---

Introduction

Modern imaging techniques like ultrasound, computerized tomography (CT) and magnetic resonance imaging (MRI) enable the detection of adrenal tumors with high sensitivity. Already in 1979, Korobkin et al., using a prototype CT scanner, predicted the “discovery of asymptomatic adrenal disease” (1). Several decades ago, both collection of adrenals for anatomical studies and large autopsy series have demonstrated a high incidence of adrenal nodules in routine autopsies (2-6). Adrenal tumors not suspected prior to radiological investigation are called ‘clinically inapparent adrenal mass’, ‘incidentally detected adrenal masses’ or ‘adrenal incidentalomas’. It is not surprising that due to the widespread use of imaging techniques these so-called “diseases of modern technology” (7) will be diagnosed more frequently, but also might put burden on the patient due to insecurity and will have increasing economic consequences for the health system (8).

Clinical Problems in Patients with Adrenal Incidentalomas

In a patient with a suspected adrenal incidentaloma several questions have to be answered. The most important ones are:

1. Is the adrenal mass incidentally found or does the patient show symptoms or comorbities (e. g. arteriel hypertension) which might be caused by the adrenal mass?

2. Is it an adrenal or extra-adrenal mass?

3. Does the adrenal mass represent a metastasis of an unknown or known primary tumor?

4. Is the adrenal mass hormonally active?

5. Is there evidence of adrenocortical carcinoma?

The following review will focus on these and other aspects of the incidentally detected adrenal mass and will give expert guidelines for diagnostic and therapeutic evaluation of affected patients.

Definition

By common consensus, an incidentally detected adrenal mass is defined as an adrenal tumor not suspected prior to the imaging procedure which led to its discovery (9). Thus, by definition, clinical symptoms and/or signs of an adrenal tumor must not precede the diagnosis of an incidentally detected adrenal mass. For example, patients with incidentally detection of an adrenal tumor in an abdominal CT scan performed for staging of nonadrenal malignant disease cannot be considered to have an incidentaloma, as adrenal metastases are a likely possibility prior to the investigation. In patients who have died from an extra-adrenal primary tumor the prevalence of adrenal metastases at autopsy is in the range of 25-100%, depending on the type of cancer (10). Another example is a patient with newly diagnosed arterial hypertension. They should also undergo a biochemical investigation to exclude an adrenal cause (pheochromocytoma, primary hyperaldosteronism). If this was missed, an adrenal mass detected during subsequently performed ultrasound study should not be called an incidentaloma.

Epidemiology

The precise incidence and prevalence of adrenal nodules is not known, as no population based studies are available. However, in post mortem investigations, the frequency of adrenal tumors is 1.4 - 8.7 % of the cases (2-4,6,11,12). Most of these tumors were small adrenal adenomas. The prevalence of tumors with a diameter > 1.5 cm is 1.8 %, and that of tumors with a diameter > 6 cm is 0.025 % (2). These data are consistent with the results using CT, where a prevalence between 0.6 and 4.4 % has been reported (13-16). Most of these tumors are very small (diameter < 1 cm) and in a study using CT, 79 % of the incidentally detected adrenal masses had a diameter < 2 cm (15). Pooling the data of several studies representing more than 30.000 patients the mean prevalence is around 1 % (Table 1).

Table 1. Prevalence of adrenal tumors in autopsy series and CT studies (reprinted with permission from (9))

First author	year	Design	N	adrenal tumors %
Russi (2)	1944	retrospective autopsy study	9000	1.45
Commons (4)	1948	retrospective autopsy study	7437	2.86
Shamma (3)	1958	retrospective autopsy study	220	1.8
Kokko (6)	1967	retrospective autopsy study	1495	1.41
Hedeland (5)	1968	prospective autopsy study	739	8.7
Reinhard (12)	1994	prospective autopsy study	498	5.0
Total			19389	2.38
Glazer (14)	1982	retrospective CT-study	2200	0.6
Garz (13)	1982	retrospective CT-study	12000	0.5
Kley (15)	1990	prospective CT-study	2568	4.4
Stark (16)	1994	prospective CT-study	13818	0.8
Total			30586	1.0

In some studies, a slight female preponderance has been reported (17,18). This might be explained by a referral bias (e. g. more imaging studies are recommended for women due to a higher prevalence of biliary disease). Probably the prevalence of adrenal masses is equal in both sexes, which is supported by autopsy studies (2,4,5,19).

The frequency of incidentally detected adrenal masses increases with age (2,4,18,20). In individuals > 50 years, the prevalence ranges between 3 – 7 % (21,22).

Some studies have reported a higher prevalence of incidentalomas on the right side (17). However, this observation can be explained by the better ultrasound visualization of the right adrenal gland. Consistent autopsies series and CT investigations indicate that both adrenals are affected equally (18). Bilateral incidentally detected adrenal masses have been reported in 2 - 10 % of the cases (18,23). Table 2 shows the prevalence of adrenal tumors with regard to clinical settings (24).

Table 2. Prevalence of adrenal tumors with regard to age, imaging and malignancy (adrenal metastasis and adrenal cortical carcinoma (ACC))

Autopsy < 30 years > 70 years	2.1 % < 1 % > 7 %
ultrasound for general health screening:	0.1 %
CT and MRI in patients with nonendocrine complaints	0.42 %
Patients with a previous diagnosis of cancer	4.3 %
Likelihood of adrenal metastasis in cancer patients	~ 75 %
Likelihood of ACC in noncancer patients	< 5 %
tumor size < 4 cm	2 %
tumor size < 4 cm	6 %
tumor size 4.1-6 cm	25 %

Clinical Characteristics

Per definition, no clinical symptoms or signs of adrenal disease should be present at the time of diagnosis in patients with incidentally detected adrenal masses. However, a more detailed questioning and a careful second physical examination might reveal evidence of subtle hormone excess (such as recent weight gain, skin atrophy, episodic headaches etc.). Moreover, arterial hypertension and obesity are significantly more prevalent in patients with incidentalomas (20,21). The hypothesis that the higher number of incidentalomas in patients with hypertension is due to the fact that these individuals have imaging procedures more often (e. g. to screen for aortic aneurysm) is not supported by autopsy studies (11,25). In a study, patients with incidentalomas who were suffering from subclinical Cushing´s syndrome (SCCS) were significantly more obese (18). In addition, patients with incidentalomas more frequently suffer from diabetes mellitus type 2 (2,5) and it has been postulated that hyperinsulism in these individuals leads to an increased proliferation of adrenal cells (26). Moreover, some of the few prospective studies showed improvement of metabolic parameters following surgery in patients with SCCS compared to medical treatment (27,28), but another trial failed to prove this in long term follow up (29). Taking these findings together there seems to be an association of incidentalomas with features of the metabolic syndrome (obesity, arterial hypertension, NIDDM, dyslipidemia, dyscoagulation).

Differential diagnosis

The histological characterization of incidentalomas reveals many different pathologic entities as shown in Table 3. However, the exact prevalence of these entities is not known, as only a small proportion of patients require surgical removal of the tumor. By far, most of the adrenal incidentalomas are benign, hormonally inactive adrenal adenomas (> 70%) (30). Nevertheless, surgical removal of incidentalomas always has to be contemplated because adrenal cortical carcinomas are rare but highly malignant tumors, with an estimated annual incidence of 0.5 – 2 cases per million population (21,31). However, data on incidence are mainly based on the National Cancer Institute survey from the early 1970s and probably underestimate the true incidence (32). Surprisingly, in some clinical studies, their frequency was reported to be even 4 - 5 % of all incidentally detected adrenal masses [16, 17, 29]. Although these findings are inconsistent, it seems that only a small percentage of adrenal cortical carcinomas present as incidentalomas. For example, a significant percentage of these patients had complained about abdominal pain and thus these tumors do not meet the above mentioned definition of incidentaloma (18). Moreover, 60 % ACCs show hormonal excess, which often leads to clinical sings (e. g. hirsutism, overt cushing syndrome) (32). In addition, large extra adrenal tumors, mostly primary retroperitoneal neoplasms, are sometimes misclassified as tumors of adrenal origin.

Table 3: Causes of adrenal mass

Hormone excess – 5 – 20 %

Adenoma (aldosterone or cortisol)

ACC (any adrenal hormone)

Pheochromocytoma

Congenital adrenal hyperplasia

Massive macronodular adrenal disease

Nodular variant of adrenal cushing

No hormonal excess

Adenoma

Myolipoma

Neuroblastoma

Ganglioneuroma

ACC

Metastasis

Cyst

Hemorrhage

Granuloma

Amyloidosis

Extraadrenal: infiltrative

Bilateral etiologies

Pathogenesis

Adrenal nodularity increases with age and can affects both glands. This led to the hypothesis that adrenal tumors are a manifestation of the aging adrenal [11]. The additional proof of capsular arteriopathy in these subjects support the concept that the nodules represent focal hyperplasia in response to focal ischemic loss of cortical adrenal tissue (11). However, it has been shown by clonal analysis that the vast majority of adrenal tumors are of monoclonal origin. Therefore, adrenal tumors are the result of clonal expansion following after a somatic oncogenic mutation (33,34). Only some nodules arise from polyclonal focal nodular hyperplasia. These adenomas may have an oligo- or multicellular origin under the putative action of extra adrenal or local growth factors (33). In functional hypersecreting tumors, adrenal cortical carcinomas and familial endocrine syndromes a stepwise tumorigenesis by known affected genes is partially understood (23,35). On the other hand, the genetic background of the “silent” adrenal tumors, which are mostly monoclonal lesions is poorly understood (33,34).

Another hypothesis of the pathogenesis of adrenal adenomas emerged from the frequent finding of macronodular adrenals (80 %) in patients with congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency (Figure 1). If these patients do not receive adequate steroid therapy they are even at an increased risk of developing adrenocortical cancer (36,37). Using abdominal CT scans in families with CAH, Jaresch et al. found adrenal nodules in 45 % of the heterozygous carriers (38). These observations led to the concept that incidentalomas are somehow related to 21-hydroxylase deficiency(39). However, molecular analysis of adrenal tumors revealed that mutations in the 21-hydroxylase gene are rare and are not linked to an increase in ACTH-induced 17α-hydroxyprogesterone (40). Furthermore, after surgical removal of the tumor the elevated ACTH response disappears, which is a proof that the exaggerated 17α-hydroxyprogesterone secretion is restricted to the tumor (41,42). In conclusion, heterozygosity for CAH does not explain the pathogenesis of the majority of adrenal incidentalomas.

Recent investigations of hormon-hypersecreting adrenal tumors showed that aberrant expression and pathologic activation of several G-protein coupled receptors such as those of vasopressin (43), gastric-inhibitory peptide (44), vasoactive intestinal polypeptide (45), luteinizing hormone (46) and catecholamines (47) could trigger or promote some adrenal tumors. Apparently, hypersecretion of neuropeptides, neurotransmitters, growth factors and cytokines that normally participate in the paracrine regulation of the adrenal cortex constitute an unregulated trophic stimulus (31,35).

It was mentioned before that a significant proportion of adrenal nodules may be regarded as a manifestation of the metabolic syndrome (21). In a prospective series of patients with adrenal incidentalomas we found that all 13 consecutive cases exhibited elevated insulin in the fasting state or after a glucose challenge (26). Pathogenetically, this is probably explained by the potent cell growth promoting activity of insulin. Through both insulin and IGF receptors insulin acts specifically on the adrenal cortex and stimulates steroidogenesis and cell proliferation (48-50).

Diagnostic approach

Most lesions of the adrenal gland are benign and hormonally silent (80 – 90 %). Therefore, it is important to limit costs and risks for the patient by avoiding all unnecessary diagnostic interventions. On the other hand, careful diagnostic evaluation holds the potential of early detection of a harmful disease and can lead to a curative therapy e.g. in pheochromocytoma or adrenal cortical carcinoma. Therefore, diagnostic evaluation has to focus on two main questions:

Is their any pathologic hormone secretion, even in the absence of “typical” clinical signs and how high is the risk of malignancy in the individual case?

Endocrinologic investigation

The proportion of incidentalomas with significant endocrine activity increases with tumor size (Figure 2). Thus significant hypersecretion of adrenal masses with a diameter < 1 cm is exceptionally rare. However, aldosterone producing adenomas have to be considered even when the tumor is small, which may cause severe hypertension. At the time of diagnosis up to 20 % of all incidentalomas are significantly hormonal active and little is known about the long term-course of the remainder (Figure 3) (18,21,41,42). In a study by Barzon et al., follow-up of 75 patients with "silent" incidentalomas showed that approximately 10 % of these individuals developed endocrine hypersecretion within 2 years. According to these results, both mass size of 3 cm or more at diagnosis and exclusive radiocholesterol uptake, indicating higher risks to develop hyperfunction, should prompt to plan a more thorough endocrine follow-up (51). Similar conclusions were made from Libe et al. and Vassilatou et al. (52,53). These results emphasize the necessity of regular reevaluation; we would like to suggest a period within one to two years.

Subclinical Cushing’s Syndrome (SCCS)

Autonomous cortisol secretion by adrenal incidentaloma is reported in 5 – 20 % of all patients (17,18,21,26,54-56) (Figure 4). For this entity, the terms subclinical Cushing’s syndrome or subclincal autonomous glucocortioid hypersecretion (SAGH) have been established. Depending on the amount of glucocorticoids secreted by the tumor, the clinical significance ranges from only slight attenuated diurnal cortisol rhythm to complete atrophy of the contralateral gland with long-lasting adrenocortical insufficiency after adrenalectomy (57,58). Thus, subclinical Cushing´s syndrome must be excluded in every patient scheduled for surgery to avoid a postoperative adrenal crisis (57). The best means to uncover autonomous cortisol secretion is the short (1 mg, overnight) dexamethasone suppression test, which rarely fails to detect subclinical Cushing´s syndrome (59-62). A suppressed serum cortisol concentration (< 1.8 µg/dl or 50 nmol/l; apparently this threshold is dependent on the cortisol assay) excludes significant cortisol secretion by the tumor. In a second step, serum cortisol > 1,8 µg/dl requires a confirmatory high-dose dexamethasone suppression test (8 mg) (60). If serum cortisol concentrations are again not suppressible, subclinical Cushing`s syndrome is diagnosed. To facilitate therapeutic decisions, a corticotropin-releasing hormone (CRH) test, the analysis of the diurnal cortisol rhythm (serum or late-night salivary) and the measurement of urinary free cortisol should be performed (60). Patients with low ACTH concentrations not responding to CRH may develop adrenal insufficiency after surgery and require adequate substitution therapy (Figure 5) (56). Increased values of urinary free cortisol are a late finding usually associated with emerging clinical signs of Cushing´s syndrome. Thus, diagnosis of SCCS should not exclusively be established by urine analysis (56).

Pheochromocytoma

The proportion of pheochromocytoma among adrenal incidentalomas has been reported to range from 0 to 11 % (17,18,60). As many of these individuals have no typical clinical symptoms, like sweating, headache, hypertension or tachycardia, a careful biochemical evaluation is necessary (18). Diagnosis should be confirmed or excluded in all incidentally detected adrenal masses, by determination of fractionated metanephrines (metaneprhines and normetanephrines separately) in plasma, urine or both if available. The sensitivity in plasma is slightly higher (95 %) in comparison to urine with better specificity in the latter (63). The accuracy of metanephrines in plasma using EIA (enzyme immunoassay) is nowadays good, but HPLC or GC-MS are still the gold standard. (64). A suppression test (clonidine test) is rarely required. In patients with elevated catecholamine secretion we advocate a I-metaiodobenzylguanidine (I-MIBG) scintigraphy for preoperative detection of possible metastases, which has a sensitivity of 90 % (65). 123I-MIBG seems to have better imaging qualities than 131I-MIBG. However, because the low availability the clinical usefulness is considered debatable, this imaging technique has not been recommended by the NHI-State-of-the-science conference statement (30). 18-Fluro Dopa PET scan shows a higher sensitivity, but is much more expensive. Proper precautionary medication during surgery is necessary because patients with pheochromocytoma may develop life threatening hypertensive crisis (63,65).

Conn’s syndrome

A high prevalence of normokalaemic Conn´s syndrome of 3 – 13 % was reported in series of hypertensive individuals by means of an elevated plasma aldosterone to renin ratio (66-69). Probably due to the exclusion of individuals with hypertension per definition this entity seems to be rare among individuals with incidentally detected adrenal masses (18,21). Hypertension and only slightly decreased potassium (< 3.9 mmol/l) levels already should raise suspicion of primary hyperaldosteronism. In a recently published multicenter Italian study, 60 % of the individuals with aldosterone producing-adenoma had potassium levels in this range, the remainder were normokalaemic (18). In patients with adrenal incidentaloma, who are normotensive and have a serum potassium above 3.9 mmol/l, no further evaluation is necessary (67). Otherwise, determination of plasma aldosterone concentration (PAC) together with plasma renin concentration (PRC) is required. As PRC is easier in preanalytics handling (no cooling of samples on ice), this parameter is preferred nowadays over plasma renin activity (PRA: ng/ml/h).

As several antihypertensive drugs interfere with mineralocorticoid regulation the discontinuation of these substances is recommended, to allow interpretation of the measurement (67,70). Beta-blockers and clonidin can result in false positive results and should be paused for at least one week. Loop diuretics, AT II blockers and the renin inhibitor aliskiren should be stopped one week, spironolactone, eplerenone and drospirenone at least four weeks before a screening test; these substances can lead to false negative results. Alpha Blockers (e. g. urapidil) and calcium antagonists (e. g. amlodipin) might be used alternatively in the time period before testing, although these substances also increase the risk of false negative results (67,71). As hypokalemia itself can reduce aldosterone secretion, normalization of potassium levels before testing is mandatory. If the PAC is elevated and the PAC (ng/l) / PRC (ng/l) ratio is higher than 20 in two independent samples, hyperaldosteronism should be suspected (67,72). As for PAC and PRC several assays exist, adherence to specific cut-off values and ratios is crucial. The diagnosis of primary hyperaldosteronism has to be confirmed by either the failure of fludrocortisone to suppress aldosterone (0.1 mg fludrocortison every 6 hrs. for 4 days, PAC on day 5 greater than 50 ng/ml, sampling mid-morning after 2 hours upright position) or after acute saline suppression (2 liters of 0.9 % NaCl solution infused i. v. over a period of 4 h; failure to suppress PAC < 50 ng/l) (67,70). As these tests are contraindicated in patients with congestive heart failure, the captopril test might be a good alternative in this group: determination of PAC in the morning, followed by intake of 25 mg of Captopril. Two hours later measurement of PAC; indicative of PAH is an unchanged PAC, whereas in essential hypertension PAC is suppressed (73,74).

Patients with an elevated PAC and PAC/PRC ratio require further evaluation in a specialized center. Investigations may even include bilateral adrenal vein catheterization to prove that the incidentaloma is indeed the source of the mineralocorticoid excess and to determine whether aldosterone is produced by one or both adrenals (67,70,75).

DHEA-S and other sexual steroids

Measurement of DHEA-S has been recommended in the work-up of incidentalomas. In a high percentage of patients with adrenal tumors low serum dehydroepiandrosterone-sulfate (DHEA-S) has been reported. This may be used as evidence of an adrenocortical origin of the tumor (76,77). Otherwise, high DHEA-S concentrations are found in a substantial proportion of adrenocortical carcinomas (32). In a recently published retrospective study DHEA-S values were elevated in 17 % of all patients with cortical carcinoma (28 % in younger individuals, age < 50 years). The sensitivity and specificity were 17 % and 93 %, respectively; negative and positive predictive values were 95 % and 10 %, respectively (18). The pathogenesis of low DHEA-S levels in adrenal adenomas has not been elucidated yet. Suppression of ACTH by silent hypercortisolism has been suggested but remains to be proven (54). Possibly low DHEA-S merely indicates premature adrenal aging which favors the development of adrenal nodules (11). The value of DHEA-S in the assessment of adrenal tumors is low, because DHEA-S shows a steep age-related decrease in normal subjects with wide intersubject variation (78). Moreover, some adrenocortical cancers may also exhibit decreased serum DHEAS, which is suggestive of a benign adenoma (17).

However, in suspicion of an adrenocortical carcinoma (ACC), DHEA-S, 17-hydroxyprogesteron, testosterone (in female), androstendion and ostradiol (in male) is recommended, because of the high percentage of hormonal hypersecretion of this tumor.

Adrenal imaging

At this time, most of the adrenal incidentalomas are detected by ultrasound, which reflects the widespread use and technical improvement of this technique (18). Additional imaging by CT is usually performed, as detailed characterization of an adrenal mass by ultrasound alone cannot be obtained. An exception is the highly echogenic adrenal myelolipoma, which does not need further imaging investigation. On CT, adrenal adenomas typically appear homogeneous, exhibit a density lower than 20 Houndsfield units (HU) and well defined margins (21,79,80). However, direct comparison of CT results and histological confirmed diagnosis has shown an overlap in density of scans, with a range of -19 – 43 (mean 19 HU) in adenomas and 31 – 43 (mean 31 HU) in ACCs (81). Therefore, a cut-off of 10 HU (24) in unenhanced scans and 35 HU with a 10 – 15 min delay on contrast enhanced CT (“washout” > 50%) seems to be practicable to distinguish between adenomas and metastases or ACCs (30,82). Adrenal cortical carcinomas are generally larger, inhomogeneous, show soft tissue density and calcifications. Irregular margins and central necrosis or hemorrhage increase the probability of malignancy. However, benign pheochromocytoma may also present as a large inhomogeneous tumor with hemorrhage (83).

Although chemical shift MRI is commonly performed, it probably does not provide additional information beyond that which is already available on unenhanced CT. In difficult cases, MRI may be helpful to further differentiate benign from malignant adrenal tumors, although again a clear discrimination cannot be achieved (82,84). Adenomas exhibit a decreased intensity on T2-weighted MRI compared to non-adenoma. Gadolinium-enhanced dynamic studies have found strong enhancement and slow wash-out in malignant neoplasms, while adenomas showed rapid wash-out (82,85). However, there are contrary results in the differentiation of benign and malignant lesions (86). The recently developed chemical shift MRI (CSI) is the recommended method of choice today. Benign tumors like adrenal adenomas with high lipid content demonstrate a typical signal intensity loss on chemical shift imaging relative to the liver (82,86) . Recently a sensitivity of 91 % and specificity of 94 % was reported using CSI in comparison to histopathology (84).

Additionally, 131I-methylnorcholesterol scan was recommended as a valuable screening tool for subclinical Cushing´s syndrome, as evidence was provided that unilateral uptake is related to functioning adenomas (87). Adrenocortical scintigraphy with NP-59 and 75Se-methylnorcholesterol has been advocated for analysis of adrenal incidentalomas (21). Because of low sensitivity, cost and limited availability these techniques are used infrequently and can’t be recommended as standard diagnostic tools (51,88).

18-Fluoro-2-deoxy-D-glucose positron emission tomography ( FDG-PET) has a high sensitivity in malignancy, but is not specific. It should be used only in high suspicion of ACC or metastasis in order to plan further treatment. Moreover, FDG-PET might be useful in adrenocortical tumors to clarify whether the mass is malignant or benign (CT scan: high unenhanced density (> 10 HU) and inappropriate washout (< 50%) (89). An adrenal to liver maxSUV ratio of less than 1.45 is predictive of a benign lesion. In addition, the recently developed ligand 11C-metomidate, detects nonnecrotic adrenocortical tumors (benign and malignant) and is negative in other masses (90). This modern and expensive technique might be indicated rarely to distinguish adrenal from nonadreanal tumors.

In summary, conclusive adrenal imaging may be established by CT in most patients and by MRI in an additional number. For the few remaining patients needing further characterization, nuclear medicine techniques are advantageous as additional imaging modalities (89,90).

Tumor size

The size of incidentalomas has most often been used as an important tool to discriminate benign from malignant lesions (18,91). It was shown in several studies that the probability of malignancy increases with the tumor size (17,21). The rationale behind this approach is the observation that most ACCs are very large or significant larger than adenomas at the time of diagnosis (18,32,92). Adrenal cortical carcinoma accounts for 2 percent of tumors less than or equal to 4 cm, 6 percent of tumors 4.1–6 cm, and 25 percent of tumors greater than 6 cm (24). However, it is obvious that using only tumor size for discrimination of incidentalomas, some smaller carcinomas will be misclassified, because carcinomas also start small. There is no evidence that early detection of ACC improves overall survival, which is less than 50 percent in five years (24,32).

Follow-up of patients with nonfunctional adrenal mass suggests that 5-30 percent of masses increase in size (24,93). It is recommended to perform a follow up imaging study 6 – 12 month after initial evaluation (24,51,93) . Additional follow up studies are not necessary, since tumor growth thereafter is extremely unlikely in a “stable” adrenal mass, which was concluded from the NHI consensus conference in 2002 (24). However, general benefit of follow up has been questioned recently (94), whereas other groups suggest longer follow up (53,93).

If clear tumor growth is demonstrated, surgical removal of the tumor is warranted. Surgical removal of suspicious tumors larger than 3 cm or even 2 cm has been recommended in some cases (51,88).

Fine needle aspiration (FNA)

In patients with adrenal incidentaloma and no history of malignancy, FNA has no proven efficacy, as histologic differentiation between benign and malignant adrenal tumors is difficult [(24,95). However, recently it was shown in an ex vivo puncture approach, that with appropriate immunohistochemistry, a differentiation between malignant and benign adrenal masses may be possible (96,97). FNA is not free of side effects and may lead to pneumothorax, frank retroperitoneal bleeding or needle track metastasis in the case of adrenocortical carcinoma (98,99). The additional improvement of the new imaging techniques (CT and MRI) in characterization of adrenal masses should restrict FNA to few indications (84). In autopsy series adrenal metastases have been found in 25 – 72 % in the case of an extraadrenal malignoma (10,84). In this group of patients FNA is a good tool (sensitivity and specificity around 90 %) and should be performed if the presence of adrenal metastasis may alter the therapy or prognosis. Thus, we restrict FNA to patients with known malignancy, in which the adrenal mass is the only evidence of possible metastasis [(100,101). In addition, pheochromocytoma always has to be excluded prior FNA as it may cause hypertensive crisis and even death (102,103).

Stepwise diagnostic approach

The authors perform a stepwise endocrinologic investigation in patients with incidentalomas as follows:

Table 4. Stepwise endocrinological investigation in adrenal incidentaloma (modified after Allolio (9), with permission)

Step I: screening test fractionated, plasma free metanephrines or 24 h urinary fractionated metanephrines (both, if available) Serum cortisol after dexamethasone suppression (1 mg at 11 PM orally) Serum potassium and repeated blood pressure measurements. In case of spontaneous hypokalemia or arterial hypertension measurement of serum aldosterone and plasma renin concentration

Step II: confirmatory test (only if the corresponding test results are abnormal) 123-I-MIBG scintigraphy (or 131 I-MIBG) high-dose dexamethasone (8 mg) suppression test. If pathologic: CRH test, analysis of diurnal cortisol secretion, 24 h cortisol excretion (see Tab. 5) Fludrocortisone suppression test or saline infusion test, 24-h excretion of aldosterone and metabolites (i.e. 18-glucuronide aldosterone). In selected cases bilateral adrenal vein catheterization with determination of aldosterone and cortisol

Step III: reevaluation repeat screening tests after 2 years in patients with a tumor size of > 3 cm (cortisol, metanephrines)

In summary, the most important questions in a given patient with an adrenal incidentaloma (section 2) may be answered as follows (Figure 6).

Therapeutic considerations

To prevent serious morbidity, all hormonally active incidentalomas must be surgically removed (18,104). This strategy is undisputed for Conn´s adenomas and pheochromocytomas. However, it remains controversial whether all patients with subclinical Cushing´s syndrome benefit from adrenal surgery (24,60), as progress from subclinical disease to overt Cushing´s syndrome may occur only in a minority of cases (51,58,93). As autonomous cortisol secretion by the tumor may range from a small percentage of the daily requirements to borderline hypersecretion with suppression of the contralateral adrenal, metabolic benefits of surgery will vary accordingly. In patients with subclinical Cushing´s syndrome undergoing unilateral adrenalectomy, a permanent weight loss in obese individuals, a reduction of hypertension and an improvement of glycemic control in diabetics are frequent findings (41,42,51). Moreover, recently a decrease in bone mineral density and altered bone metabolism in patients with SCCS (diagnosed by elevated urinary free cortisol) was shown, which indicates that these individuals are at higher risk of osteoporosis (105-108). Although these results were not obtained in large prospective studies, they indicate that surgical therapy should be advocated in more patients.

Adrenal surgery has a low, significant morbidity and mortality (109) which has to be taken into account if an operation is considered. As there are no prospective, randomized trials comparing open with laparoscopic adrenalectomy both approaches are acceptable (24). The newly developed endoscopic adrenalectomy (109) with its low morbidity (5 – 10 %) and mortality (< 1) may justify earlier intervention. However, it has to be emphasized that these operations are mostly performed in specialized centers by surgeons experienced in laparoscopic techniques (110). There are no studies that demonstrate a consistent benefit of one laparoscopic approach (transabdominal or retroperitoneal) over another (24).

Generally, surgery should be considered in patients with SCCS who have elevated late-nigth salivary cortisol and elevated urinary free cortisol because progression to overt Cushing`s syndrome is imminent. When the diagnosis of subclinical Cushing`s syndrome is established by high-dose dexamethasone suppression, adrenalectomy is recommended in younger individuals (age < 50 years), in patients who have metabolic disease of recent onset possibly related to the incidentaloma (e. g., obesity, hypertension, diabetes), or if they have osteopenia or osteoporosis (Table 5) (56).

Table 5. Relative and Absolute Indications for Adrenalectomy in Patients with Suspected Subclinical Cushing´s Syndrome (SCCS) based on the Clinical Presentation and Biochemical Findings

Biochemical Classification	Clinical Presentation*	Therapy
A: Decompensating SCCS Abnormal dexamethasone suppression test (8 mg), elevated urinary free cortisol, and/or late-night salivary cortisol	Asymptomatic	Consider surgery
	Possibly symptomatic	Consider surgery
B: Compensated SCCS Abnormal dexamethasone suppression test (8 mg), normal urinary free cortisol, and late-night salivary cortisol	Asymptomatic	Conservative
	Possibly symptomatic	Consider surgery
C: SCCS unlikely Normal dexamethasone suppression test (8 mg), but elevated urinary free cortisol or low morning ACTH or an abnormal late-night salivary cortisol	Asymptomatic	Conservative
	Possibly symptomatic	Conservative
*Asymptomatic patients do not have symptoms and signs of Cushing´s syndrome and have normal blood pressure, glucose metabolism, and body weight. Possibly symptomatic patients do not have symptoms and signs of Cushing´s syndrome but have hypertension, diabetes mellitus, diffuse obesity, or osteopenia. (modified from: Reincke, M. (2000). “Subclinical Cushing's syndrome.” Endocrinol Metab Clin North Am 29(1): 43-56)

Because of the increasing risk of malignancy, all tumors with a diameter 6 > cm are surgically removed after adequate endocrinologic investigation. Smaller tumors (4 – 6 cm) may also be removed, if there is strong suspicion of malignancy after adrenal imaging and scintigraphy (21,88). Additionally, in the case of clearly elevated serum DHEAS or other sexual steroids, surgical therapy should be considered, especially in patients younger than 50 years (18). All tumors demonstrating significant growth during follow-up should also be removed.

NIH Consensus development program: "incidentaloma"

In February 2002 a NIH state-of-the-science conference on “Management of Clinically Inapparent Adrenal Mass” (“Incidentaloma”) developed recommendations for the work-up and treatment of patients with adrenal incidentalomas (http://consensus.nih.gov/ta/021/021_intro.htm) (24,30). A summary of these recommendations is shown in Figure 7.

Summary

Due to the advances and broader use of modern imaging, adrenal incidentalomas are detected with increasing frequency. Their prevalence peaks between the 50th and 60th year (3 – 7 %) and is particularly high in patients with features of the metabolic syndrome (arterial hypertension, obesity, insulin resistance). A biochemical investigation (serum and/or urinary metanephrines, 1 mg dexamethasone suppression test, plasma aldosterone / renin concentration) reveals a subtle, but significant endocrine activity of these tumors in up to 20 % (e. g. pheochromocytoma and subclinical Cushing´s syndrome). The risk of malignancy is low, but it increases with the size of the mass. Therefore, the diameter of the tumor remains the most useful parameter for clinical decision making. Hormonal significantly active incidentalomas and all tumors > 6 cm should be surgically removed. The newly developed endoscopic adrenalectomy with its low morbidity and mortality may justify earlier intervention. Smaller tumors require both imaging and endocrine follow-up. If tumor growth is detected, the adrenal mass must be surgically removed. Imaging studies beyond 12 month in a stable adrenal mass are not recommended. Fine-needle aspiration should be restricted to oncologic patients, in whom the adrenal mass is the only evidence of possible metastasis. The utility of adrenal scintigraphy remains controversial.