Ewa Cyrańska-Chyrek M.D., Department of Endocrinology, Metabolism, and Internal Medicine, University of Medical Sciences, Przybyszewskiego Str. 49, 60–355 Poznań, phone: 609 725 558, 61 869 13 30, e-mail: ewacyranska@gmail.com

Introduction

Adrenal incidentaloma (AI) includes all lesions with diameter ≥ 1 cm found incidentally during imaging examinations not associated with suspected adrenal pathology. Due to greater availability of imaging techniques (USG, CT, MRI) and various diagnostic recommendations, a systematic increase in AI detection has been observed worldwide. Adrenal lesion detection during CT examination in oncological patients is estimated at 1.3% [1, 2] and 4% [3]. Furthermore, aging and screening regarding early detection of lung cancer, including chest CT, conducted in Poland, constitute additional factors enhancing this trend.

Every adrenal incidentaloma requires a detailed hormonal and imaging assessment. This results in high hospitalisation and additional examinations fees, as well as patient anxiety. AI diagnosis includes radiological phenotypical evaluation (particularly in CT, or in case of contraindications – in MRI) and biochemical assessment of tumour hormonal activity, which includes cortisol circadian rhythm, salivary cortisol concentration, 24-hour urinary free-cortisol and metanephrines test, short cortisol dexamethasone suppression test, ACTH level assessment, aldosterone and ARO concentration, and adrenal androgen level.

Each diagnostic step is associated with limitations and method imperfections, which may result in false positive and false negative results. The following paper aims to summarise the aetiology of the most common diagnostic mishaps, which frequently lead to misdiagnoses, an increase patient’s anxiety and, as a consequence, the introduction of improper therapy or its discontinuation.

Adrenal incidentaloma radiological assessment

After CT, abdominal ultrasonography (USG) is the most effective method in finding asymptomatic adrenal adenomas. Healthy adrenal glands are not visible in the examination due to their small size. In fact, in adults they do not exceed 2–5 cm, with branches reaching 6–10 mm [4]. It is estimated that ultrasound assessment is effective for tumours of about 3 cm. The sensitivity of the abdominal ultrasound is 76% (lower than CT and MRI), although it has high specificity of 92% [5]. Additionally, sensitivity decreases in the left adrenal gland, in obese patients and in lesions < 2 cm.

What is more, structures imitating adrenal adenomas may appear, also known as phaeochromocytomas. Therefore, only CT can verify the USG images with erroneously suspected adrenal lesions. The following disorders may turn out to be the imitating structures: superior renal pole tumour, renal cysts, lymph nodes (bilaterally), on the right-focal liver lesions, hepatic flexure, inferior vena cava dilatation, whereas on the left – accessory spleen, lobular spleen, tail of pancreas, splenic artery aneurysm, and stomach diverticulum.

Computed tomography (CT) is the first-line examination in AI diagnosis. The following morphological criteria are analysed: the size, shape, radiation attenuation coefficient (in Hounsfield units [HU]), tumour homogeneity or heterogeneity, and the presence of regressive changes and calcifications. Additionally, tumours smaller than 3 cm are usually benign, whereas those larger than 6 cm are more frequently malignant [6].

The majority of adrenal incidentalomas have a benign imaging phenotype, i.e. their diameter is not larger than 3–4 cm, they have smooth margins, and are characterised by low density, where density ≤ 10 HU indicates adenoma or a cyst, while density ≤ (–) 50 HU points to myelolipoma (sensitivity 71%, specificity 98% [7]). The malignancy features are polycyclic, blurred margins, non-homogenous structure (including cavitation and calcifications), as well as rapid progression of size (> 6–10 mm per year).

However, lesions with smooth margins (round and oval) may also be malignant, e.g. malignant phaeochromocytoma, or small metastases from cancers outside the adrenals. Although calcifications are most common in adrenal cortex cancers, they may also appear in the tubercular lesions and in neuroblastomas [8]. On the other hand, benign phaeochromocytomas may constitute a non-homogenous focus, comprising cavitation and calcifications. Hence, phaeochromocytoma phenotypic diversity is reflected in the name “imaging chameleon” [9]. Moreover, non-homogenous structure and large size may also constitute features of adrenal adenoma following haemorrhage or a very large myelolipoma lesion.

About 30% of all adrenal adenomas are lipid-poor (with density between 10–30 HU). Frequently, in these cases, single-phase adrenal CT is insufficient for a definitive diagnosis concerning tumour character (adenoma or non-adenoma) and requires two-phase examination with lesion density assessment before (0’) and after administration of iodine contrast medium (in the first minute [1’] – enhancement assessment and 10’ or 15’ washout assessment). In fact, owing to this method it is possible to characterise the tumour on the basis of an additional parameter, i.e. absolute and relative of water soluble contrast washout (WSC). Furthermore, visualising the potential infiltration of the nearby structures, as well as the presence of metastatic lesions is also possible.

Sangwaiya MJ et al. [10] analysed the imaging results of 323 adrenal adenomas. They concluded that the water soluble contrast washout analysis (with the cut-off point at 40%) reached a sensitivity of 76.9%, specificity of 93.75%, and accuracy of 77.7% in the differentiation between adenomas and other adrenal tumours. Apart from specificity, these values are significantly lower than those presented in the previously published papers (98%, 92%, 98%, respectively) [11, 12].

Magnetic resonance imaging (MRI) is the secondline examination, performed when contrast CT is inconclusive in the diagnosis of the adenoma type or in cases of suspected phaeochromocytoma. A large amount of lipids suggests a benign lesion. The advantage of this examination is the possibility of detecting even the smallest lipid amounts, consequently allowing differentiation between adenomas and other lesions in 90% of cases.

A retrospect study by Haider MA et al. showed that adrenal adenomas of density > 10 HU (thus, lipidpoor), which had not been classified in the single-phase CT, were correctly diagnosed in the chemical shift MRI in 67% of cases [13].

Another advantage of this method is identifying phaeochromocytomas, as they show increased signal intensity in relation to the liver and the spleen (as they are hyperintense) in T2-dependent images. Due to this feature, MRI is crucial in the differentiation between phaeochromocytomas and carcinomas. In the literature descriptions of pleomorphic adenomas are available with cortex and adrenal medulla tissue, therefore comprising lipids and simultaneously overproducing catecholamines [14].

Adrenal collision tumours constitute a separate issue because they represent the coexistence of foci of different origin in the adrenal adenoma area [15]. Most commonly, two benign lesions coexist, i.e. adenoma and myelolipoma [16], where metastatic tumours in the adenoma (mainly from lung and breast cancer, or melanoma) are less frequent [17, 18]. However, phaeochromocytoma foci have also been found in lipid-rich adenomas.

Biochemical diagnosis of adrenal incidentalomas

According to the literature, hormonally inactive adenomas constitute 75–80% of AI cases. Among the hormonally active lesions, the most frequently diagnosed tumours are adrenal cortex carcinomas. They secrete glucocorticoids autonomously and present ACTH-independent or subclinical ACTH-independent Cushing’s syndrome (where incidentalomas constitute about 5%). Another 5% of AI cases are phaeochromocytomas that overproduce catecholamines. About 1% of adrenal incidentalomas autonomously produces aldosterone (aldosteronoma), which causes hyperaldosteronism, a.k.a. Conn’s syndrome. Adrenal cortex carcinomas make up less than 5% of incidentaloma cases, and metastases to the adrenals constitute about 2.5% of AI cases.

According to the indications, in all AI patients, regardless the clinical picture and concomitant diseases, hypercortisolaemia and phaeochromocytoma tests are necessary. Diagnosis should also include hyperaldosteronism testing in patients with diagnosed hypertension, which should be obligatory in resistant hypertension cases as well as in patients with history of hypokalaemia episodes [21, 22]. Hormonal diagnosis is not required in lesions with imaging phenotype of cysts or myelolipomas [23].

ACTH-independent Cushing’s syndrome testing

Cortisol circadian rhythm assessment

Cortisol secretion is pulsatile, in a circadian rhythm, as a response to ACTH stimulation. The highest level is reached in the morning, there is a 50% decrease at noon, and the lowest level is estimated late in the evening. Increased cortisol late-night concentration is assumed to be an early symptom of hypercortisolaemia.

Total cortisol serum concentration level may be falsely increased or falsely decreased in certain clinical states, concomitant diseases, and administration of some drugs (Table I).

Due to the possible non-specific late-night cortisol increase, which is stress-induced, cortisol level testing is recommended during or after the second day of hospitalisation. In order to reduce patients’ anxiety exacerbating hypercortisolaemia, optimisation of phlebotomy conditions has become more important; thus, the blood drawing time should be shorter and the patient should not be abruptly awaken.

What is more, in cases of concomitant diseases, where catabolic processes and loss of proteins are predominant, falsely decreased results of total cortisol concentration may be obtained.

Urinary free-cortisol concentration

Daily urinary cortisol output reflects plasma free-cortisol concentration. Lower than threefold upper limit of normal may occur in pseudo-Cushing syndrome cases, which may be induced by prolonged anxiety, depression, alcoholism, or pregnancy. Vitally, only a positive urinary free-cortisol output result obtained three times is evidence against the diagnosis of endogenous hypercortisolaemia. In fact, in other cases the possibility of subclinical Cushing’s syndrome with periodic increased cortisol output (cyclic Cushing’s syndrome) [24].

Daily cortisol urinary output assessment is characterised by high specificity (96%) but low sensitivity (40–50%) [25]. Cortisol concentration in urine may be falsely increased or falsely decreased in the cases presented in Table II.

Late-night saliva cortisol concentration

Currently, measuring late-night cortisol concentration in saliva constitutes a widely accepted screening method in the diagnosis of hypercortisolaemia where its increased levels (> 4.0 nmol/L [145 ng/dL] at 11 p.m.–12 p.m.) are considered to be an early symptom. The obtained result correlates with free cortisol concentrations, regardless of binding protein concentration changes [27]. This is the most vital element in the differentiation between true endogenous hypercortisolaemia and elevated cortisol level as a result of the increase in CBG concentration, e.g. in the course of oestrogen therapy or during pregnancy. The examination consists of chewing a saliva-absorbing tampon for 1–2 minutes. It is recommended to perform the examination on two consecutive days (at about 11 p.m. – midnight). The samples are placed in special Salivettes, where cortisol concentrations remain constant in room temperature for as much as a week [28, 29]. A tampon placed in a refrigerator may be stored for several weeks.

A meta-analysis conducted by Carrol T et al. showed outstanding sensitivity (92–100%) and specificity (93–100%) of the examination. The diagnostic process may take place at home, in an environment devoid of additional stress, without generating additional hospitalisation costs, and with repeatable tests results [31]. It is recommended in children, in suspected cyclic hypercortisolaemia, and in cases of falsely increased results due to CBG elevation (at the beginning of pregnancy, and oestrogen therapy). However, there are cases where falsely increased cortisol levels in saliva may still appear (Table III).

Measuring salivary cortisol constitutes also a highly useful test in patients following neurosurgery in the course of ACTH-dependent Cushing’s syndrome because it allows for an effective assessment of surgery, as well as for early detection of the relapse.

ACTH concentration

The secretion of adrenocorticotropic hormone (ACTH) is pulsatile, with short half-life (t½ is 3–8 minutes). Due to large circadian fluctuation it is not employed as a screening test in the diagnosis of adrenal diseases. In fact, it is fundamental to obtain a proper blood sample for analysis, i.e. venous blood should be drawn into an EDTA tube, cooled, and delivered to the laboratory. Another possibility is to separate plasma and deliver frozen plasma to the laboratory. Glass test tubes should not be used because ACTH is adsorbed in the glassware. During phlebotomy, the patient should not be stressed, and the blood should be preferably drawn using a previously inserted cannula. Improper sample storage leads to fast ACTH breakdown and to falsely decreased results. Some causes of falsely increased and decreased ACTH results are presented in Table IV.

ACTH secretion is not always entirely blocked despite the autonomous cortisol secretion by the hormonally active adenoma. Therefore, on the sole basis of ACTH concentration it is impossible to differentiate between ACTH-independent and ACTH-dependent Cushing’s syndrome (particularly in the 10–40 pg/mL range).

Low-Dose Dexamethasone Suppression Test

The test consists of the observation of cortisol secretion suppression after oral administration of 1 mg dexamethasone at 11–12 p.m. The cortisol serum concentration is measured on the following day in the morning after overnight fasting (at 8–9 a.m.). According to the AACE and AAES 2009 recommendations, cortisol concentration > 138 nmol/L (5 ug/dL) in asymptomatic adrenal incidentaloma patients is an indication for further diagnosis. Cortisol concentrations < 50 nmol/L (< 1.8 ug/dL) exclude hypercortisolaemia. Until recently, lowering cortisolaemia to 138 nmol/L was thought to be within the norm. However, lowering the cut-off point to 50 nmol/L significantly increased the test sensitivity: 58% vs. 75–100% respectively [32]. It is estimated that with the current criteria, the test has high sensitivity but relatively low specificity (88%) [33]. Holleman F et al. suggested 95 nmol/L [3.4 ug/dL] as a threshold value in cortisolaemia in 1 mg DXM test, and with this value they obtained a sensitivity of 100% and higher specificity 994%0 [34]. Nevertheless, Morelli V et al. suggested 83 nmol/L (3.0 μg/dL) as a criterion level [35].

It is crucial to notice that a single abnormal result with DXM in adrenal adenoma patients should not constitute the basis for the diagnosis of hypercortisolaemia, or result in surgical intervention.

False positive results may be the consequence of increased cortisol binding protein (CBG) concentrations, e.g. in hyperestrogenism or in pseudo-Cushing’s syndrome (as characterised above). Analogically, in patients with lower CBG concentration (as above) and/or suffering from hypoalbuminaemia, there is a risk of obtaining false negative test results.

In patients treated with hepatic microsomal enzymes induction drugs (mainly CYP3A4), accelerated dexamethasone metabolism occurs, which may result in a false positive result [36] (Table V).

In the administration of medications suppressing CYP3A4, false negative results may be obtained. The most commonly used CYP3A4 inhibitors are listed in Table VI.

Another factor contributing to false positive results may be impaired dexamethasone absorption, e.g. during treatment with antacids, which in turn leads to glucocorticoid absorption impairment from the digestive tract. Meikle AW et al. suggested simultaneous measurement of cortisol and DXM serum concentrations, which would assist in assessing the DXM absorption and thus optimise the test [37].

Proper phaeochromocytoma diagnosis

Phaeochromocytoma cells comprise Catechol-O-methyltransferase (COMT), which is a catecholamine to methoxy catecholamine transformation catalyst. Moreover, methoxy catecholamine synthesis in tumorous cells is constant and therefore their concentration test constitutes a better tumour presence marker than catecholamine concentration tests, which are secreted episodically.

The awareness of outside factors influencing catecholamine and its metabolite concentration levels both in blood and in urine is crucial in the proper interpretation of biochemical trials. Stressful situations and some diseases, such as myocardial infarction, heart failure, hypoxia, or stroke, may enhance sympathetic system activity and hence result in false positive results. Overproduction of methoxy catecholamines takes place also in clinical states, e.g. increased exertion, pain, shock or alcohol abuse. Additionally, a number of drugs increase methoxy catecholamine urinary concentration (Table VII).

False negative test results may also be caused by reserpine, ingesting tyrosine hydroxylase inhibitors prior to testing which is L-tyrosine to dihydroxyphenylalanine catalyst, DOPA (bananas, nuts, citrus fruit, tea, coffee, other caffeine-containing products, and vanilla), as well as iodine contrast mediums. What is more, biochemical tests may be performed only on the fifth day after administration of iodine contrast medium. In fact, not adhering to this rule may lead to false negative results.

Despite the fact that catecholamine plasma concentration has high sensitivity (96-100%) and specificity (89%), it is not easily accessible in Poland. Free methoxy catecholamine concentration within normal levels indicates a diagnosis outside phaeochromocytoma, although tumours secreting only dopamine should also be borne in mind. According to the guidelines of the Polish Society of Hypertension, daily methoxy catecholamine urinary output should be repeated three times, because in some cases a single test is not diagnostically conclusive due to episodic catecholamine release [38].

Adrenaline and noradrenaline plasma concentration has a considerably lower sensitivity (69%) and specificity (86%). Moreover, four hours before phlebotomy, the patient must not smoke because it may result in a false positive result.

Daily vanillylmandelic acid urinary output test may present false positive results owing to the non-specificity of the most frequently employed methods. The following aspects influence the result: diet including tyrosine hydroxylase inhibitors (two days prior to sampling the patient should refrain from eating bananas, vanilla, soft-cheese, confectionery, pastry, chocolate, coffee, or citrus fruit), some drugs, and extensive exertion, as well as strong emotional stress before the test, which is analogous to the abovementioned factors. Furthermore, increased output of this metabolite is found in only 60–70% of phaeochromocytoma patients, which may add to the false negative results.

Chromogranin A

Identification of chromogranin A (CgA) outside gastroenteropancreatic neuroendocrine tumours (GEP-NET) may influence the diagnosis of phaeochromocytoma and other disorders (neuroblastoma, small cell lung cancer, familial medullary thyroid carcinoma, prostate cancer) [39]. CgA concentration does not depend on sex or age, but its concentration increases after a meal. Therefore, the blood sample should be collected during fasting, at least eight hours after the last meal.

It is vital to confirm that the patient is not using drugs that might influence the production and release of CgA from enterochromaffin cells vastly distributed in the alimentary canal. The aforementioned drugs include proton pump inhibitors (e.g. omeprazole), when CgA concentration may increase to the levels observed in the GEP-NET with liver metastases. This effect may depend on the duration and individual patient response, hence, if possible, these medications should be discontinued at least 7–14 days before CgA testing. Additionally, elevated CgA level may occur in the course of treatment with H2 receptors inhibitors (ranitidine, cimetidine), as well as during glucocorticoid administration, which may result in twofold CgA concentration elevation. In fact, disorders associated with elevated CgA levels may accompany a number of internal diseases [40] (Table VIII).

Due to numerous limitations and low diagnostic value, CgA level assessment is not recommended in the diagnosis of suspected phaeochromocytoma.

Even clinically silent, undiagnosed phaeochromocytomas may cause a hypertensive crisis, for instance during surgery. Taking into account the fact that in 25% of cases phaeochromocytoma may be genetic (especially in young patients, with bilateral lesions or localised outside the adrenals), it is crucial to consider genetic tests in these patients, including: multiple endocrine neoplasia type 2 (MEN), von Hippla-Lindaua Disease (VHL), as well as phaeochromocytoma-paraganglioma syndrome (PPS).

Reliable diagnosis in hyperaldosteronism (aldosterone, plasma renin activity [PRA], aldosterone-to-renin ratio assessment)

In incidentaloma patients with hypertension and/ or hypokalaemia and normokalaemia, diagnosis of hyperaldosteronism (Conn’s syndrome, PA [primary aldosteronism]) should be attempted [41]. Aldosteronetorenin ratio (ARR) constitutes a screening test that reflects the aldosterone concentration level to renin activity ratio. In fact, ARR >20–30 with increased aldosterone concentration suggests hyperaldosteronism. Moreover, aldosterone concentration should be measured in normonatraemic conditions (high-sodium diet for three days, with confirmed 200 mmol/d sodium urinary output and normokalaemia).

Plasma renin activity constitutes an indirect indicator of renin secretion, i.e. it established the amount of produced angiotensin I (ANG I) from 1 ml of blood within one hour of incubation in controlled conditions. In healthy subjects the PRA value depends on body position and salt intake. Nevertheless, there are a number of physiological and pharmacological stimuli that modify aldosterone and PRA level, thus influencing indirectly ARR (Table IX).

It is worth remembering that in women receiving oral contraceptives or hormone replacement therapy, PRA is preferred over DRC (direct renin concertation) because these medications may decrease DRC (ARO to a lesser degree) and may falsely elevate ARR level. What is more, in the luteal phase in women, due to progesterone anti-mineralocorticoid activity, aldosterone and PRA are elevated, which, as a consequence, may lead to falsely decreased ARR results. In fact, it constitutes one of the causes of unnecessary diagnostic measures. Therefore, separate norms in terms of aldosterone, PRA, and ARR are recommended not only for men and women, but also for separate menstrual cycle phases [42].

The influence of administered medications must be taken into account when interpreting ARR [43, 44]. Prior to planned diagnostics it is necessary to modify the hypotensive therapy. It entails the discontinuation of drugs that may substantially influence the results. In order to avoid false negative results, 4–6 weeks before the examination the following medication should be discontinued: aldosterone antagonists (spironolactone, eplerenone), potassium-sparing diuretics (amiloride, triamterene), and liquorice-based substances. Additionally, 2–4 weeks before examination the following drug groups should be abandoned: ACE-I, sartans, renin inhibitors (aliskiren), dihydropyridine calcium channel blockers, β-blockers, α2-adrenomimetics (false negative results). If possible, they should be replaced with long acting calcium channel blockers (verapamil retard, prazosin – Polpressin, terazosin – Hytrin, α1-adrenolytics, clonidine, Iporel; hydralazine), which hardly influence ARR values.

In order to avoid a false negative ATT result, progestogen therapy should also be discontinued for at least four weeks, and NAIDs should also be withdrawn. Decreased PRA and aldosterone values are confirmed in the following clinical situations: after 65 y. a., in patients with impaired renal function, diabetic nephropathy, Addison’s disease, tuberculosis, and congenital adrenal hyperplasia).

Adrenal androgen concentration (DHEA-S, DHEA, testosterone)

Adrenal cortex cancers constitute about 5% of AI cases where hormonal activity is confirmed in 2/3 of patients [45]. They most frequently cause rapid ACTH-independent Cushing’s syndrome, hyperandrogenism (in women) or feminisation syndrome (in men). In these cases, sexual hormone testing is crucial, such as androstenedione, testosterone, or dehydroepiandrosterone sulphate (DHEA-S), and in men with feminine features also estradiol level should be tested.

The main adrenal androgens are DHEA and DHEA-S because 95% of them are produced in the adrenal glands. Testosterone is secreted by the adrenals in small amounts. In endocrinological diagnostics, DHEA-S is employed due to the lack of circadian fluctuations and long half-time (> 20 hours).

Adrenal androgen testing may be recommended in women with hyperandrogenism, i.e. amenorrhoea, infertility, hirsutism, alopecia, acne. In these cases, total testosterone and 17(OH) progesterone should be tested. In fact, high androgen levels (TC > 200 ng/dl, DHEA-S > 800 ug/dL, 17-OH-progresterone) may accompany adrenocortical carcinoma and require the diagnosis of polycystic ovarian syndrome and congenital adrenal hyperplasia.

Vitally, in some cases of adrenocortical carcinoma, which is hormonally inactive, DEAH-S concentration may be within normal limits. Moreover, elevated or decreased DHEA-S is diagnosed in various clinical situations (Table X).

In order to differentiate between androgenism causes in adrenal tumour, dexamethasone androgen suppression test may be conducted, where cortisol, testosterone, and DHEA-S levels will be analysed. The test is performed two days after DXM introduction in 0.5 mg oral dose every six hours at 8 a.m., 2 p.m., 8 p.m., and 2 a.m. ACTH is a hormone stimulating androgen secretion in the zona reticularis of the adrenals. DHEA-S level is assessed on the third day two hours after the administration of the last dose. The absence of the decrease in DHEA-S and testosterone secretion indicates autonomous androgens secretion. On the other hand, a DHEA-S decrease > 60% and testosterone > 40% suggests adrenal secretion of androgen excess [45], whereas the decrease in DHEA-S concentration > 60% and testosterone < 40% points to ovarian origin of androgen excess.

Summary

A unified diagnostic scheme regarding incidentalomas has not been established to date. Taking into account the immense increase in the diagnosis of incidentalomas and the necessity of conducting numerous expensive diagnostic tests in the majority of healthy patients, establishing transparent and effective diagnostic paths has become urgent.

The awareness of biochemical and imaging test limitations, and the knowledge of false positive and false negative result sources, allows for the optimisation of the diagnostic process. Simultaneously, the abovementioned factors may contribute to a decrease in unnecessary and frequently repeated tests. Additionally, it may imply avoiding the costs of unjustified deep diagnostics.

On the other hand, understanding factors influencing false negative test results will not lead to a loss of the doctor’s alertness in further medical care, and will motivate them to pose a correct diagnosis.