Androgen-secreting adrenal tumors

Fernando Cordera, MD, Clive Grant, MD, Jon van Heerden, MD, Geoffrey Thompson, MD, and William Young, MD, Rochester, Minn

Background. Adrenal tumors that secrete androgens exclusively are extraordinarily rare. The aim of this study was to characterize patients with pure androgen-secreting adrenal tumors.

Methods. A retrospective chart review from January 1946 through November 2002 identified 11 female patients with pure androgen-secreting adrenal tumors.

Results. The mean age was 23.4 years (range, 1-52). The most common presenting symptoms were hirsutism, acne, and clitoral enlargement. Elevated 17-ketosteroids were found in seven of nine tested patients. Computed tomogram, ultrasound, or both localized tumors in six of seven patients. All tumors were surgically resected, one laparoscopically, all without complications. Five of the 11 tumors were malignant. Mean weight and mean maximal diameter for benign and malignant tumors were 44 g and 4.2 cm and 232 g and 9.8 cm, respectively. Mean hospital stay was 8.5 days, with excess androgen production resolved in all patients. Recurrence and disease-related death occurred in only one patient who had pulmonary metastases at diagnosis. The remaining patients had no recurrence of tumor at mean follow-up of 11.7 years (range, 0.5-32 years).

Conclusions. Pure androgen-producing tumors are extremely rare. Approximately 50% are benign, and surgical resection provides excellent treatment if the tumors are not metastatic at the time of diagnosis. (Surgery 2003;134:874-80.)

From the Department of Surgery and Internal Medicine, Mayo Clinic, Rochester, Minn

IN THE ADULT FEMALE, VIRILIZATION CAUSED BY A TUMOR most commonly originates in the ovary, such as an arrhenoblastoma or theca cell tumor. Less fre- quently, adrenocortical carcinomas present a mixed clinical and biochemical picture of Cushing’s syndrome plus virilization. Adrenal tumors that exclusively secrete androgens without excess corti- sol are rare. In contrast, small adrenal tumors are common. Reports of autopsy series have demon- strated an incidence of these tumors in approxi- mately 6% of patients which is highly age dependent (range: 1.4% to 15%).1 Incidentally discovered adrenal tumors are discovered in 0.4% to 2.7% of patients undergoing abdominal com- puted tomography (CT).2 Most adrenal tumors are clinically silent, but when functional usually secrete cortisol, aldosterone, or catecholamines. Virilization is frequently encountered in childhood adrenal tumors3 and represents the most common feature of adrenocortical carcinomas in patients of this age group.3 However, there are conflicting data on this

matter in the adult population. Some authors re- port that most of these lesions are benign and others suggest that more than 70% are malignant.4,5

OBJECTIVE

We intended to determine the frequency, symp- tomatology, imaging, surgical and pathology char- acteristics, and prognosis of patients with purely androgen-secreting adrenal tumors.

METHODS

A retrospective chart review was performed of all patients with androgen excess symptomatology associated with an adrenal tumor between January 1946 and November 2002. Patients with docu- mented simultaneous production of cortico- steroids were excluded, as were those with the diagnosis of polycystic ovarian syndrome and adrenal hyperplasia. Follow-up data were obtained through return visits to our institution and through telephone calls to those patients who had not visited our institution within 6 months before the preparation of this study. The Mayo Institutional Review Board approved this study.

RESULTS

Demographics and presentation. Eleven patients with a pure androgen-secreting adrenal tumor were

Presented at the 24th Annual Meeting of the American Association of Endocrine Surgeons, San Diego, California, May 11-14, 2003.

Reprint Requests: Clive, S. Grant, MD, 200 1st St SW, Rochester, MN 55905.

@ 2003, Mosby, Inc. All rights reserved.

0039-6060/2003/$30.00 + 0

doi:10.1016/S0039-6060(03)00410-0

Table I. Summary of 11 patients with pure virilizing adrenal tumors
PtAgeSymptom delay (mos)17-KSPlasma testosteroneImagingOp dateAdrenalPathologyTumor sizeTumor weightOutcomeF/U time (mos)
12784normalelevatedCT true pos10/1/02RightAdenoma682A & W6
2186elevatedelevatedCT true pos8/26/00LeftCancer8149A & W29
347300N/AelevatedCT true pos2/23/00RightAdenoma535A & W36
424elevatedelevatedCT true pos12/21/98LeftCancer680A & W47
516N/AN/ACT true pos8/18/95LeftAdenoma2.520A & W86
65212normalelevatedCT false neg5/26/93LeftAdenoma1.510A & W118
723elevatedN/AUS true pos6/1/82RightAdenomaN/A25A & W245
8218elevatedN/AAbd x-ray10/26/77RightCancer11250A & W245
92912elevatedelevatedEXU12/10/76RightAdenomaN/A93A & W311
10274elevatedN/AAbd x-ray6/19/59RightCancer13530DOD29
1129120elevatedN/ANone9/26/46RightCancer11150A & W387

identified, all of whom were females (Table I). Four were children, three of whom were age 2, and one age 1. The mean age of all patients at the time of the diagnosis was 23.5 years (range 1-53 years). The most common presenting symptoms were hirsutism and acne (90% each) and clitoral enlargement (60%). Four (36%) adult women also had irregular menses or amenorrhea. Hypertension was present in three patients (30%).

Laboratory. Laboratory studies to establish the diagnosis varied. Elevated serum testosterone was identified in six patients (patients #1 and #6 had pure testosterone-secreting adenomas). Dehydro- epiandrosterone sulfate (DHEA-S) was elevated in five and androstenedione was elevated in three patients. All patients had normal serum cortisol levels. Urinary 17-ketosteroids were elevated in seven of the nine patients tested, and urinary ketogenic steroids were normal in all five tested patients.

Imaging. Ultrasonography and CT scan identi- fied an adrenal tumor in six of seven and five of six patients, respectively (Fig 1). Other imaging studies localizing the tumor included plain abdominal X- rays in two, excretory urogram in two, and mag- netic resonance in one. One patient underwent an abdominal/retroperitoneal exploration without preoperative imaging, solely on the basis of clinical and laboratory findings.

Surgical and postoperative management. All 11 patients underwent surgical resection of the tumor by seven surgeons. One adrenalectomy was performed laparoscopically; the remaining 10 operations were performed as open procedures: two by a lateral retroperitoneal approach and eight through an anterior approach. The procedures included seven right total adrenalectomies and four left total adrenalectomies. Two concomitant procedures were performed, including a lap- aroscopic cholecystectomy in the patient who

underwent the laparoscopic adrenalectomy and a uterine myomectomy in another patient. There were no intraoperative or postoperative com- plications.

The mean hospital stay was 8.5 days (range 2-36 days). Dismissal laboratory results confirmed reso- lution of excess production of androgens in all patients. Of the five patients diagnosed with a malignant tumor, two received chemotherapy. The agents used were 5-FU, o,p’ DDD (mitotane) and Chloromycetin in one patient and mitotane in another. In addition the first patient received radiation therapy. She was the only patient with documented metastases, which were present at the time of diagnosis and were limited to the lung. She underwent a wedge resection of the pulmonary metastasis 2 months after the adrenalectomy and did well initially, but died of disease recurrence 17 months after the initial adrenalectomy.

Pathology. Pathology examination was available for all 1 specimens (Fig 2). Six tumors were classified as benign adenomas and five were interpreted as carcinomas. The mean weight and diameter for benign tumors was 44 g (range 10-93 g) and 4.2 cm (range 1.5-6 cm), respectively. The mean weight and diameter for malignant lesions was 232 g (range 80-530 g) and 9.8 cm (range 6-13 cm). Only one malignant tumor showed capsular invasion, the remainder appeared well circum- scribed without capsular invasion. Criteria for mak- ing a diagnosis of malignancy varied but included a weight >100 g, a diameter ≥10 cm, nuclear pleomorphism, and frequent mitoses as well as local invasion and metastases.

Outcome. Follow-up was current on all 10 surviving patients. Overall, the mean follow-up time was 11.7 years (range 6 months-32 years). With the exception of the single patient who had adrenal carcinoma with pulmonary metastases at

Fig 1. CT scan on patient #2 (adrenal carcinoma) with radiology interpretation of an "indeterminate heteroge- neous smoothly marginated noncalcified soft tissue mass involving the left adrenal gland measuring 7.5×5.5×6.5 cm. This mass abuts but does not appear to involve surrounding tissues. No evidence for metastatic spread. The right adrenal gland is normal."

presentation, the remaining four with cancer and all the patients with benign tumors have not shown any evidence of disease recurrence and are alive and well through their last follow-up.

DISCUSSION

The rarity of pure androgen-secreting tumors may be put into perspective when compared with other adrenal tumors. From 1960 to 1996, 120 patients with adrenocortical carcinomas were op- erated at the Mayo Clinic. In just 10 years since 1992, in excess of 350 adrenalectomies have been performed for primary adrenal disease. By contrast, we have operated on only 11 patients with purely androgen-secreting tumors in more than 50 years.

Symptoms of pure androgen-secreting adrenal tumors can be separated into three categories: hirsutism, virilization, and disruption of normal menstrual cycles. Hirsutism refers to excessive hair growth in specific androgen-sensitive areas of the body, including the face (upper lip, chin, side- burns), chest, areola, linea alba, lower back, buttocks, inner thighs, and external genitalia. Virilization includes clitoral enlargement, deepen- ing of the voice, temporal hair loss, and develop- ment of masculine somatic characteristics. Some element of both hirsutism and virilization were found in virtually all 11 of our patients. Menstrual abnormalities also occurred in five of our seven postpubertal women.

Symptomatology of pure virilizing adrenal tumors varies depending on the sex and age of

presentation. In children, these tumors produce a clinical picture similar to that of congenital adrenal hyperplasia with accelerated growth veloc- ity and muscle development, acne, penile/clitoral enlargement, and the precocious development of pubic and axillary hair.

Pure virilizing adrenal tumors in the adult, whether benign or malignant, are rare. These tumors secrete the androgens DHEA and the sulfated form DHEA-S, androstenedione, and testosterone (Fig 3). Normally, the adrenal cortex secretes DHEA, androstenedione, and estrone (formerly measured as 17-ketosteroids, ie, 17-KS), whereas the testis and ovary normally secrete testosterone and estradiol. Masculinization from an adrenal tumor typically results from secretion of large amounts of androstenedione or DHEA (causing elevated 17-KS) that are peripherally metabolized to testosterone. Very rarely, pure testosterone-secreting adrenal tumors have been described (as in our patients #1 and #6). Mattox and Phelan6 collected only 47 such tumors since it was first described in 1925. Usually, a high level of testosterone in a virilized female denotes an ovarian tumor. These tumors usually do not secrete DHEA or androstenedione.

Once elevated levels of serum androgens or urinary 17-KS have been detected, a dexametha- sone suppression test may be performed. Such dynamic hormone testing, involving attempts to either stimulate or suppress the hormone secre- tion, has been suggested to differentiate adrenal from ovarian sources.4 However, adrenal tumors have been stimulated by human chorionic gonad- otropin,7 yet failed to be stimulated by adrenocor- ticotropic hormone (ACTH) or suppressed by dexamethasone.8 Moreover, ovarian tumors have been stimulated by ACTH.9 Therefore, dynamic endocrine testing has been challenged as a means to distinguish between the two sources of hormone hypersecretion.

The best method currently to distinguish be- tween an ovarian and an adrenal tumor as a cause of virilization is by modern radiologic techniques such as computed tomography CT, ultrasonography or MRI, or venous sampling. Only a single 1.5-cm tumor was overlooked in our seven patients imaged with CT or ultrasonography. With current adrenal imaging protocols including 3-mm cuts, even that 1.5-cm tumor would almost certainly be discovered today. In addition, the appearance of an adrenal tumor on imaging studies may suggest whether the tumor is benign or malignant, identify metastases, and may help to plan the therapeutic approach to the patient.

Fig 2. A, Gross 6-cm cross-section of androgen-secreting adenoma (patient #1) that shows a homogeneous adrenocortical adenoma. B, C, D, Gross cross-section, low- and high-power histologic sections of androgen- secreting adrenocortical carcinoma (patient #2). In contrast to Figure 2A of a benign adenoma, note areas of necrosis and hemorrhage in the gross section. The low-power histologic section demonstrates invasion into the tumor capsule. The high-power section shows nuclear pleomorphism and mitoses.

A

B

METRIC

C

Surgical resection is the primary treatment for patients with adrenal virilizing tumors. Tradition- ally, a transperitoneal open anterior adrenalec- tomy for the resection of these tumors has been recommended. This remains the recommended approach for malignant adrenal tumors. However, since the first reports describing laparoscopic adrenalectomy in 1992 by Petelin10 and Gagner,1 this technique has gained wide acceptance for removal of most benign adrenal tumors. Virilizing adrenal tumors have been successfully resected in this manner. Multiple reports have demonstrated that when compared to the open approach, the laparoscopic technique results in decreases in hospital stay, narcotic requirement, blood loss, and recovery time, and overall increased patient satisfaction.12 Currently, laparoscopic adrenalec- tomy is recommended for most adrenal tumors smaller than 6 cm in diameter that are suspected to be benign. Laparoscopic adrenalectomy is curren- tly not favored for adrenal cortical carcinomas,13

although most of them are large and would not even warrant consideration of a laparoscopic approach.

Postoperatively, patients should be closely mon- itored biochemically, with periodic determinations of adrenal androgen levels, urinary steroids, and imaging studies. Recurrent symptoms or biochem- ical abnormalities should prompt a search for locally recurrent or metastatic disease. Metastases from virilizing adrenal cortical cancers primarily involve liver, lung, and regional lymph nodes, the majority of which appear within 1 year of tumor resection.

The advent of CT scans in the 1970s revealed clinically what had been known for years by autopsy studies: small adrenal tumors are relatively com- mon. Their frequency increases with advancing age, they are associated with diabetes and hyper- tension, and a small minority autonomously hyper- secrete clinically important hormones. However, any hope that adrenocortical carcinomas might be

Fig 3. Steroid synthesis pathway illustrating the four major categories of cortical hormones: progestogens, corticoids, androgens, and estrogens.

Steroid Synthesis Pathway

Cholesterol

DHEA-S

Progestogens

Androgens

Pregnenolone

17-hydroxy- pregnenolone

DHEA

Androstenediol

Progesterone

17 hydroxy- progesterone

Androstene- dione

Testosterone

11 deoxy- corticosterone

11-deoxy- cortisol

Estrone

Estradiol

Corticosterone

Cortisol

Estriol

18-hydroxy- corticosterone

Corticoids

Estrogens

Aldosterone

diagnosed earlier by discovering these adrenal in- cidentalomas has been largely disappointing.

Five of our 11 virilized patients had adrenocor- tical carcinoma without either clinical or biochem- ical evidence of cortisol hypersecretion. Only one of our five cancer patients presented with metasta- ses, and she eventually died of disease. The re- maining four were alive and free of disease at last contact (average, 14.75 years). With an estimated incidence of only 0.5 to 2 cases per million per year, large series of patients with adrenocortical carci- noma are uncommon (Table II). Adrenocortical carcinomas that present with mixed virilizing and Cushing’s syndrome are typical, but pure virilizing adrenocortical carcinomas are uncommon, com- prising 5% to 10% of the cases in most large series. Wooten and King14 reviewed 1,891 patients in the English literature through 1993. They found a female-to-male ratio of 3:2, functioning-to-non- functioning of 3:2, and a virtually even split of right and left adrenal cancers. Adrenocortical carcinoma remains a serious and often fatal disease. Un- fortunately, from 30% to 50% of these patients present with advanced stage IV disease with a dismal expected 5-year survival of less than 5%.

In contrast to the often pessimistic picture of adrenocortical carcinoma in adults, reports of pediatric patients with adrenocortical carcinoma are generally more favorable.15,16 All of our patients age 18 and younger (three cancer, two

adenoma) are alive and well 10.9 years post- operatively. Almost all adrenocortical cancers in children are functional, predominantly pre- senting with pure virilization (40%-60%) followed by mixed virilizing and Cushing’s syndrome (Table III). Purely aldosterone secret- ing carcinomas are very rare at any age. Children with localized disease, completely resected, with normal hormonal values postoperatively have had 90% long-term survival.15

Radiotherapy for malignant adrenal tumors has not been shown to improve survival. The principal medical therapy used in patients with adrenocorti- cal carcinoma has been o,p[prime]-DDD (mito- tane), an adrenolytic agent. The response rate has been about 35%,14 but it too has not been shown to prolong survival. Its use in children is not docu- mented. Other agents that interfere with adrenal steroid synthesis, such as ketoconazole, amino- glutethimide, and metyrapone may relieve symp- toms of steroid excess but do not improve survival.

Prognosis for benign adrenal tumors is excellent with resolution of both hormone excess and reversal of the target organ effects. At the opposite end of the spectrum, the usual prognosis for adrenocortical carcinoma is generally poor, with an overall 5-year survival of 20% to 25%. Factors that have been associated with poor prognosis specifically for virilizing tumors include: incom- plete resection, weight >100 grams, size >200 cm3,

Table II. Large series of adrenocortical carcinoma
First authorTotal number patients# Virilized% Virilized5-year survival stage III (%)5-year survival stage IV (%)
Kendrick135800300
Del Gaudio5190105
Imai828452
Henley17623525<5
Crucitti18129108217
Bodie 198291115<5
Soreide 209977<10
Icard2115674300
Table III. Androgen-secreting adrenal tumors: children and adults
First authorTotal # patientsPatientsVirilizationCushing'sMixedNonfunctioningBenignMalignant
Wajchenberg147Children13090047
Adults43153
Sandrini1558Children2322941741
Adults0
Laszlo221Adult11
Forsbach231Child11
Derksen414Children14212
Imai85Child111
Adult444
Del Gaudio38Adults826

age >3.5 years at diagnosis, preoperative symptom duration of more than 6 months, and a marked increase in urinary 17-ketosteroids and 17-hydrox- ysteroids.14-16 However, pure virilizing carcinomas in general appear to have a better prognosis than other adrenal carcinomas. In the present study four of five patients diagnosed with adrenocortical carcinoma showed long-term survival.

Pure androgen-producing adrenal tumors are rare, almost uniquely found in females, and characteristically present with a combination of hirsutism, virilization, and menstrual irregulari- ties. The two most common sources of androgen hypersecretion are the adrenal glands and the ovaries. Identifying which is responsible for the symptoms is best done by obtaining serum and urine tests for androgens and their metabolites, together with modern abdominal imaging tech- niques such as CT and ultrasound scans. Surgical resection is the treatment of choice. For tumors that appear benign, measuring no more than about 5 to 6 cm, a laparoscopic approach seems reasonable. After resection, hormone levels re- turn to normal, the end-organ effects are re- solved, and long-term cure is expected when the tumor is benign. Similarly, in patients with virilizing adrenocortical carcinomas, complete

resection offers an excellent likelihood for long- term disease-free cure.

REFERENCES

1. Dobbie J. Adrenocortical nodular hyperplasia: the ageing adrenal. J Pathol 1969;99:1-18.

2. Grumbach MM, Biller BM, Braunstein GD, Campbell KK, Carney JA, Godley PA, et al. Management of the clinically inapparent adrenal mass (“Incidentaloma”). Ann Intern Med 2003;138:424-9.

3. Latronico A, Chrousos G. Adrenocortical tumors. J Clin Endocrinol Metab 1977;82:1317-24.

4. Derksen J, Nagesser S, Meinders A, et al. Identification of virilizing adrenal tumors in hirsute women. N Engl J Med 1994;331:968-73.

5. Del Gaudio A, Del Gaudio G. Virilizing adrenocortical tumors in adult women: report of 10 patients, 2 of whom each had a tumor secreting only testosterone. Cancer 1993;72:1997-2003.

6. Mattox J, Phelan S. The evaluation of adult females with testosterone producing neoplasms of the adrenal cortex. Surg Gynecol Obstet 1987;164:98-101.

7. Givens J, Andersen R, Wiser W, Coleman SA, Fish SA. A gonadotropin-responsive adrenocortical adenoma. J Clin Endocrinol Metab 1974;38:126-33.

8. Imai T, Tobinaga J, Morita-Matsuyama T, Kikumori T, Sasano H, Seo H, et al. Virilizing adrenocortical adenoma: in vitro steroidogenesis, immunohistochemical studies of steroidogenic enzymes, and gene expression of corticotro- pin receptor. Surgery 1999;125:396-402.

9. Werk E, Sholiton L, Kalejs L. Testosterone-secreting adrenal adenoma under gonadotropin control. N Engl J Med 1973;289:767-70.

10. Petelin J. Laparoscopic adrenalectomy. Proceedings of the International Symposium of the Canadian Society for Endoscopic and Laparoscopic Surgery. Saskatoon, Saskatch- ewan, Canada, 1992.

11. Gagner M, Lacroix A, Bolté E. Laparoscopic adrenalectomy in Cushing’s syndrome and pheochromocytoma. New Engl J Med 1992;327:1033.

12. Thompson G, Grant C, van Heerden J, Schlinkert RT, Young WF Jr, Farley DR, et al. Laparoscopic versus open posterior adrenalectomy: a case-control study of 100 patients. Surgery 1997;122:1132-6.

13. Kendrick M, Lloyd R, Erickson L, Farley DR, Grant CS, Thompson GB, et al. Adrenocortical carcinoma: surgical progress or status quo? Arch Surg 2001;136:543-9.

14. Wooten M, King D. Adrenal cortical carcinoma: epidemi- ology and treatment with mitotane and a review of the literature. Cancer 1993;72:3145-55.

15. Sandrini R, Ribeiro R, De Lacerda L. Childhood adreno- cortical tumors. J Clin Endocrinol Metab 1997;82:2027-31.

16. Wajchenberg B, Pereira M, Medonca B, Latronico AC, Campos Carneiro P, Alves VA, et al. Adrenocortical carcinoma: clinical and laboratory observations. Cancer 2000;88:711-6.

17. Henley D, van Heerden J, Grant C, Carpenter P. Adrenal cortical carcinoma-a continuing challenge. Surgery 1983; 94:926-31.

18. Crucitti F, Bellantone R, Ferrante A, Boscherini M, Crucitti P, et al. The Italian registry for adrenal cortical carcinoma: analysis of a multiinstitutional series of 129 patients. Surgery 1996;119:161-70.

19. Bodie B, Novick A, Pontes J, Straffon RA, Montie JE, Babiak T, et al. The Cleveland Clinic experience with adrenal cortical carcinoma. J Urol 1989;141:257-60.

20. Soreide J, Brabrand K, Thoresen S. Adrenal cortical carcinoma in Norway, 1970-1984. World J Surg 1992;16:663-8.

21. Icard P, Chapuis Y, Andreassian B, Bernard A, Proye C. Adrenocortical carcinoma in surgically treated patients: a retrospective study on 156 cases by the French Association of Endocrine Surgery. Surgery 1992;112:972-80.

22. Loszio FA, Toth S, Kocsis J, Pavo, Szecsi M. Testosterone- secreting gonadotropin-responsive adrenal adenoma and its treatment with the antiandrogen flutamide. J Endocrinol Invest 2001;24:622-7.

23. Forsbach G, Guitron-Cantu A, Vazquez-Lara J, Mota-Morales M, Diaz-Mendoza ML. Virilizing adrenal adenoma and primary amenorrhea in a girl with adrenal hyperplasia. Arch Gynecol Obstet 2000;263:134-6.

DISCUSSION

Dr Eberhard A. Mack (Madison, Wis). I enjoyed your paper very much. Since this is such an unusual entity, I just rise to add one case of ours that was a truly virilizing tumor in a young female. This was before CT scanning was available, and this patient with hirsutism, clitoral

hypertrophy, and amenorrhea was studied extensively over an 11-year period, including two laparotomies and biopsy of the affected adrenal. At that time, angiographic techniques yielded a tumor in the left adrenal that turned out to be a virilizing adrenal ganglioneuroma.

We performed intraoperative samplings, and the testosterone from the adrenal vein was over 17,000 with a peripheral of 450. After resection of the left adrenal, the testosterone value normalized. Two or three years later, the patient sent a birth announcement of a baby girl.

Dr John A. Chabot (New York, NY). Do you use these same criteria to make judgments about treating secret- ing adrenal tumors that aren’t secreting exclusively androgens? As you know, making the decision about which of these are cancers and which aren’t, whether using histology or other clinical criteria, is very, very difficult. Some people say until a metastasis shows up you can’t decide whether one of these is a cancer or an adenoma.

What criteria do you use to make clinical decisions about approaching other adrenal cortical secreting tumors? Is the 6-cm cut-off the same? What criteria do you use to say it is a benign-looking tumor? When you use both criteria, benign versus malignant? What criteria are you using to determine which of these tumors are cancers and which are adenomas?

Dr Cordera. To determine whether these androgen- secreting tumors are benign or malignant may be difficult. Tumor sizes in diameter and weight have previously been major criteria. Weight greater than 100 g was considered suggestive if not diagnostic of carci- noma. Lymph node or distant metastases are definitive. More recently, certain histologic criteria are given more importance, such as frequency of mitoses, nuclear pleomorphism, necrosis, and capsular or vascular in- vasion.

Regarding nonsecreting adenomas identified by ab- dominal imaging studies, we try to make a decision about the likelihood of malignancy based on the combination of tumor size and its phenotype. This includes the tumor shape, any border irregularities or local invasion, tissue inconsistency, vascularity, presence of necrosis or calcifi- cation, and enlargement if prior imaging studies are available. The size threshold to strongly consider adre- nalectomy for nonfunctioning adrenal tumors remains about 4 cm, but many other factors are also considered. In the case of adrenocortical hormone-secreting tumors primarily causing Cushing syndrome, the presence of a tumor greater than 5 to 6 cm, especially if the patient has a mixed clinical picture of Cushing and virilization, would strongly suggest malignancy. Pure aldosterone- secreting cancers are extremely rare.