Silent corticotroph adenomas: Further clinical and pathological observations

Silent Corticotroph Adenomas: Further Clinical and Pathological Observations JOHN A. LOPEZ, MD, B. K. KLEINSCHMIDT-DEMASTERS, MD, CHUN-I SZE, MD, WHITNEY W. WOODMANSEE, MD, AND KEVIN O. LILLEHEI, MD Adrenocorticotroph cell pituitary adenomas immunoreactive for adrenocorticotropic hormone (ACTH) but unassociated with preoperative signs of hypercortisolism constitute between 6% and 43% of all ACTH adenomas. Few large series have been published. At our referral center for pituitary diseases, we have encountered 12 patients with silent ACTH adenomas, none of whom exhibited definite clinical features of hypercortisolism preoperatively. Two patients presented with apoplexy, and in 2 patients preoperative neuroimaging studies mimicked craniopharyngioma. Pathological examination revealed 8 adenomas with variably basophilic cytoplasm (type I, including 1 each with coarse basophilic granules and Crooke’s hyaline change) and 4 with predominantly chromophobic cytoplasm (type II). Diffuse versus patchy (30% to 50% of cells) immunostaining best distinguished these 2 types; calcitonin staining was focal or negative in both. Two patients had unexpected postoperative courses consistent with acute cortisol insufficiency; 1 patient suffered from a severe flu-like illness, and the other had dizziness and was found to have a serum cortisol level of < 1.0 ␮g/dL. Both patients improved after

cortisol replacement followed by a slow taper. Another patient developed 2 separate pituitary adenomas, a silent ACTH adenoma followed by a pure prolactinoma resected months later. Clonality studies demonstrated that the 2 tumors had arisen from different clonal populations. These cases offer additional insights into clinical, neuroimaging, histological, and biological features of silent ACTH adenomas. Because 2 of these patients seemed to require postoperative cortisol supplementation that otherwise would not have been given, clinicians should be notified about ACTH immunostaining in adenomas from patients without preoperative diagnoses of Cushing’s disease, to optimize postoperative care. HUM PATHOL 35: 1137-1147. © 2004 Elsevier Inc. All rights reserved. Key words: Crooke’s hyaline, cortisol, adrenocorticoptropic hormone, adrenocorticotroph, pituitary adenoma. Abbreviations: ACTH, adrenocorticotropic hormone; EM, electron microscopy; FSH, follicle-stimulating hormone; LH, luteinizing hormone; MRI, magnetic resonance imaging; PAS, periodic acidSchiff; PCR, polymerase chain reaction.

Silent corticotroph cell pituitary adenomas are tumors that, by definition, are not associated with recognized clinical features of Cushing’s disease preoperatively. Laboratory data consistent with hypercortisolism must also be lacking in these patients. Unlike many types of pituitary adenomas in which the preoperative serum values predict, and often parallel, the immunohistochemical staining pattern found in the adenoma tissue, the finding of adrenocorticotropic hormone (ACTH) immunostaining in the tumor is unexpected, underscoring the role of the pathologist in making the diagnosis. Although some attempts have recently been made to develop extensive and sophisticated endocrinologic tests that will allow clinical diagnosis preoperatively, these tests are not currently used in practice.1 Horvath et al originally described 3 subtypes of silent adrenocorticotroph adenomas of human pituitary tumors in 1980,2,3 but further revised this classification schema in 1988 to remove the type III tumors from the silent ACTH adenoma spectrum.4 Silent ACTH type I tumors are basophilic adenomas with a sinusoidal pattern and variable, but usually strong,

ACTH staining in a large percentage of adenoma cells. The original authors underscored the need for electron microscopy (EM) for classifying ACTH adenomas into types I, II, and III. However, EM is not as necessary for diagnosing silent type I adenomas, because, from a morphological standpoint, silent type I tumors fully mimic functional ACTH adenomas. For silent type I ACTH adenomas, the diagnosis is made only after the clinician and pathologist involved in the case recognize the discordance between the absence of preoperative clinical and laboratory features referable to Cushing’s disease in the patient and the positive, strong immunoreactivity for ACTH in the postoperative biopsy specimen. Type II silent ACTH adenomas are chromophobic tumors on periodic acid-Schiff (PAS)– orange G stains with a diffuse or, less commonly, sinusoidal pattern and angular cells that closely mimic chromophobic functional ACTH adenomas. These tumors manifest more patchy immunoreactivity for ACTH. EM is not absolutely necessary for identifying this type either, although it often offers the additional insights of subtle differences between functioning chromophobic corticotroph cell adenomas and the morphologically similar silent type II chromophobic adenomas. Clinically, silent type I and II ACTH adenomas are macroadenomas that present in similar fashion to other endocrinologically inactive sellar masses, with visual field defects (in 61% of cases), headaches (in 50% of cases), and hypopituitarism (in 26% of cases).5 Because the patients manifest no clinical symptoms of hypercortisolism (Cushing’s disease), the clinician is unlikely to perform a dexamethasone-suppression test or other

From the Departments of Neurosurgery, Pathology, Neurology, and Medicine (Endocrinology), University of Colorado Health Science Center, Denver, CO. Accepted for publication April 22, 2004. Address correspondence and reprint requests to B. K. Kleinschmidt-DeMasters, MD, University of Colorado Health Sciences Center, Department of Pathology, B-216, Denver, CO 80262. 0046-8177/$—see front matter © 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2004.04.016

1137

HUMAN PATHOLOGY

Volume 35, No. 9 (September 2004)

screening tests for Cushing’s syndrome, such as a 24hour urinary cortisol excretion.5 These tumors have an unusually high incidence of infarction, leading to apoplexy.3,5 Some studies have found that these tumors behave more aggressively after recurrence but do not recur more frequently,6 whereas other studies emphasize a significant likelihood of recurrence.5,7 In either case, making the correct diagnosis of a silent ACTH usually affects postoperative clinical management, in terms of either closer follow-up intervals or possible postoperative adjuvant radiation therapy.7 Most of the literature on silent type I and II ACTH adenomas consists of case reports,7-16 although 3 excellent large series reporting 17 patients,5 28 patients,6 and 27 patients7 have recently appeared. Recent insights into the biology of these tumors have also been gained by studies studying expression of pro-opiomelanocortin gene expression in both functioning and silent ACTH adenomas.17,18 Over the past 10 years, in our practice we have encountered cases of silent ACTH pituitary adenomas with additional interesting clinical, neuroimaging, morphological, and biological features not highlighted in the aforementioned recent large series. Some of our patients with silent ACTH adenomas seemed to manifest subtle alterations in the ACTH– cortisol axis, resulting in low postoperative cortisol levels similar to those seen postoperatively in patients with known functioning ACTH adenomas. We chose to focus on the specific clinical details in these individually clinically challenging patients and provide clinical scenarios for the most complex patients in this report, paralleling what was done in the original 1980 study by Horvath et al.2 We also emphasize the morphological spectrum of silent type I ACTH adenomas, ranging from epithelioid cells to coarsely granulated cells to moderate-sized cells, with 1 case even showing striking Crooke’s hyaline changes.16 In type II adenomas, very different zones were noted morphologically and immunohistochemically within the tumors, suggesting postclonal modifications within subpopulations of adenoma cells in these neoplasms. In addition, we report double pituitary adenomas in 1 patient with both a silent ACTH adenoma and a prolactinoma and provide clonality assessment. We also attempt to gain additional insight into the biology of these tumors by assessing immunoreactivity for calcitonin. Horvath et al2 suggested that subsets of ACTHsecreting cells in the normal anterior pituitary gland co-secrete this hormone, in addition to ACTH, betalipotropin, and endorphins, and that assessment of staining for this product might further subcategorize silent ACTH adenomas, but to our knowledge such a study has not yet been published. We follow up on this suggestion from the 1980 study and test whether calcitonin staining is present in appreciable amounts in our type I or II tumors and whether it might distinguish subsets of silent ACTH adenomas beyond the current classification system.

MATERIALS AND METHODS Twelve patients with pituitary adenomas that showed significant amounts of immunostaining for ACTH but proved to have no clinical manifestations of Cushing’s disease were identified from the files of the Departments of Pathology and Neurosurgery, University of Colorado Health Sciences Center. Ten of the 12 patients were seen and operated on by the same neurosurgeon specializing in pituitary adenomas (K.O.L.), and clinical details of the cases were reviewed by a single endocrinologist with expertise in pituitary diseases (W.W.W.). Particular attention was paid to the clinical presentation and detailed laboratory testing in these cases, both preoperatively and in retrospect, because despite the fact that the Health Sciences Center is a referral center for patients with pituitary adenomas, the clinicians had no suspicion that the patients harbored ACTH-immunoreactive adenomas until notified of this by the pathologist. All tissue specimens were submitted in toto and fixed in 10% buffered formalin and in some cases in Bouin’s fixative for optimization of immunohistochemical staining. Tissues were processed routinely, cut into 5-␮-thick sections, and stained with hematoxylin and eosin, Gomori’s reticulin, and PAS-orange G histochemical stains. Immunohistochemical staining for anterior pituitary hormones was performed using the peroxidase–antiperoxidase complex technique with light hematoxylin counterstaining. Antisera and their sources included prolactin and ACTH (both polyclonal; Signet, Dedham, MA), growth hormone and thyroid-stimulating hormone (TSH) (polyclonal and monoclonal, respectively; both from Dako, Carpinteria, CA), follicle-stimulating hormone (FSH) and luteinizing hormone (LH) (both polyclonal; Ventana Medical Systems, Tucson, AZ), and alpha-subunit (monoclonal; AMAC, Westbrook, ME). Immunostaining for calcitonin (polyclonal; Dako) was undertaken on both formalin-fixed and Bouin’s-fixed material. Tissue fixed in glutaraldehyde and processed for electron microscopy was available in 1 case. Clonality analysis was performed with the clonality-X chromosome inactivation assay protocols used at the National Cancer Institute.19,20 In brief, paraffin-embedded tissue blocks from previous surgical material were obtained from the histology laboratory. To prepare the tissue for DNA extraction, 4-␮-thick tissue sections were cut, deparaffinized, and digested by proteinase K overnight at 37 °C, followed by whole-genome polymerase chain reaction (PCR) amplification using the human androgen receptor gene (HUMARA) CAG trinucleotide repeat microsatellite primers. Then 20 mL of PCR product was denatured and subjected to electrophoresis in 6% acrylamide gel. The experiments were run in replicate to verify results. Demographic data, light microscopy appearance, and results of histochemical and immunohistochemical staining are summarized in Table 1. A summary of clinical, light microscopy, and immunohistochemical details of cases is provided in the following sections, with more detailed clinical scenarios given for individual patients of greatest clinical interest.

RESULTS Patient Demographics and Clinical Features The study population comprised 9 females and 3 males ranging in age from 33 to 76 years (mean age, 50.1 years). Clinically, 11 of the 12 patients exhibited

1138

SILENT CORTICOTROPH ADENOMAS (Lopez et al)

TABLE 1. Morphological and Immunohistochemical Features ACTH Immunostaining; Percentage of Cells Staining and Pattern

Immunostaining for Other Hormones; Adenoma Type

Ill-defined sinusoidal pattern, moderately large cells with rounded cytoplasm, mixed basophilic/chromophobic Sinusoidal pattern, plump epithelioid cells, moderately basophilic

100%, strong; many cells showing perimeter staining along cell membranes 100%, strong, staining throughout cytoplasm

None, including negative FSH, LH (repeated); type I

3; 33, F

Sinusoidal pattern, plump epithelioid cells, strongly basophilic with coarse PASpositive granularity

100%, strong, staining throughout cytoplasm

Rare prolactin positive cells; type I

4; 76, F

Ill-defined sinusoidal pattern, moderately large cells with rounded cytoplasm, mixed basophilic/chromophobic Sinusoidal pattern, plump epithelioid cells, strongly basophilic

100%, strong, many cells showing perimeter staining along cell membranes 100%, strong, staining throughout cytoplasm

Rare prolactin positive cells, type I

Sinusoidal pattern, moderately large cells, weakly basophilic Sinusoidal pattern, moderately large cells with rounded cytoplasm, mixed basophilic/chromophobic Ill-defined sinusoidal pattern, plump epithelioid cells, strong basophilia at cell perimeter indicative of Crooke’s hyaline change Mixed sinusoidal/diffuse pattern, variably large cells, chromophobic with scattered weakly basophilic cells Diffuse pattern, moderatesized cells, chromophobic

100%, strong, staining throughout cytoplasm

Rare prolactin-positive cells; type I

100%, strong, mixed cytoplasmic and cell membrane staining

Rare prolactin-positive cells; type I

100%, strong, striking perimeter staining along cell membranes

None; type I

50%, strong, staining throughout cytoplasm

None, type II

50%, moderate, mixed cytoplasmic and cell membrane staining 30–50%, strong, staining throughout cytoplasm

Rare prolactin- and GHreactive cells; type II

Patient, Age (Years), Gender 1; 54, M

2; 36, F

5; 68, F

6; 40, F 7; 48, F

8; 48, F

9; 39, F

10; 49, F

Morphological Appearance

11; 56, M

Mixed sinusoidal/diffuse, moderate-sized cells, chromophobic with scattered weakly basophilic cells

12; 54, F

Mixed sinusoidal/diffuse, moderate sized cells, weakly basophilic

30% to 50%, strong, staining throughout cytoplasm

magnetic resonance imaging (MRI) evidence of macroadenomas. Only 1 patient presented with a microadenoma (patient 3), with her tumor found secondary to an apoplectic event. Of the 11 patients with macroad-

Rare prolactin positive cells; type I

Very rare prolactin positive cells; type I

Scattered prolactinreactive cells, with very rare individual cells immunoreactive for GH, TSH, and ASU; type II None; type II

Additional Neuroimaging and Histological Features

Two separate lesions by neuroimaging studies; separate prolactinoma arising from different clone by enzyme analysis Neuroimaging differential of adenoma versus Rathke cleft cyst; zones of cholesterol clefts indicative of old hemorrhage, corresponding to clinical apoplexy

Cystic/solid mass on neuroimaging studies thought to be craniopharyngioma, hemorrhagic contents with cholesterol cleft formation and focal calcification Cystic mass on neuroimaging studies

Majority of tissue showing necrosis and hemorrhage, corresponding to clinical apoplexy Cystic mass on neuroimaging studies

Cystic/solid mass on neuroimaging studies thought to be craniopharyngioma Cystic/solid mass on neuroimaging studies

enomas, 10 had findings referable to mass effect. Five patients presented with visual complaints (patients 4, 5, 8, 9, and 11), 1 patient presented with headache (patient 6), and 4 patients presented with preexisting hy-

1139

HUMAN PATHOLOGY

Volume 35, No. 9 (September 2004)

popituitarism (patients 1, 5, 6, and 10). The patient with a macroadenoma not presenting with symptoms of mass effect (patient 2) had a tumor co-secreting prolactin and presented with amenorrhea. Two of the patients (patients 3 and 8) had clinical apoplectic events, and 3 patients (patients 3, 5, and 8) exhibited MRI and pathological evidence of hemorrhage within the tumor. Endocrine evaluation of the 12 patients revealed 4 patients (patients 1, 2, 10, and 12) with normal 24-hour urine free cortisol levels, 1 patient (patient 6) with a normal morning serum cortisol level, and 4 patients (patients 7, 8, 9, and 11) with normal random serum cortisol levels. Only patient 5 had an elevated serum ACTH level, of 106 pg/mL (normal, 0 to 60 pg/mL), along with a high-normal random serum cortisol level. This patient had no clinical signs of Cushing’s disease. Two patients (patients 3 and 4) did not undergo evaluation of the pituitary–adrenal axis but did not exhibit any clinical signs of Cushing’s disease. Of particular note, 2 of the 12 patients (patients 1 and 2) demonstrated postoperative signs and symptoms consistent with adrenal insufficiency. The clinical courses of these 2 patients, as well as of 2 patients with apoplexy (patients 3 and 8), 1 patient with confusing preoperative imaging studies simulating craniopharyngioma (patient 5), and 1 patient with a confusing clinical history (patient 10), are described in more detail in the Clinical Synopses section. None of our 12 patients has demonstrated recurrence to date, although 6 cases have had less than a 2-year follow-up, making the postoperative time period insufficiently long for any meaningful assessment. Silent ACTH adenomas constituted 22% of all ACTHsecreting adenomas resected at our institution over the past 7 years. Light Microscopy Features and Immunohistochemical Staining Patterns Details of immunostaining patterns, including distribution of ACTH immunostaining and co-staining for additional anterior pituitary hormones, are given in

Table 1. Briefly, 8 cases were classifiable as type I, and 4 cases were classifiable as type II. The most frequent co-staining pattern was that of rare cells immunoreactive for prolactin, as seen in 6 of 8 type I adenomas and in 2 of 4 type II adenomas. Three cases (patients 3, 5, and 8) exhibited histological evidence of hemorrhage or cholesterol cleft formation; 2 of these had corresponding known clinical apoplexy. Calcitonin immunostaining results were available for 10 of the 12 cases. In 2 patients (patients 1 [type I adenoma] and 10 [type II adenoma]), a diffuse, illdefined blush staining was identified, predominantly in Bouin’s-fixed tissue. Very rare, weakly staining adenoma cells were seen in patients 4 (type I), 11 (type II), and 12 (type II). The 5 remaining cases (patients 2, 3, 4, 8, and 9) were completely negative for calcitonin in both formalin-fixed and Bouin’s-fixed specimens of the same tumor. Hence, differences in immunostaining could be discerned between type I and type II adenomas; calcitonin staining did not further distinguish these 2 types. Detailed Clinical Synopses of Illustrative Cases Patient 1

This 54-year-old man was diagnosed incidentally with a pituitary macroadenoma after a cervical MRI for neck and arm pain disclosed an incidental tumor. The patient had presented with erectile dysfunction 1 year earlier and was diagnosed with hypogonadism. This was attributed to primary testicular failure, because the low testosterone level was associated with elevated FSH and LH levels. The patient denied most symptoms of pituitary dysfunction except for a gradual increase in weight by 4 to 5 pounds annually for the past few years. When questioned, he noted a long-standing dorsal cervical fat accumulation that he attributed to body-building. On physical exam, the patient had a round face; a thick, short neck; dorsal cervical fat accumulation; and a weight of 258 pounds and height of 5 feet, 11 inches.

™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 FIGURE 1. Photomicrograph of the classic epithelioid type I silent ACTH adenoma, with a focal papillary-like arrangement around abundant blood vessels. This type is morphologically indistinguishable from functioning ACTH adenomas. (B) Patient 2. Immunostaining for ACTH from the same tumor shown in (A) demonstrated strong, diffuse reactivity in nearly 100% of cells; staining extends throughout the cytoplasm, sparing only the faintly blue nucleus. (Hematoxylin and eosin, original magnification ⫻ 200.) (C) Patient 2. Immunostaining for prolactin from the same tumor shown in (A) and (B) showed rare reactive cells, often smaller and less epithelioid than the dominant adenoma population. (Peroxidase-antiperoxidase technique for ACTH with light hematoxylin counterstaining; original magnification ⫻ 200.) (D) Patient 3. Type I adenoma with coarse, PAS-positive granules stuffing the cytoplasm. (Peroxidase-antiperoxidase technique for prolactin with light hematoxylin counterstaining; original magnification ⫻ 400.) (E) Patient 3. The same patient illustrated in (D) also had a zone of fibrosis and cholesterol clefts within her tumor, tombstones of a previous apoplectic episode. (PAS– orange G stain; original magnification ⫻ 1000.) (F) Patient 3. Another unusual pattern seen in a type I adenoma was Crooke’s hyaline change, with PAS-positive granules (upper right) and ACTH immunostaining (lower left) peripherally located at the cell membranes. (Hematoxylin and eosin; original magnification ⫻ 200.) (G) Patient 9. Type II adenoma showing a focal sinusoidal pattern and larger cells (top of photograph) in one area of the adenoma juxtaposed to a diffuse, patternless area with small cells (lower part of photograph); this morphological variability suggests postclonal modifications in the tumor. (PAS– orange G stain and peroxidase-antiperoxidase technique for ACTH with light hematoxylin counterstaining; original magnification ⫻ 400.) (H) Patient 9. The same area of the type II adenoma shown in (G) demonstrates variable, patchy immunoreactivity for ACTH, with more cells staining in the sinusoidal area (top). (Hematoxylin and eosin; original magnification ⫻ 200.) (I) Patient 10. Another type II adenoma, illustrated at low magnification, demonstrates that although patchy ACTH immunostaining was present in many areas, extreme variability in staining also could be seen from one tissue fragment to another. (Peroxidase-antiperoxidase technique for ACTH with light hematoxylin counterstaining; original magnification ⫻ 200.)

1140

SILENT CORTICOTROPH ADENOMAS (Lopez et al)

1141

HUMAN PATHOLOGY

Volume 35, No. 9 (September 2004)

Visual fields were full, and fundoscopic examination was unremarkable. He had dry skin and visible striae in his axilla and upper arms, presumably due to his past body-building activities, but no abdominal striae. Initial pituitary laboratory values were normal except for a slightly low total T3 level of 88 ng/dL (normal, 90 to 190 ng/dL), low testosterone level of 93 ng/dL (normal, 200 to 850 ng/dL), and slightly high LH level of 9.1 ␮U/mL (normal, 1.0 to 9.0 mU/mL) and FSH level of 34 ␮U/mL (normal, 2 to 18 mU/mL). A random serum cortisol level was 10 ␮g/dl. A 24-hour urine free cortisol level was normal at 64 ␮g/24 hr (normal range, 10 to 90 ␮g/24 hr). MRI scan of the pituitary revealed a well-circumscribed, 1.3 cm ⫻ 1.4 cm intrasellar mass with a cystic component. The initial endocrine assessment was a macroadenoma, most likely a gonadotrophic cell adenoma, given the patient’s high FSH and LH levels in association with low testosterone level. The patient subsequently underwent transsphenoidal resection of the pituitary adenoma and was discharged from the hospital 2 days later with a rapid dexamethasone taper that ended on postoperative day 3. Pathological analysis of the tumor revealed a basophilic adenoma with immunoreactivity only for ACTH. FSH and LH immunostaining were repeated twice and were negative. The pathologist promptly notified the neurosurgeon and endocrinologist of the findings. The endocrinologist subsequently contacted the patient to check on his clinical status. The patient reported feeling well but did note mild dizziness. He was asked to come into clinic for endocrine laboratory studies; these revealed an undetectable serum cortisol level (⬍1.0 ␮g/dL) on postoperative day 5. He was started on prednisone 5 mg/day, with subsequent resolution of his symptoms. Before his postoperative endocrine follow-up visit, the patient self-discontinued his prednisone with no apparent adverse effects. Patient 2

This 36-year-old female was first diagnosed with a pituitary adenoma in October 2000, when she presented with an elevated prolactin level of 74 ng/dL (normal, 2 to 15 ng/dL) and amenorrhea. Physical examination revealed no features of Cushing’s disease, but the patient did complain of an annual weight gain of 10 pounds over the previous 6 years. Screening for possible Cushing’s syndrome revealed a normal 24hour urine free cortisol level of 70 ␮g/24 hr. MRI revealed bilateral pituitary adenomas: a solid rightsided mass, 0.9 ⫻ 1.6 cm, and a cystic left-sided tumor, 1.2 ⫻ 1.1 cm. A trial of the dopamine agonist cabergoline was initiated. Although her prolactin level responded to this therapy, the patient did not tolerate the medication. A repeat MRI showed that the right-sided mass was essentially unchanged, whereas the mass on the left had significantly decreased in size to 0.98 ⫻ 0.6 cm. Invasion of the right cavernous sinus by the rightsided mass was now noted. Surgery was undertaken to remove the right-sided tumor and to explore the leftsided abnormality. Intraoperatively, the tumor mass on

the right side was identified immediately, and a gross total resection was achieved. However, an extensive search conducted to find the cystic region on the left side of the gland proved negative. The pathology showed a basophilic pituitary adenoma with a sinusoidal pattern and focal papillary appearance (Fig 1A), strongly and diffusely immunoreactive for ACTH (Fig 1B), with rare cells immunoreactive for prolactin within the adenoma (Fig 1C). Five days postoperatively, the patient experienced apparent chemical meningitis with low back pain, a fever of 101°F, and lower extremity weakness. She was started on steroid replacement therapy, after which her symptoms dramatically improved. She was then tapered off the steroids, and her symptoms recurred, but to a lesser extent. Fatigue was the overriding symptom at this time. Dexamethasone was restarted after reviewing her pathology report, and this again resulted in significant clinical improvement. Her symptoms were attributed to steroid withdrawal after treatment for Cushing’s disease. Approximately 1 month after surgery, her morning cortisol level was low, at 1 ␮g/dL (after holding dexamethasone), and her prolactin level remained elevated, at 44 ng/mL. She continued to complain of fatigue and amenorrhea and was changed to prednisone. Seven months after the patient’s initial surgery, her pituitary–adrenal axis appeared to be recovering, as documented by corticotropin-stimulation test results, and thus her prednisone dosage was tapered. Her amenorrhea and hyperprolactinemia persisted. Due to her intolerance of dopamine agonists and a continued abnormal lesion in the left pituitary, she underwent repeat transsphenoidal pituitary exploration. A left-sided pituitary adenoma was removed and revealed strongly positive prolactin immunoreactivity. Clonality analyses showed that the 2 tumors removed at the 2 separate surgical procedures had arisen from entirely separate clones, based on HUMARA assays (Fig 2). Patient 3

In 2000, this 33-year-old woman was diagnosed with supraventricular tachycardia and underwent 3 ablative procedures via cardiac catheterization. One procedure was complicated by complete heart block, which was subsequently treated with placement of a pacemaker. During hospitalization for heart block in 2001, she developed an acute onset of right-sided weakness, which was thought to be due to cardiogenic emboli. MRI revealed a hemorrhagic pituitary mass, and a follow-up scan 6 weeks later showed that the mass had possibly enlarged. On review of systems, the patient reported galactorrhea, amenorrhea, easy bruising, and a 50- to 60pound weight gain since late 2000. Her weight gain was responsive to dieting. Her past medical history was also remarkable for the onset of idiopathic diabetes insipidus 10 years earlier, which was treated with desmopressin, and seizures since 1999, which was treated with topiramate. She had delivered a normal child in 1997

1142

SILENT CORTICOTROPH ADENOMAS (Lopez et al)

Rathke’s cleft cyst. The patient underwent transsphenoidal surgery in January 2002. The pathology was a basophilic adenoma with coarse basophilic granules (Fig 1D). The tumor was immunoreactive for ACTH, with rare cells immunoreactive for prolactin. Focally, the adenoma demonstrated fibrosis and striking cholesterol cleft formation (Fig 1E), consistent with remote hemorrhage and paralleling her initial scan that had shown hemorrhage in the tumor. Postoperatively, she continued to feel poor and self-discontinued her steroids. Approximately 10 weeks postoperatively, she had a random serum cortisol level (off steroids) of 13 ␮g/dL (normal, 2 to 16 ␮g/dL), suggesting she was not suffering from steroid withdrawal or adrenal insufficiency. The patient was subsequently lost to follow-up. Patient 5

FIGURE 2. Gel illustrating that the 2 adenomas removed months apart at 2 separate surgeries on patient 2 were derived from different clones. The amplified PCR products are stained with ethidium bromide; lane 1 shows a 100-bp DNA ladder, with lane 2 representing a Hec/A cell line control. Lanes 3 and 4 represent PCR products from the adenoma treated with (lane 3) or without (lane 4) Hha-1 restriction enzyme digestion. Both lanes show 2 slightly diffused bands (arrow). In this method, clonal cell populations will show “loss” of the nonmethylated allele after restriction enzyme digestion. In this experiment, PCR products from the patients showed 2 bands and were interpreted as polycolonal.

and breast-fed until 1999. Prolactin levels in 2001 were elevated on multiple occasions and ranged between 25 and 30 ng/dL. She had been given a 1-month trial of bromocriptine in August 2001, but suffered significant nausea and vomiting, without resolution of the amenorrhea. Several endocrine evaluations at multiple institutions revealed elevated prolactin levels with otherwise normal endocrine laboratory study and neurologic exam findings. Preoperative screening for Cushing’s disease was not performed, because the patient did not exhibit any clinical signs or symptoms of hypercortisolemia. Serial brain imaging studies showed that the pituitary lesion had enlarged from 0.5 cm to 0.9 cm. The preoperative diagnoses were pituitary adenoma versus

This 68-year-old female first developed a change in her personality in the spring of 1997 soon after an abdominal hernia repair. MRI of her brain in April 1998 revealed a partially cystic/solid suprasellar and intrasellar lesion consistent with either a craniopharyngioma or a pituitary adenoma. Visual field examination revealed bitemporal visual field deficits. Her past medical history was significant only for primary hypothyroidism, sigmoid polyposis, and diverticulitis. An endocrine workup revealed hyperprolactinemia (144 ng/ mL), an elevated ACTH level of 106 pg/mL (normal, 0 to 60 pg/mL), and a high-normal cortisol level. Her general physical exam was normal. She had no clinical signs of Cushing’s disease, and so no further screening for hypercortisolemia was performed. Her preoperative diagnosis was craniopharyngioma versus a prolactinoma. She underwent transsphenoidal surgery in August 1998, with gross total resection of her pituitary mass. Intraoperatively, “machine oil”–like contents yielded the impression of a craniopharyngioma. Pathology revealed a strongly basophilic pituitary adenoma with ACTH immunoreactivity and rare cells immunoreactive for prolactin. Significant portions of the adenoma were admixed with hemorrhage and cholesterol clefts, and focal calcification was also found in the adenoma. She did well postoperatively and was tapered off steroids within 4 weeks without clinical deterioration. Patient 8

This 48-year-old female presented with the sudden onset of left retro-orbital eye pain and blurred vision. Her past medical history was significant for hypertension, hypercholesterolemia, shortness of breath on exertion, and cessation of menstruation in the past year. Although the pain resolved fairly quickly, her visual difficulties worsened over 48 hours when she did not immediately seek medical attention. At that point she consulted an ophthalmologist, who noted a visual field cut and ordered neuroimaging studies. Ten days after the onset of problems, the patient saw a neurosurgeon, who diagnosed a pituitary adenoma and obtained en-

1143

HUMAN PATHOLOGY

Volume 35, No. 9 (September 2004)

docrinology studies, which were normal. By the time of the patient’s visit with the neurosurgeon, her vision had improved, and transsphenoidal resection was electively undertaken. Intraoperatively, hemorrhagic and necrotic debris was encountered. The adenoma was mostly necrotic and had foci of resolving hemorrhage; cytologically, the cells exhibited eye-catching Crooke’s hyaline changes with strong basophilia and ACTH immunostaining at the cell perimeters (Fig 1F). The patient’s postoperative recovery was uneventful. Patient 10

This 49-year-old female with a history of primary hypothyroidism first noted intermittent tingling in her left arm and leg 18 months before her diagnosis. Electromyography and nerve conduction velocity study results were normal. A head MRI was subsequently obtained in the spring of 1990 that was remarkable only for some white matter abnormalities of unclear significance. She was started on a course of physical therapy, and the tingling in her arm and leg improved. She experienced a seizure in late July of 1990, consisting of stiffening of her left leg and flexion of her left arm, with some momentary loss of consciousness. She was hospitalized. An electroencephalogram was consistent with a generalized seizure disorder, and she was started on antiepileptic medication. In the fall of 1990, a follow-up MRI scan demonstrated an enlarged sella turcica with an intrasellar and suprasellar mass. An endocrine workup revealed an elevated T4 level of 16.2 ␮g/dL (normal, 4.9 to 10.7 ␮g/dL), a low T3 resin uptake of 26.4% (normal, 28% to 36%), a thyroid-stimulating hormone level of 3.7, an FSH level of 16.9, an LH level of 38, a prolactin level of 18.1 ng/mL (normal, 3.8 to 23.2 ng/mL), a cortisol level of 12.0 ␮g/dL (normal, 5 to 25 ␮g/dL), and a urine free cortisol level ⬍5 ␮g/24 hr. Her corticotropin stimulation test was normal (ie, stimulation cortisol rose from 18.0 to 38.3 at 30 minutes). The patient underwent transsphenoidal resection of her pituitary tumor in April 1991. The pathology revealed a chromophobic adenoma with ACTH immunoreactivity in 50% of cells in an irregular patchy pattern of distribution, similar to that seen in patient 8 (Fig 1H). Focally, however, there were significant differences from one adenoma tissue fragment to the next, with areas showing ⬎90% of cells immunoreactive and others showing ⬍25% of cells immunoreactive (Fig 1I). Additional rare cells showed immunoreactivity for prolactin and growth hormone, although the latter finding was equivocal. Electron microscope examination revealed elongated, polygonal, angular tumor cells with abundant mitochondria, small amounts of rough endoplasmic reticulum, and Golgi complexes. The cytoplasm was sparsely populated by spherical to teardrop-shaped secretory granules 100 to 430 nm in size (Fig 3). Cytoplasmic microfilaments were not discernable, consistent with a silent corticotroph cell adenoma, type II. Her postoperative 24-hour urine free cortisol

FIGURE 3. Electron microscopic examination of patient 10 showed elongate polygonal, angular tumor cells with abundant mitochondria, small amounts of rough endoplasmic reticulum, and Golgi complexes. The cytoplasm was sparsely populated by spherical to tear-dropped shaped secretory granules 100 to 430 nanometers in size (inset), marginating near the cell membrane. Cytoplasmic microfilaments were not discernable, consistent with a silent corticotroph cell adenoma, type II. (Original magnification ⫻ 3000; inset, ⫻ 15,000.)

level was normal. The patient did well postoperatively, but her seizure disorder persisted. DISCUSSION Silent ACTH adenomas are interesting tumor types that have been postulated to either secrete structurally abnormal ACTH that is endocrinologically inactive but detectable by immunohistochemistry5 or radioimmunoassay,12 or to secrete ACTH intermittently21 or at low levels continuously. Defective intracellular processing of the ACTH molecule has also been suggested.22 Given the few reports in the literature on silent ACTH adenomas, it might be anticipated that additional insights will emerge as more centers gain experience with larger numbers of these patients. We encountered clinical, biological, and morphological features in our 12 patients with silent ACTH adenomas, type I (8 patients) and type II (4 patients), that have not been previously emphasized. First, although the tumors could be grouped into silent types I and II, based on published criteria that emphasize architectural pattern and basophilic versus chromophobic staining,2,3,5 the morphological range was more broad than previously described in the larger series. Most type I adenomas, as expected, were cases with large epithelioid cells identical to functioning densely granulated ACTH adenomas and columnar or papillary-like sinusoidal patterns (Fig 1A, B, and C). However, we also noted a case with exceptionally coarse PAS-positive granules (Fig 1D) and a case with Crooke’s hyaline changes (Fig 1F). Although the latter was not described in the 3 recent large series,5,6,7 2 silent ACTH

1144

SILENT CORTICOTROPH ADENOMAS (Lopez et al)

adenomas with Crooke’s hyaline change have recently been reported.16 The degree of basophilia of the tumors cells with PAS– orange G stain is also broad, suggesting that the diffuse and ubiquitous ACTH immunostaining pattern of type I (Fig 1B) versus patchy and partial pattern of ACTH immunostaining of type II silent adenomas (Fig 1H and I) is an easier and more reproducible criterion for subclassification than EM. Our findings corroborate the findings of Bradley et al,6 in which tumors were not specifically subtyped by EM, but rather were classified based on extent of diffuse or patchy ACTH immunostaining. Most of our silent type I adenomas were also associated with rare cells exhibiting prolactin immunoreactivity, although in some instances the cells appeared morphologically distinct and not identical to the adenoma cells (Fig 1C). In no instance did the numbers of prolactin-immunoreactive cells within the adenoma appear to be sufficient to account for the elevated serum levels of prolactin seen in 7 of our 12 patients. We favor the explanation for clinical hyperprolactinemia in these patients offered by Scheithauer et al,5 that there may be “inhibition of dopamine secretion by tumoral endorphin production.” Silent ACTH adenoma tumors apparently may even induce the production of a second population of autonomous prolactin-secreting cells through this endorphin production, as evidenced by our finding of 2 separate pituitary adenomas in our patient 2. These 2 tumors were found to arise from different clones. The first tumor was a type I basophilic adenoma with strong diffuse ACTH immunostaining in nearly 100% of cells (Fig 1A and B) and rare scattered cells immunoreactive for prolactin (Fig 1C). The tumor resected later was a pure prolactinoma with classic features. Double pituitary adenomas associated with silent ACTH adenomas is a novel finding of the present study. The combination of these 2 specific hormone types (ACTH and prolactin) as a double pituitary adenoma is not likely due to chance. Despite the fact that double pituitary adenomas are rare in surgical practices, a recent report by Meij et al23 identified 3 patients with double ACTH- and prolactin-secreting pituitary adenomas. Both their cases 1 and 3 had clinically overt Cushing’s disease, and the prolactin cell adenoma was silent in their case 3. Their case 2 had nodular corticotroph hyperplasia as the cause of her Cushing’s disease and also a silent prolactin cell adenoma. Clonal analyses were not undertaken in that study, but one might surmise that these tumors were also due to separate clones. The results of their study, taken in conjunction with the findings of double adenomas in our patient 2, suggest that either these 2 tumor types rarely induce the formation of one another or share common instigating events. Either the ACTH or the prolactin cell adenoma may be clinically silent. We also noted interesting morphological features in the type II adenomas in our series. In contrast to the 100% immunoreactivity for ACTH in the type I adenomas, all 4 type II cases demonstrated 30% to 50% of

cells immunoreactive for ACTH. In several cases, morphologically different-appearing areas (Fig 1G) and irregular immunostaining (Fig 1H and I) suggested probable postclonal modifications within subsets of cells in the adenoma. Postclonal modification may account for the reports in the literature of clinically silent ACTH adenomas evolving over time into clinically active tumors producing Cushing’s syndrome.22 The findings of clinical or subclinical apoplexy in these tumors have been well documented by others,5,7,15 and we also identified neuroimaging and pathological features of hemorrhage or infarction in our tumors, including the formation of cholesterol clefts (Fig 1E). Additional neuroimaging feature noted in our series were several cases in which preoperative neuroimaging features mimicked those of craniopharyngiomas or Rathke’s cleft cyst (Table 1). The most important clinical finding in our study, however, was that in several of our patients with silent ACTH adenomas, the ACTH– cortisol axis appeared to be dysfunctional. Even though there appeared to be a lack of functional ACTH secretion or Cushing’s disease preoperatively, removal of the tumor seemed to result in postoperative low cortisol levels. The features in several of our patients strengthen the argument that partially active or intermittently secreted hormone product is produced in these clinically silent ACTH adenomas. Our patient 2 suffered a well-documented cortisol collapse after removal of her adenoma, despite the fact that she exhibited no symptoms of hypercortisolism preoperatively, except possibly a 6-year history of weight gain. Due to this weight gain, she had in fact undergone a preoperative 24-hour urine cortisol test that was in the normal range, as was her serum cortisol level. Because the diagnosis of Cushing’s disease was not made preoperatively, postoperatively she was not treated as if she had suppressed pituitary corticotroph function. Patients with known preoperative Cushing’s disease are given higher doses of corticosteroids over longer periods after surgery and usually are maintained for a time on steroids after hospital discharge. Patient 2 was not given this therapy and suffered severe postoperative symptoms, ameliorated only by steroid replacement. To our knowledge, only 1 study has mentioned postoperative impairment of the pituitary–adrenal axis in clinically silent ACTH adenomas requiring glucocorticoid replacement.13 Patient 2 was an index case for us and has prompted pathologists at our institution to notify the neurosurgeon and endocrinologist whenever the diagnosis of a clinically unsuspected corticotroph cell adenoma is made. A second patient (patient 1) may also have been in the initial phases of cortisol collapse postoperatively when, in a phone call several days after surgery, he disclosed severe fatigue. The phone call was placed as a direct result of the neuropathologist informing the clinician about the diagnosis of a silent ACTH. Prompt medical intervention with steroid replacement was instituted before the patient became seriously ill. Patient 1 had exhibited a buffalo hump preoperatively, but he

1145

HUMAN PATHOLOGY

Volume 35, No. 9 (September 2004)

did not meet the criteria for Cushing’s disease. His serum cortisol level was low (⬍1.0 ␮g/dL) on postoperative day 5. It is standard practice at our institution to administer a short (3-day) decadron taper perioperatively in patients with nonfunctional adenomas, with discontinuation by the time of discharge. This rapid taper would be insufficient for a patient with known Cushing’s disease and known suppression of the pituitary–adrenal axis. Given this patient’s silent ACTH adenoma, his low postoperative serum cortisol level may also have been a direct consequence of the removal of his ACTH tumor, similar to the scenario noted earlier for patient 2. Patient 3’s clinical history illustrates that silent ACTH adenomas may show differences in the level of ACTH secretion over time. Although this patient did not manifest Cushing’s disease at the time of her examination, in retrospect she may have had Cushing’s disease for a brief period earlier in her complicated course when she gained 60 pounds over a 6-week period. Furthermore, she may have “cured” herself after her apoplectic event, similar to other reported cases in the literature of patients whose cushingoid features remitted after apoplexy and spontaneous tissue destruction.10 Identification of Cushing’s disease can be clinically challenging. It is important to emphasize that none of the patients in our series met the clinical criteria for Cushing’s disease, despite careful examination by experienced observers at an academic referral center for pituitary disorders. In 1 patient, 1 of 2 clinicians suspected Cushing’s disease, but the clinician negated the diagnosis when she found normal cortisol levels in that patient. Eleven of 12 patients had had endocrine laboratory study results that were inconsistent with hypercortisolism preoperatively. In addition, 3 patients had normal preoperative 24-hour urine free cortisol levels. One patient had had an elevated ACTH level, but because the neuroimaging studies were confusing and the patient was not cushingoid in appearance, this elevated level was attributed to stress. Although all patients met the definition in the literature of clinically silent ACTH adenomas, and many had laboratory documentation of normal cortisol level, nevertheless the diagnosis of Cushing’s disease cannot be biochemically excluded in any patient without extensive preoperative dexamethasone-suppression testing.5,6 However, this testing is, for obvious reasons, almost never conducted on patients without a clinical suspicion of Cushing’s disease and was not performed on any of our 12 cases. The impetus for correct diagnosis of silent ACTH adenoma by the pathologist not only is related to treatment in the immediate postoperative period with prompt steroid replacement, but also has consequences for long-term therapy. Scheithauer et al5 suggested that silent corticotroph adenomas behave aggressively and have a high recurrence rate, underscoring the need for long-term follow-up and consideration for postoperative radiotherapy. Webb et al7 found that 37% of their patients experienced recurrence (median follow-up of 60 months), with a 41.7% recurrence rate among those

who had not received postoperative radiotherapy. However, Bradley et al6 concluded that silent corticotroph adenomas do not recur more frequently, but do behave aggressively if they do recur. Our follow-up time is too short to allow us to corroborate or refute these findings from other studies. Based on the findings of Webb et al,7 postoperative radiation therapy should be strongly considered in patients with silent ACTH adenomas. In contrast, other types of clinically silent pituitary adenomas often manifest more indolent growth patterns, and postoperative radiation therapy can often be avoided, as we have shown previously.24 In conclusion, we have reported on 12 patients with silent ACTH adenomas and variable, sometimes complex, clinical symptoms. In these tumors, preoperative clinical and laboratory findings either are confusing for or negate the diagnosis of an ACTH adenoma, even for experienced clinicians at referral centers for pituitary diseases. Neuroimaging studies may suggest a diagnosis of craniopharyngioma rather than adenoma, even to experienced neuroradiologists. The diagnosis of silent corticotroph cell adenoma can be made only by a pathologist who combines knowledge of the immunostaining results with corroborating clinical information. Whereas immunostaining for pituitary hormones has long been recommended for all pituitary adenomas, it is essential for making the diagnosis in silent ACTH adenomas. Extensive immunostaining of tumors is admittedly costly and, especially for smaller hospitals, may be prohibitive. It may be tempting for the pathologist to conduct immunostaining only in cases of endocrinologically active pituitary adenomas associated with known serum abnormalities on preoperative endocrinologic testing. In fact, the most important cases on which to use immunostaining may be endocrinologically silent pituitary adenomas, such as these silent ACTH adenomas. Correct diagnosis is especially important for the uncommon patient who develops cortisol insufficiency in the immediate postoperative period due to partial or intermittent suppression of the pituitary–adrenal axis. Indeed, had the pathologist not alerted the clinicians to the diagnosis of an ACTH immunoreactive tumor in a prompt manner, some of our patients (ie, patients 1 and 2) might not have received optimal postoperative cortisol replacement therapy.21 Acknowledgment. We thank Drs. S. Levy and A. Prall from Intermountain Neurosurgery, Denver, CO, for referring patients 6 and 9 to us. We also gratefully acknowledge the photographic expertise of Lisa Litzenberger, the typing of Andrew Rex, the proofreading of Cindy McNair, and the histotechnology assistance of David Davis.

REFERENCES 1. Ambrosi B, Colombo P, Bochicchio D, et al: The silent corticotropinoma: Is clinical diagnosis possible? J Endocrinol Invest 15: 443-452, 1992 2. Horvath E, Kovacs K, Killinger DW, et al: Silent corticotropic adenomas of the human pituitary gland: A histologic, immunocytologic, and ultrastructural study. Am J Pathol 98:617-638, 1980

1146

SILENT CORTICOTROPH ADENOMAS (Lopez et al) 3. Kovacs K, Horvath E: Tumors of the Pituitary Gland. Washington, DC, Armed Forces Institute of Pathology, 1986, p 157 4. Horvath E, Kovacs K, Smyth HS, et al: A novel type of pituitary adenoma: Morphological features and clinical correlations. J Clin Endocrinol Metab 66:1111-1118, 1988 5. Scheithauer BW, Jaap AJ, Horvath E, et al: Clinically silent corticotroph tumors of the pituitary gland. Neurosurgery 47:723-729, 2000 6. Bradley KJ, Wass JA, Turner HE: Non-functioning pituitary adenomas with positive immunoreactivity for ACTH behave more aggressively than ACTH immunonegative tumours but do not recur more frequently. Clin Endocrinol (Oxf) 58:59-64, 2003 7. Webb KM, Laurent JL, Okonkwo DO, et al: Clinical characteristics of silent corticotrophic adenomas and creation of an internet-accessible database to facilitate their multi-institutional study. Neurosurgery 53:1076-1085, 2003 8. Ueyama T, Tamaki N, Kondoh T, et al: Large and invasive silent corticotroph-cell adenoma with elevated serum ACTH: A case report. Surg Neurol 50:30-31, 1998 9. Braithwaite SS, Clasen RA, D’Anglelo CM: Silent corticotroph adenoma: Case report and literature review. Endocrinol Pract 3:297301, 1997 10. Nagaya T, Seo H, Kuwayama A, et al: Pro-opiomelanocortin gene expression in silent corticotroph-cell adenoma and Cushing’s disease. J Neurosurg 72:262-267, 1990 11. Dickstein G, Arad E, Shechner C: Late complications in remission from Cushing disease. Recurrence of tumor with reinfarction or transformation into a silent adenoma. Arch Intern Med 157:2377-2380, 1997 12. Ueyama T, Tamaki N, Kondoh T, et al: Large and invasive silent corticotroph-cell adenoma with elevated serum ACTH: A case report. Surg Neurol 50:30-31, 1998 13. Sakaguchi H, Koshiyama H, Sano T, et al: A case of nonfunctioning pituitary adenoma resembling so-called silent corticotroph adenoma. Endocr J 44:329-333, 1997 14. Sano T, Kovacs K, Asa SL, et al: Pituitary adenoma with

“honeycomb Golgi” appearance showing a phenotypic change at recurrence from clinically nonfunctioning to typical Cushing disease. Endocr Pathol 13:125-130, 2002 15. Abe M, Sawabe Y, Mochizuki Y, et al: Corticotroph cell adenoma without typical manifestations of Cushing’s disease presenting with cavernous sinus syndrome following pituitary apoplexy. Endocr J 48:503-507, 2001 16. Roncaroli F, Faustini-Fustini M, Mauri F, et al: Crooke’s hyalinization in silent corticotroph adenoma: Report of two cases. Endocr Pathol 13:245-249, 2002 17. Lloyd RV, Fields K, Long J, et al: Analysis of endocrine active and clinically silent corticotropic adenomas by in situ hybridization. Am J Pathol 137:479-488, 1990 18. Stefaneanu L, Kovacs K, Horvath E, et al: In situ hybridization of pro-opiomelanocortin (POMC) gene expression in human pituitary corticotrophs and their adenomas. Virchows Arch A 419: 107-113, 1991 19. Allen RC, Zoghbi HY, Moseley AB, et al: Methylation of Hpall and Hhal sites near the polymorphic CAG repeat in the human androgen-receptor gene correlates with X chromosome inactivation. Am J Hum Genet 51:1229-1239, 1992 20. Enomoto T, Fujita M, Inoue M, et al: Analysis of clonality by amplification of short tandem repeats. Carcinomas of female reproductive tract. Diag Mol Pathol 3:292-297, 1994 21. Atkinson AB, Chestnutt A, Crothers E, et al: Cyclical Cushing’s disease: Two distinct rhythms in a patient with a basophil adenoma. J Clin Endocrinol Metab 60:328-332, 1985 22. Tan EU, Ho MSJ, Rajasoorya CR: Metamorphosis of a nonfunctioning pituitary adenoma to Cushing’s disease. Pituitary 3:117122, 2000 23. Meij BP, Lopes MB, Vance ML, et al: Double pituitary lesions in three patients with Cushing’s disease. Pituitary 3:159-168, 2000 24. Lillehei KO, Kirschman DL, Kleinschmidt-DeMasters BK, et al: Reassessment of the role of radiation therapy in the treatment of endocrine inactive pituitary macroadenomas. Neurosurgery 43:432439, 1998

1147