; 131: 121-123. Jordan RM, Kendall JW, Kerber CW: The primary empty sella syndrome. Am J Med 1977; 62: 569-580. Corbey RS, Leguin RM, Rolland R: Hyperprolactinemia and secondary amenorrhea. In: Crosignani PG. Robyn C, eds. Prolactin and human reproduction. New York: Academic Press, 1977; 203-215.
The American
Journal of Medicine
Volume 72
873
Diagnosis of Pituitary Tumors by Hormone Assays and Computerized Tomography
LUBOMIR J. VALENTA, M.D., Ph.D. ROBERT D. SOSTRIN, M.D. HARVEY EISENBERG, M.D. JAMES A. TAMKIN, M.D. ALAN N. ELIAS, M.D. Irvine and Redondo Beach, California
From the Department of Medicine, University of California, Irvine, and the Medical Imaging Center, South Say Hospital, Redondo Beach, California. This work was supported in part by TL Enterprises, Agoura, California. Reprint requests should be addressed to Dr. Lubomir J. Valenta. Diabetes and Endccrine Clinic of North County, 2131 El Camino Real, Oceanside, California 92054. Manuscript accepted December 1, 198 1.
One hundred and seventy patients with endocrine abnormalities, 29 males and 141 females, were studied by computerized tomography (CT) scanning of the pituitary and by pituitary hormone assays including basal and stimulated hormone concenbations in the Mood. Pituitary macroadenoma was observed in 40 of the 170 patients, microadenoma in 49, primary empty seiia in 46, secondary empty seiia in 25, including 20 of the 40 patients with macroadenoma after surgical removal or radiation therapy of the tumor. Supraseiiar tumor was present in four patients. in 23 women with menstrual abnormalities and 3 men with impotence and hyperproiactinemia, the results of CT scanning were normal. 01 the 40 patients with macroadenoma, 24 presented with a hyperfuncttoning syndrome (acromegaiy or hyperproiactinemia). Among the 49 patfents with microadenomas, 46 had the amenonheagaiactorrhea spndrome, one male presented with acromegaiy and one with Cushkg’s disease, and one female also had Cushfng’s disease. Prtmary empty seiia was associated with the amenorrhea-gaiactorrhea syndrome in females and impotence in males. Only four female patients demonstrating an empty seiia were asymptomatk. Distinctive biochemkai findings were identified in certain groups of patients with different pituitary pathologic features. CT scanning was found very useful in the assessment of the size of the tumor, its supraseiiar or paraseiiar extension and response to therapy, and in differential diagnosis of tumor and empty seiia. Disappointing results were obtained in patients with endocrine abnormalities and normal resufts of CT scanning, and in patients foliowing surgical treatment for a pftuftary tumor, when the identification of recurrent tumor was impossibie from a single examination. it is concluded that CT scanning is the most advanced technique in the diagnosis of anatomic pituitary abnormalities, and it should replace the traditional invasive diagnostic procedures. Until recently, preoperative diagnosis of pituitary tumors depended on invasive procedures such as pneumoencephaiography or cerebral angiography. These procedures helped to assess the extent of the tumor, especially in regard to its supraseliar extension, and also to differentiate solid tumors from nonsurgical lesions such as the “empty” sella. They also were being used to differentiate intrasellar and suprasellar abnormalities. Since the advent of high-resolution computerized tomography (CT), this technique is rapidly replacing the traditional invasive diagnostic procedures [l-l 11. The present work reports our experience with the use of high-resolution CT in the differential diagnosis of pituitary tumors and compares the anatomic ab-
June 1982
The American Journal of Medicine
Volume 72
961
CT SCANNING FOR PITUITARY
TABLE I
TUMORS-VALENTA
ET AL.
Hormone Radioimmunoassay Characterization Interassay Coeliicient of Variation
Assay Sensitivity
Charcoal Second AB Second AB Second AB
10.3 7.5 13.0 6.4
10 pg/ml 0.75 rig/ml 0.5 ng/ml 1 @J/ml
Second AB
7.5
1 mlU/ml
Second AB
9.3
0.5 mlU/ml
Second AB
9.2
PEG Second Second Second Second
7.6
1 pg/dl 0.5 pgldl 12.5 ng/dl 10 ngldl 15 pg/ml 20 ng/dl
Bound/Free Separalion
Hormone Assay ACTH Growth hormone Prolactin Thyroid-stimulating hormone Luteinizing hormone Follicle-stimulating hormone
Cortisol (A.M.) T4 T3
Testosterone Estradiol Progesterone AB = antibody; PEG = polyethylene
glycol.
normalities
pattern
with the hormonal
obtained
in the
same patients. PATIENTS AND METHODS One hundred and seventy patients were studied, 29 males and 141 females.
Pftuitary Reserve Test. Pituitary reserve test was performed after an overnight fast in a patient who rested in a quiet room for at least 30 minutes following placement of a heparinized catheter into the cubital vein. Fasting (Time 0) sample was then obtained and analyzed for serum glucose and, by specific radioimmunoassays, for plasma ACTH, serum cortisol, growth hormone, prolactin. thyroid-stimulating hormone, luteinizing hormone, follicle-stimulating hormone, testosterone and free testosterone in males, estradiol and progesterone in females, T4 (thyroxine) and Ts (triiodothyronine). Sex hormone-binding globulin and thyroxine-binding globulin were also measured for total binding capacity. Subsequently, the patient received a mixture of 500 pg of thyrotropin-releasing hormone, 100 pg of gonadotropin releasing hormone (Factret@) and 0.2 U/kg of body weight of regular insulin as an intravenous bolus. Blood was then collected 20, 30 and 60 minutes later for plasma ACTH, serum cortisol, growth hormone, prolactin, thyroid-stimulating hormone, luteinizing hormone, follicle-stimulating hormone and glucose determinations. Similar protocols have been described previously by others [12-171.
Analytical Techniques.
Serum glucose was measured by
the glucose oxidase assay. Hormones were measured by specific radioimmunoassays. Details of the techniques are summarized in Table I. CT Scanning. CT scanning was performed using the General Electric (GE) CT/T 8800 model. All scans were obtained following intravenous infusion of 150 ml of Conray-60. The patient was placed in a prone position with the head hyperextended within the gantry. A computer-generated digital
862
4.8
AB i- PEG AB AB AB
June 1982 The American Journal of Medicine Volume72
9.0 6.7 7.3
scout radiograph (Scout View@) was taken to determine appropriate gantry angle and position. Contiguous 1.5 mm thick sections were preselected and direct coronal scans obtained. Similarly, with the patient in a supine position, axial sections through the sella turcica were taken. Sagittal reconstructions were made from the axial images. When cisternography was indicated, approximately 5 to 6 ml of the contrast agent metrizamide in a concentration of 170 mg/lOO ml was injected intrathecally into the lumbar cistern. The patient was placed in the Trendelenburg position for approximately 60 seconds and then returned to a recumbent position. Immediately after that, CT scanning was performed as just described.
RESULTS Normal Results of CT Scanning of the Pituitary. Normal results of CT scanning of the pituitary are illustrated in Figure 1. On direct coronal views, normal pituitary gland was seen filling the sella turcica as an area of increased density, which was higher than the density of the brain tissue. The upper surface of the gland was slightly concave. Maximal height was between 5 and 7 mm. Laterally, the pituitary could be separated from cavernous sinuses that were of even higher density. A pituitary stalk was visualized frequently, surrounded by a low-density area corresponding to the suprasellar cistern. The contrast material (Conray60) in the suprasellar blood vessels allowed for visualization of the internal carotid atieries and their branches, namely arteria cerebri anterior and media, analogous to cerebral angiography (Figure 1). On sagittal reconstructions, the bony structure of the sella turcica was similarly outlined as on the traditional
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
Figure 7. A, CT scan of normal pituitary, coronal projection; B, schematic representation of structures seen in A. The white areas represent the bone structures, i.e., the anterior clinoid processes and the sphenoid bone with the sphenoid sinus, forming the floor of the sella turcica. Stippled area corresponds to the pituitary interconnected with the hypothalamus by the pituitary stalk. The third ventricle is seen as a low-density diamond-shaped area above the pituitary stalk. The following vasculatufe can be identified: the “siphon” of the internal carotid artery next to the pituitary gland; continuation of the carotid artery as the supfaclinoid segment, with the take off of the posterior communicating artery; finally division of the carotid artery into the MI segment of the middle cerebral and the AI segment of the anterior cerebralartery. The cavernous sinuses are lateraland caudal to the siphon of the carotid artery.
lateral skull films. It was filled with the electrondense pituitary gland separated from the hypothalamus by the suprasellar cistern of low electron density. Axial sections were less reliable than the coronal and sagittal views for the evaluation of intrasellar contents. Suprasellar tomograms should show an area of low density corresponding to the suprasellar cistern. Intrasellar tomograms visualize pituitary tissue of a high density. Normal Results of Pituitary Reserve Testing. Normal
TABLE II
basal and stimulated values of pituitary hormones as obtained in a large series of healthy subjects are summarized in Table II. Pituitary Macroadenoma. The diagnosis of a pituitary macroadenoma was made whenever a solid pituitary lesion was present that exceeded 1 cm in diameter. This was observed in 40 patients, 26 females and 14 males. The age of the patients at the time of diagnosis varied between 17 and 82 years for the females and between 40 and 83 years for the males.
Pituitary Reserve Test: Normal Values*
Baseline (Time 0) lest Glucose Cortisol ACTH Growth hormone Prolactin Females Males Thyroid-stimulating hormone Luteinizing hormone Follicle-stimulating hormone l
Units mg/dl pg/dt og/mt ng/ml nglml ng/ml pU/ml mlU/ml mlU/ml
Mean f SD
Range
87.5 f 10.7 13.4 f 4.0
70-103 -
37.8 f 32.4 2.0 f 1.3 10.0 10.9 2.9 9.2 9.5
f f f f f
6.2 5.8 1.4 4.5 5.4
MaximalResponse Mean f SD Range
27.5 23.5 159.5 21.0
l-26 5-23 l-6 3-24 3-20
68.5 38.2 13.2 66.3 24.2
f 7.1 f 4.6 f 75.5 f 13.3
17-36 17-41 51-430 6-53
f f f f f
22-130 17-97 6-23 34-150 8-64
32.2 17.5 4.8 24.3 16.2
The data were obtained in at least 30 healthy, age- and sex-matched subjects.
June 1982
The American Journal of Medicine
Volume 72
963
CT SCANNING FOR PITUITARY TUMORS-VALENTA ET AL.
CT scan of a macroadenoma with suprasellar extension. The tumor presents as an enhancing lesion (arrows) that is demonstrated in three different projections: A, direct coronal; 6, direct axial, suprasellar; C, sagittal reconstruction.
Flgure 2.
Eight patients, five males and three females, had acromegaly. Their fasting serum growth hormone level was between 14 and 68 ng/ml. In four cases, the growth hormone level did not increase significantly during the pituitary reserve test. In the remaining four patients, three- to six-fold increase in growth hormone was observed within an hour following thyrotropin-releasing hormone administration. Hyperprolactinemia was diagnosed in 21 patients with macroadenoma, five males and 16 females. All females presented with the syndrome of amenorrheagalactorrhea. All males but one complained of impotence but none of them demonstrated galactorrhea or gynecomastia. Two of the males also had acromegaly, and their serum growth hormone levels were elevated to 55 and 68 ng/ml, respectively. Serum prolactin levels
064
June lQ82
The
American Journal of Medlclne
Volume 72
in the remaining patients with macroadenoma varied between 236 and 800 ng/ml. None of the patients responded to thyrotropin-releasing hormone stimulation by an increase in serum prolactin. In eleven patients, four males and seven females, no hypersecretion of any of the pituitary hormones was demonstrated. However, all but two patients with macroadenoma demonstrated some deficiencies of the pituitary functions. This consisted of nonmeasurable basal growth hormone and absent or blunted growth hormone response to stimulation in 26 of 40 patients, and low basal and stimulated levels of luteinizing hormone and follicle-stimulating hormone in all but two. The basal and stimulated levels of cortisol were low in two patients only. Five patients showed borderline low serum T4 and T3 levels and inappropriately low thy-
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
CT scan of a pituitary macroadenomaof homogeneomlyincreased density extending into the left parasellar area (arrows). F/gufe 3.
roid-stimulating hormone levels, but only two of them were clinically hypothyroid. By CT scanning, as well as by conventional skull x-ray, the sella turcica was enlarged in all patients with macroadenoma. The CT scan showed a tumor of density higher than the brain tissue, more or less homogenous (Figures 2 and 3). The lesion was intrasellar in 18 patients In 14, a supraseiiar extension was observed with an irregular mass deforming the suprasellar cistern (Figure 2). Ten of these patients had defects of the visual fields that regressed after treatment. Finally, the eight remaining patients with macroadenoma demonstrated lateral extension of the tumor (Figure 3). In one patient with long-standing hypothyroidism, a pituitary macroadenoma was demonstrated by CT scanning (Flgure 4A). Substantial regression of the tumor occurred following eight months of levothyroxine replacement therapy (Figure 4B). A history suggestive of pituitary apoplexy was obtained in one patient, a 62 year old female with enlargement of the sella turcica. Her CT scan is illustrated in Ftgure 5. The lowdensity mass lesion inside the sella, with peripheral enhancement, is demonstrated. These findings are similar to those in previously published cases of pituitary apoplexy [9, lo]. Attachment of the pituitary stalk to the periphery of the lesion differentiates such conditions from the empty sella, in which the stalk can be seen inside the lowdensity space [ 151. Three of the patients with macroadenoma received radiation treatment only. Twenty-six patients have been thus far treated by surgery, which confirmed the presence of a solid lesion of a diameter greater than 1 cm.
Pituitary Microadenoma. Pituitary microadenoma was diagnosed when the diameter of the demonstrable pituitary mass lesion was less than 1 cm. The smallest lesion identified was 4 mm in diameter. Microadenoma was observed in 49 of 170 patients, 47 females and two males. The females were between 17 and 46 years old, average 28. The males were 32 and 65 years old. Thus, the patients with microadenoma were mostly younger than the patients with macroadenoma. One of the male patients presented with acromegalic features, and his fasting growth hormone level was elevated. In one male, age 65, and one female, age 21, the diagnosis of Cushing’s disease was established. All remaining females complained of menstrual abnormalities and/or galactorrhea. Their serum prolactin levels varied between 36 and 220 ng/ml, and in all 20 who were tested, an increment of at least 30 ng/ml occurred following stimulation. This was in contrast with the findings in macroadenoma producing excess prolaatin, in which all the basal values were higher than 230 ng/ml and the stimulated values were not markedly increased over the baseline. In 12 of 20 females with micrcadenomas, serum estradiol and progesterone as well as luteinizing hormone and follicle-stimulating hormone levels were low, and the response to stimulation was blwnted. In five, the response of the gonadotropins was normal, and in three, it was exaggerated. No other endocrine abnormalities were present. The diagnosis of microadenoma has been confirmed thus far in 14 patients who underwent surgery. In 20 patients with microadenoma, the lesion was, on CT scanning, of a lower density than the remaining pituitary tissue (Figure 6A). In 17, it was heterogenous,
1982
The American
Journal
of Medlcine
Volume
72
865
CT SCANNING FOR PITUITARY
TUMORS-VALENTA
ET AL
Figure 4. Pituitary macroadenoma with a suprasellar extension in a patient suffering from a long standing hypothyroidism. A, before, and 6, eight months after correction of hypothyroidism by levothyroxine.
666
June 1992
The American Journal of Medicine
Volume 72
(.T SCANNING FOR PITUITARY
F&UIS 6.
TUMORS-VALENTA
ET AL.
A, low density microadenoma
(arrow). The tumor is separated from the suprasellar cistern by a dense zone that may represent normel overlying pituitaty tissue. On surgery, 4 mm solid microa& enema was found in the location of the lesion. B, CT scan of a microadenoma with central enhancing zone and lowdensity peripheral zone (arrow). A solid encapsulated microahwma was found on surgery, of a d&meter of 6 mm.
of a high and low density (Figure 6s). In 12 patients, the lesion was isodense with the pituitary tissue (Figure 7). The tumor was mostly clearly separated from the rest of the pituitary tissue by a zone of a high or low density. However, in eight patients with an isodense lesion, exact separation from normal pituitary was impossible (Figure 7). Signs of the tumor in these cases were: increased height of the pituitary shadow, with vertical dimension exceeding 7 mm: upward convex rather than concave shape; focal thinning of the floor of the sella at the site of the bulging: displacement of the pituitary stalk to the side of the sella. Of the eight patients with such isodense microadenoma, presence of the lesion has been confbmed thus far in far who had surgery. One patient with a microadenoma and an elevated
serum prolactin level had multiple endocrine neoplasia type I. She also demonstrated an incidental finding of two meningiomas. Primary (Idiopathic) “Empty” Sella. P&nary empty sella was diagnosed by the CT scan in 46 of the 170 patients. Seven of the patients were males between ages 33 and 62, and 39 were females, ages 16 through 59, mean age 3 1. In 28 of the 46 patients, conventional x-ray films of the sella turcica and the polytomograms were interpreted as revealing no abnormalities: in the remaining 18 patients, the sella turcica was found to be abnormal, enlarged or eroded. The seven male patients complained of impotence, and their serum testosterone levels were decreased. Their serum luteinizing hormone and follicle-stimulating
June 1982
The American Journal of Medicine
Volume 72
867
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
F&N
7.
lso&n.se microadsnoma. The
tumor is not clearly separated from the rest of the pituitary. Its presence is evidenced by an increased vertical height of the pituitary, its upward convex shape and focal thinning of the floor of the sella (arrow). On surgery, a solid tumor 8 mm in diameter was present.
hormone levels were also low and showed blunted response to stimulation. Their serum prolactin levels were moderately elevated to values between 26 and 50 ng/ml, and there was a blunted response during the pituitary reserve test. Of the 39 females, two had primary amenorrhea due to ovarian dysgenesis. Their serum prolactin levels were mildly elevated, fluctuating between 30 and 40 ng/ml,
and their luteinizing hormone and follicle-stimulating hormone levels, both basal and stimulated, were normal but inappropriately low for the agonadal state (Valenta et al., to be published). Of the remaining female patients, 35 had amenorrhea and/or galactorrhea, and their basal serum prolactin levels varied between 8 and 460 ng/ml, being moderately elevated to less than 50 ng/ml in most of them. Complete pituitary reserve test was performed
Figure 8. Primary (idiopathic) empty sella. The sella turcica is slightly enlarged ofl both sag&i reconstructions (A) of axial views (B), andon direct coronal views (C). A low-density space is present inside the sella, in continuity with the suprasellar cistern. The pituitary presents as a narrow crescent-like zone of high density in the posterior-inferior part of the se/la (A, &tack arrow; C, small whtte arrow). In the axial view (B) the contents of the sella turcica are of a low density where they should show pituitary tissue of high density (B, arrow). Pituitary stalk is visualized running through the low-density space to the pituitary. This has been described as “infundibuium sign” 1151 (see C, large curved white arrow). a = anterior.
BBB
June 1982
The Amerlcen Journal of Medicine
Volume 72
CT SCANNING FOR PITUITARY
in 22 female patients. Variable prolactin responses to stimulation were seen, from blunted to exaggerated, without obvious correlation with the degree of hyperprolactinemia. One patient was biochemically hypothyroid (T4 level of 4 pg/dl and Ts level of 68 ng/dl), with the basal and stimulated thyroid-stimulating hormone levels within the normal range, which was clearly abnormal compared with her serum thyroid hormone levels. In 14 female patients, both basal and stimulated luteinizing hormone and/or follicle-stimulating hormone levels were decreased, and in four patients, both basal and stimulated luteinizing hormone levels were increased. In nine females, partial empty sella was seen in association with microadenoma. These patients are included in the group with microadenoma. Six patients of 22 tested demonstrated subnormal response of serum growth hormone to hypoglycemia, with maximal stimulated values between 4 and 8 ng/ ml. Typical finding on CT scanning of the empty sella was the extension of a low-density space inside the sella turcica. The best views for identification of the intrasellar low-density lesions were the sagittal reconstruction of the axial views (Figure 8A). However, the information could also be readily obtained from the coronal projections (Figure 8C), or from low axial views, which ordinarily show the shadow of the pituitary gland (Figure 86). Another diagnostic sign of the empty sella was the infundibulum sign [ 151, demonstrated by the course of the pituitary stalk through a space of low
TUMORS-VALENTA
ET AL.
density inside the sella to the rudimentary pituitary tissue (Figure 8C). In nine of our patients, the diagnosis of empty sella was subsequently confirmed by a metrizamide study that demonstrated penetration of the intrathecally applied contrast material inside the sella turcica. In one female, the arachnoid outpouohing extended inside the sphenoidal sinus through a defect in the floor of the sella. This was classified as a transition form between the empty sella and basal encephalocele
[=I. Secondary Empty Sella. Eighteen of the 40 patients with pituitary macroadenoma just described were reexamined after pituitary surgery, and two after supravoltage radiation therapy. Five additional patients who received surgical or, in one case, radiation treatment elsewhere were also studied. All these patients showed grossly abnormal sella turcica, enlarged and sometimes eroded, and in all, low-density space was seen inside the sella (Figure 9). However, dense, irregularly shaped material was also present, and it was impossible to decide when it corresponded to normal pituitary tissue, to residual or recurrent tumor, or to tissue reactive to surgery. When applicable, i.e., in the hyperfunctioning syndromes, hormonal studies such as serum growth hormone and somatomedin C, or prolactin levels were more reliable than CT scanning in assessment of the residual tumor. Suprasellar Lesions. Four patients with suprasellar lesions not related to the pituitary gland were examined.
Ffgure 9. CT scan of the secondary empty sella in a patient treated by transsphenoidal surgery for a pituitary macroadenoma. Sagittal reconstructions of direct axial views were obtainedat the level indicated by an interrupted line, in the mi&ie of the sella turcica. A-anterior. The sella turcica shows a focal defect in the anterior-inferior part of the floor due to previous trans-sphenoidal surgery (arrow). The contents of the sella are inhomogenous, partly of a low density in continuity with the suprasellar cistern (secondary emp& sella), partly of a high density in the bottom of the se/la and posteriorly. The high density can represent normal pituitary tissue, residual tumor or tissue reactive to surgery.
June 1982
The American Journal of Mediclne
Volume 72
889
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
Figure 10. CT scan of a suprasellar lesion, an optic chiasm-hypothaiamic glioma. On sag&al reconstruction (A) of direct axial views (6) the tumor is shown to protrude into the supraselbr cistern (arrow). Metrizamide cisternography (C) demonstrates the tumor (tu) as a solid mass deforming the cistemal system. At this level, the pituitary can be seen as a narrow rim of tissue (arrow) at the floor of the sella turcica.
level was normal but her follicle-stimulating hormone level was low, and the response to stimulation was blunted. Craniotomy confirmed the presence of glioma in the area of optic chiasm. Preoperative CT scan in this patient is illustrated in Figure 10. Both enhanced contrast and metrizamide study showed a well-delineated suprasellar mass. Two more patients, one with a suprasellar meningioma and one with a neuroglioma, were examined in whom the results of basal endocrine studies were normal. Normal Results of CT Scanning in Patients with Endocrine Disease. In 26 patients, three males with
One of them presented with oligomenorrhea, mild elevation of serum prolactin and low serum gonadotropin levels. She was shown to have a meningioma of the diaphragma sellae with suprasellar extension. The tumor was removed surgically. The second patient had neurofibromatosis, and two years prior to admission, amenorrhea and galactorrhea had developed. She was shown to have normal basal and stimulated ACTH, cortisol, thyroid-stimulating hormone and growth hormone levels. Her serum prolactin level was slightly elevated to 27 ng/ml and increased to 88 ng/ml following stimulation, which is normal. Her luteinizing hormone
TABLE III
Pltultary Lesions In 170 Patients: Correlations of the CT Scan wlth Clinical Findinas
Totals Cl Scan Macroadenoma Primary empty seila Secondary empty &la Supraseiiar lesions Normal Totals
Acromegaiy F Y
Patients
M
F
40 49
14 2
26 47
5 1
4 23 141
7
4 26 170
3 29
Hyperpoiactinemia’ M F 5
3 -
-
3
1 13
16 46
2 23 137 .--
Gushing’s Disease F M -
1
Hypop tiuitarism F M 12
1
_t
23 127
---
3 -
1
2 1
_.
74
-
_
SA
Hyperproiactinemia with the syndrome of amenorrhea-gaiactorrhea in females or impotence in males. t Biochemicalevidenceof hypogonadotropic hypogonadism is included, but it might be a consequence of hyperprolactinemia. which was present in ail of these patients. * Figures in parentheses indicate the number of patients included under macroadenoma in whom a diagnosis of secondary empty seiia was made after pituitary surgery. l
070
June 1962
The American Journal of Medicine Volume 72
CT SCANNING FOR PITUITARY
impotence and 23 females with menstrual disturbances and/or galactorrhea, normal results of CT scanning of the pituitary were obtained. Most of these patients demonstrated variable degrees of hyperprolactinemia, which, however, did not exceed 50 ng/ml. In some of them, normal serum prolactin levels were present. Correlations of the Clinical and Anatomic Diagnosis. Correlations of the results of CT scanning and clinical findings are summarized in Table III and IV. Acromegaly was diagnosed in 10 patients, eight of whom demonstrated macroadenoma on the CT scan, and one microadenoma. In the remaining case, the CT scan was compatible with secondary empty sella following previous treatment elsewhere. The syndrome of amenorrhea-galactorrhea in females or impotence in males was present in 130 of the 170 patients. The underlying pituitary pathology was pituitary macroadenoma in 16, a microadenoma in 46, primary empty sella in 42, secondary empty sella in five and hypothalamic lesion in two. In 24 patients demonstrating hyperprolactinemia, the results of CT scanning were normal. One male and one female had Cushing’s disease and their CT scans were compatible with the presence of a pituitary microadenoma. Some evidence of hypopituitarism was present in almost all patients with macroadenoma. It consisted mostly of growth hormone, luteinizing hormone and follicle-stimulating hormone deficiency. In hyperprolactinemic states associated with microadenoma, low basal and stimulated luteinizing hormone and folliclestimulating hormone levels were also observed, but were never associated with growth hormone deficiency. In the empty sella syndrome, hyperprolactinemia was present in 35 of the 39 cases, gonadotropin deficiency was not infrequent, blunted growth hormone response to hypoglycemia was observed in six of 22, and inappropriately low thyroid-stimulating hormone response to thyrotropin-releasing hormone was observed once. COMMENTS
The diagnosis of pituitary tumors previously relied on indirect and/or invasive methods such as x-ray study of the sella turcica, pneumoencephalography and cerebral angiography. Some of the small pituitary tumors not causing enlargement of the sella turcica could be suspected on polytomograms of the sella from focal defects or bulging of the floor of the sella. However, a poor correlation has been noted between such changes and the presence of microadenoma [ 171. Similar difficulties have been encountered in differentiating the empty sella from a solid pituitary lesion [ 181. Compared with the previously employed diagnostic techniques, CT scanning is noninvasive, it visualizes directly the soft tissue, as well as the supra- and para-
TABLE IV
TUMORS-VALENTA
ET AL.
Hormonal Findings in 170 Patients of the Present Series
Acromegaly: Elevated fasting growth hormone, in 50 percent further increase in response to thyrotropin-releasing hormone. Elevated prolactin with blunted response in some. Hypogonadotropic hypogonadism in most cases. Anatomic substrate: macroadenoma. Hyperprolactlnemic slates: Menstrual irregularities and/or galactorrhea in females, impotence in males: l Serum prolactin higher than 200 ng/ml; absent or blunted prolactin response to stimulation; hypopituitarism, especiaily involving growth hormone, luteinizing hormone, follicle-stimulating hormone. Ariaatomic substrate: macroadenoma producing prolactin l Serum prolactin between 50 and 200 ng/ml, some response to stimulation preserved; low or normal luteinizing hormone, follicle-stimulating hormone with variable response to stimulation; absence of other signs of hypopituitarism. Anatomic substrate: microadenoma (prolactinoma) l Serum prolactin normal or moderately elevated (mostly below 50 ng/ml); variable response to stimulation; hypopituitarism (luteinizing hormone, follicle-stimulating hormone, growth hormone) more frequent than with microadenomas. Anatomic substrate: primary empty sella l Serum prolactin slightly elevated or normal, responding to stimulation; signs of hypopituitarism may be present. Anatomic substrate: suprasellar lesions l Findings similar to those found in patients with microadenomas except that prolactin levels are mostly below 50 ng/ml, and normal or exaggerated responses of luteinizing hormone, follicle-stimulating hormone to stimulation are observed more frequently. Anatomic substrate: normal CT scan of the pituitary Hypopiluiiarism: Almost all patients with macroadenoma are affected, whether or not they demonstrate excessive production of growth hormone and/or prolactin. In patients other than acromegalics, the most frequently affected hormones are growth hormone, lukeinizinghormone. and fdlick-stimulating hormone. Hypopituitarism alone, without hyperfunctioning syndrome is present in about 35 to 40 percent of patients with macroadenoma. With microadenoma. low luteinizing hormone and follicle-stimulating hormone may be functional, secondary to hyperprolactinemia [21]. No other signs of hypopituitarism are present. Almost half of the patients with primary empty sella demonstrate partial hypopituitarism.
sellar vascular structures [ 191, and maximal information can be obtained during one session in an ambulatory patient. Thus the cost-effectiveness is markedly increased. However, it should be emphasized that a high-resolution CT technique should be used. The radiologic findings should be correlated with hormonal and surgical data. In the present study, the tumors were identified by CT scanning as lesions of variable density. Most of the macroadenomas, i.e., tumors found on CT scanning and upon surgery to be of a diameter greater than 1 cm, were of a density similar to that of normal pituitary tissue. A focal decrease in tissue density could be related to necrotic and degenerative changes. Central low density and peripheral enhancement have been re-
June 1992
The American Journal of Medicine
Volume 72
971
CT SCANNING FOR PITUITARY
TUMORS-VALENTA
ET AL.
ported as being typical of pituitary apoplexy [ 9, lo], and a similar image was observed in one of our patients with a history suggesting previous pituitary apoplexy. The microadenomas were found to be less than 1 cm in diameter both by CT scanning and at surgery. In 14 patients in whom surgery was performed, it confirmed the presence of a microadenoma in the location indicated by the CT scan. All these microadenomas were solid lesions. Thus low density of some of the tumors was not caused by necrotic changes. The exact reason that some of the small tumors present as low-density lesions is unclear, but possibly it is related to their low vascularity. In contrast to macroadenomas, the microadenomas did not cause enlargement of the sella turcica. They were mostly clearly separated from normal pituitary tissue and the suprasellar cistern by a peripheral zone of lower or higher density than the rest of the tumor. Again, the reasons for such distinct separation are not clear. It may be due to the presence of a capsule or compressed normal pituitary tissue. However, some of the isodense microadenomas were not visualized as clearly circumscribed lesions, and their separation from normal pituitary tissue on the CT scan was impossible. In such cases, increased height of the pituitary (greater than 7 mm in the highest vertical diameter), upward convex rather than concave shape, focal defects of the floor of the sella and displacement of the pituitary stalk were found to be useful diagnostic signs in the present as well as previous studies [ 11. The smallest microadenoma of the present series diagnosed by CT scanning and confirmed surgically was 4 mm in diameter. This seems to be the current limit of the employed technique. Smaller lesions than that would be impossible to distinguish from the heterogeneity of the normal pituitary tissue that is always present to some extent. CT scanning including the coronal and sagittal projections is an excellent technique for differentiating the empty sella from solid lesions. In questionable situations, CT scanning with metrizamide cisternography will decide. An artifactual appearance of empty sella can be created by less sophisticated equipment due to volume-averaging by the computer and the presence of a large volume “empty” space of the sphenoid sinus next to the sella turcica. This was not a problem in the present studies. In all our cases of primary empty sella, some soft tissue density could be identified inside the sella, adjacent to the contour of the bony structure, and frequently a pituitary stalk was visualized connecting this tissue with the hypothalamus. This has been described recently as the “infundibulum sign” [ 151. The character of the patients with primary empty sella was different from the classic description of a middle-aged, obese, multiparous, hypertensive woman
872
June 1982
The American Journal of Medicine
Volume 72
without any endocrine abnormalities [20]. Our patients were mostly young adults, normotensive and not necessarily obese, and almost all of them presented with hyperprolactinemia, galactorrhea and menstrual or sexual dysfunction. The changing image of the patient with empty sella is likely due to the fact that our series included patients referred primarily for endocrine abnormalities, and that the diagnosis of the empty sella is made frequently by CT scanning in the presence of normal size of the sella. Previously, such cases could not be diagnosed by conventional skull films. Disappointing results of CT scanning were obtained in our series in two groups of patients: those after surgical or radiation treatment of pituitary tumors and those with clinical abnormality but normal anatomic study. In the former group, it was often impossible to decide when there was a residual or recurrent tumor. In the latter group, the presence of underlying pathology was impossible to discover. Follow up of the clinical and laboratory findings and repeated CT scanning in both of these groups of patients remain, at the present, the only diagnostic means. The overall prevalence of pituitary abnormalities in our series was almost five times higher in females than in males. Of the newly diagnosed pituitary lesions, more than 60 percent were solid tumors, with microadenoma slightly more frequent than macroadenoma. In the remainder, almost 40 percent, of patients with endocrine abnormalities, the underlying pathology was the primary empty sella. In the females, microadenoma was most frequent, followed by the primary empty sella, and macroadenoma was present only in about 20 percent. In contrast, in males, macroadenoma was identified in about half of the patients. The most frequent endocrine abnormality in our series was the amenorrhea and/or galactorrhea syndrome in females and impotence in males, with associated hyperprolactinemia. It was present in 135 of the 170 patients (Table Ill). Ten of 170 patients had acromegaly. In 11 of 170, no hyperfunctioning syndrome could be identified, but all these patients had hypopituitarism. In addition to the diagnostic value of the hormonal assays, the combination of the hormonal patterns in our series had a differential diagnostic importance, as summarized in Table IV. In summary, both hormonal testing and CT scanning have an important role in the differential diagnosis of pituitary lesions, and they make the use of invasive techniques obsolete. ACKNOWLEDGMENT The word processing expertise of Ms. Susan Burns and Miss Barbara McIntyre is appreciated.
CT SCANNING FOR PITUITARY TUMORS-VALENTA
ET AL.
REFERENCES 1.
2.
3.
4.
5.
6.
7.
6.
9.
10.
11.
Syvertsen A, Haughton VM, Williams AL, Cusick JF: The computed tomographic appearance of the normal pituitary gland and pituitary microadenomas. Radiology 1979; 133: 365-391. Wolpert SM. Post KD, Biller BJ, Molitch ME: The value of computed tomography in evaluating patients with prolactinomas. Radiology 1979; 131: 117-l 19. Leeds NE, Naidich TP: Computerized tomography in the diagnosis of sellar and parasellar lesions. Semin Roentgen 1977; 12: 121-135. Naidich TP, Pinto RS, Kushner MJ, et al.: Evaluation of sellar and parasellar masses by computed tomography. Radiology 1976; 120: 91-99. Reich NE, Zelch JV, Alfidi RJ, Meaney TF, Duchesneau PM, Weinstein MA: Computed tomography in the detection of juxtasellar lesions. Radiology 1976; 116: 333-335. Drayer BP, Rosenbaum AE, Kennerdell JS, Robinson AG. Bank WO, Deeb ZL: Computed tomographic diagnosis of suprasellar masses by intrathecal enhancement. Radiology 1977; 123: 339-344. Wiggli U, Benz UF: Normal computed tomography anatomy of the suprasellar subarachnoid space. Radiology 1978; 128: 65-70. Fitz CR, Wortzman G, Harwood-Nash DC, Holgate RC, Barry JF, Boktt DW: Computed tomography in craniopharyngiomas. Radiology 1978; 127: 687-69 1. Post MID, David NJ, Glaser JS, Safran A: Pituitary apoplexy: diagnosis by computed tomography. Radiology 1980; 134: 665-670. Fitzpatrick D, Tolis G, M&any EE. Taylor S: Pituitary apoplexy. The importance of skull roentgenograrns and computerized tomography in diagnosis. JAMA 1980; 244: 59-61. Manelfe C, Louvet J-P: Computed tomography in diabetes
12.
13.
14.
15.
16.
17.
18.
19.
20. 21.
June 1982
insipidus. J Comput Assist Tomogr 1979; 3: 309-316. Besser GM, Ratcliffe JG, Kilborn JR, Ormston BJ, Hall R: Interaction between thyrotrophin, corticotrophin and growth hormone secretion in man. J Endocrinol 1971; 51: 699. Harsoulis P, Marshall JC, Kuhn SF, Burke CW, London DR, Fraser TR: Combined test for assessment of anterior pituitary function. Br Med J 1973; 4: 326-329. Voigt KH, Dahlen HG, Fehm HL: Simultaneous stimulation test for the anterior pituitary hormones. Horm Metab Res 1974; 6: 436-437. Haughton V, Rosenbaum AE, Williams AL, Drayer B: Recognizing the empty sella by CT: the infundibulum sign. Am J Roentgen01 1981; 136: 293-295. Valenta CJ, Elias AN: lntrasphenoidal arachnoidocele with amenorrhea-galactorrhea and a pituitary apoplexy-like syndrome. Am J Obstet Gynecol (in press). Burrow GN. Wortzman G, Rewcastle NB, Holgate RC. Kovacs K: Microadenomas of the pituitary and abnormal sella tomograms in an unselected autopsy series. N Engl J Med 1981; 304: 156-158. Haney AF, Kramer RS, Wiebe RH, Hamond CB: Hypothalamic-pituitary function and radiographic evaluation of women with hyperprolactinemia and an “empty” sella turcica. Am J Obstet Gynecol 1979; 134: 9 17-924. Hayman LN, Evans RA, Hinck VC: Rapid high dose (RHD) contrast computerized tomography of perisellar vessels. Radiology 1979; 131: 121-123. Jordan RM, Kendall JW, Kerber CW: The primary empty sella syndrome. Am J Med 1977; 62: 569-580. Corbey RS, Leguin RM, Rolland R: Hyperprolactinemia and secondary amenorrhea. In: Crosignani PG. Robyn C, eds. Prolactin and human reproduction. New York: Academic Press, 1977; 203-215.
The American
Journal of Medicine
Volume 72
873