Adrenal Tumors in Adults
CrossMark
Andre Pinto, MD, Justine A. Barletta, MD*
KEYWORDS
· Adrenal · Adrenal adenoma · Adrenocortical carcinoma · Pheochromocytoma
ABSTRACT
A Ithough most adrenal tumors are not diag- nostic dilemmas, there are cases that are challenging. This may be due to the tissue provided, for example fragmented tissue received in the setting of morcellation, or it may be due to inherently challenging histology, such as in cases with equivocal features of malignancy. Addition- ally, much has been learned about the molecular alterations of adrenal tumors, especially pheo- chromocytomas. Many of these alterations repre- sent germline mutations with significant clinical implications for patients and their families. The aim of this review is to provide an overview of the most common adrenal tumors in adults so that pathologists can tackle these interesting tumors.
OVERVIEW
This article will present an overview of the most common adrenal tumors in adults. It will start with a discussion of adrenal incidentalomas, including what they represent and when they are removed. Next adrenal cortical adenomas, the most common adrenal neoplasm, and a tumor that is generally diagnostically straight-forward will be covered. From there adrenocortical carcinomas will be re- viewed, concentrating on gross and microscopic findings and histologic features of malignancy. The article will finish with pheochromocytomas, addressing not only what they look like, but also discussing important hereditary associations.
ADRENAL INCIDENTALOMA
Adrenal incidentalomas are defined as adrenal masses larger than 1 cm that are inadvertently discovered in the course of diagnostic evaluation
or treatment of another medical condition.1 This excludes masses found in the setting of imaging performed to detect metastatic disease in a pa- tient with a known malignancy since 75% of such masses are metastases.2 Adrenal incidenta- lomas are estimated to be present in 1.5% to 9.0% of people and are found in up to 5.0% of pa- tients undergoing computed tomography of the abdomen.3 In general, the lesions are small (<3- 4 cm), men and women are equally affected, and they are most commonly detected in patients in their sixth decade of life.4 A list of underlying le- sions responsible for incidentalomas is presented in Table 1.1 There are 2 main factors to consider when deciding whether an incidentaloma should be surgically removed. The first is the functional status of the tumor, and the second is the risk of malignancy. Roughly 10% of incidentalomas are functional and fewer than 5% are malignant (see Table 1).1 A 2002 National Institutes of Health state-of-the-science statement regarding the management of patients with incidentalomas elaborated the following abbreviated conclu- sions.2 Patients with biochemical evidence of a pheochromocytoma should be surgically treated. Additionally, surgery should be considered for pa- tients with clinically apparent functional cortical tumors. Data were deemed insufficient to advo- cate for surgery or nonsurgical management of tu- mors with subclinical hyperfunctioning adrenal cortical adenomas. Because of the higher risk of malignancy with increased tumor size (adrenal cortical carcinomas account for <2% of tumors that are <4 cm, 6% of tumors that are 4.1-6 cm, and 25% of those >6 cm), it was advised that tumors larger than 6 cm should be surgically removed, those smaller than 4 cm could be followed, and those 4 to 6 cm require additional clinical data to determine whether surgery is appropriate.
Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis Street, Boston, MA 02115, USA
* Corresponding author.
E-mail address: jbarletta@partners.org
http://dx.doi.org/10.1016/j.path.2015.07.005
| Table 1 Etiology of incidentalomas | |
|---|---|
| Cause | Prevalence, %, Approximate |
| Adrenal cortical adenoma | 80 |
| Functional | 10 |
| Adrenal cortical carcinoma | 2 |
| Pheochromocytoma | 3 |
| Metastases | 1 |
| Other causesª | 15% |
a Includes adrenal cortical nodules, adrenal cysts, myeloli- pomas, hematomas/hemorrhage, infection.
Data from Cawood TJ, Hunt PJ, O’Shea D, et al. Recom- mended evaluation of adrenal incidentalomas is costly, has high false-positive rates and confers a risk of fatal can- cer that is similar to the risk of the adrenal lesion becoming malignant; time for a rethink? Euro J Endocri- nol 2009; 161(4):513-27.
ADRENAL CORTICAL ADENOMA
When evaluating an adrenal cortical adenoma, it is helpful to know the functional status of the tumor. Tumors that produce aldosterone are almost al- ways benign; whereas, production of sex steroids is an ominous sign because there are only rare re- ports of benign adenomas with sex steroid pro- duction. Production of glucocorticoids is seen with adenomas and carcinomas, although because of the much higher frequency of adrenal adenomas compared with adrenal cortical carci- noma, the vast majority of tumors that produce cortisol will be adenomas. Most adenomas are small (<5 cm) and solitary. Grossly, most ade- nomas are yellow, solid, homogeneous, and well circumscribed (Fig. 1A). Some tumors may appear heterogeneous depending on variable cyto- plasmic lipid content of tumor cells: a brighter yel- low color is seen with higher cytoplasmic lipid and a more tan color with lipid depletion. Second- ary changes, such as cystic degeneration and hemorrhage, also can occur (see Fig. 1B). Grossly, aldosterone-producing tumors may be slightly brighter yellow (“canary yellow”) compared with cortisol-producing tumors. Rarely, cortisol- producing adenomas can be dark or even black in color (see Fig. 1C). Ultimately, functional status cannot be determined with certainty by gross evaluation.
Histologically, most adenomas are well circum- scribed with a pushing border; however, in some cases a thin fibrous capsule is present (Fig. 2A). Most adenomas are similar in appearance to the zona fasciculata (see Fig. 2B). The architecture is generally nested or alveolar and less frequently
corded or trabecular. The cells are slightly larger than those of fasciculata, but have the same vacu- olated clear cytoplasm, small nuclei, and variably distinct nucleoli. The vacuolated cytoplasm is secondary to the high lipid content, which can be demonstrated by an oil-red-O stain. Some tumors have a more heterogeneous microscopic appearance (see Fig. 2C). Adenomas associated with aldosterone production are predominantly composed of cells similar to fasciculata; however, some tumors may have populations of cells that appear like fasciculata, glomerulosa, and reticularis admixed. Tumors resected in the setting of spironolactone treatment for an aldosterone- producing adenoma may demonstrate “spirono- lactone bodies,” which are cytoplasmic laminated inclusions composed of aldosterone seen in compact cells with eosinophilic cytoplasm (glomerulosa-like cells of the adenoma) (see Fig. 2D). Adenomas associated with cortisol pro- duction are often also composed of cells similar to fasciculata, but again there may be some heterogeneity with lipid-depleted cells admixed. Lipofuscin pigment is often present in these lipid- depleted cells (see Fig. 2E), and abundant lipofus- cin explains the black adenomas described previously. Functional status cannot be deter- mined with certainty with histologic evaluation. However, tumors associated with cortisol produc- tion more frequently have intracytoplasmic lipofus- cin and myelolipomatous metaplasia, whereas spironolactone bodies are virtually diagnostic of spironolactone treatment, and thus an aldosterone-producing adenoma. Additionally, non-neoplastic cortical atrophy can often be dis- cerned with cortisol-producing adenomas. In contrast, the cortex adjacent to aldosterone- producing adenomas may be normal or even show hyperplasia of the glomerulosa layer (“para- doxic hyperplasia”) with the normally patchy glo- merulosa layer forming a thick band beneath the adrenal capsule. Scattered cells or small clusters with marked nuclear atypia (Fuhrman nuclear grade 3 or 4) can be seen in benign adenomas regardless of functional status (see Fig. 2F). This atypia alone does not warrant a diagnosis of malig- nancy. In contrast, mitoses are very rarely seen in adenomas (<1 mitosis/50 high-power fields [HPFs] is typical), and atypical mitoses are virtually confined to carcinomas.
Adrenal cortical adenomas are not typically a diagnostic dilemma; however, occasionally, distin- guishing an adenoma from a non-neoplastic adre- nal cortical nodule or even normal cortex can be a challenge. Non-neoplastic adrenal cortical nod- ules are frequently seen in the setting of old age, hypertension, and diabetes. These nodules are
A
B
C
A
B
C
D
E
F
thought to be compensating for adjacent adrenal atrophy arising as a result of vascular insufficiency of adrenal arterioles. Non-neoplastic nodules also can be seen in patients with aldosterone- producing adenomas, and these nodules are thought to arise due to hypertension. In
most cases, non-neoplastic adrenal cortical nod- ules are multifocal and bilateral; however, domi- nant nodules can occur and so occasionally clinical/radiologic correlation may be required to differentiate a nonfunctional adenoma from a dominant non-neoplastic adrenal cortical nodule.
Occasionally it can be difficult to distinguish a benign adenoma from normal adrenal cortex due to the tissue provided (i.e, limited tissue in the setting of the rare adrenal core biopsy or frag- mented tissue in the setting of morcellation). If a biopsy is entirely composed of cells that are similar in appearance to fasciculata cells, this could favor adenoma; however, it should be noted that the cells of an adenoma are somewhat larger when compared with the normal fasciculata cells. As noted previously, some adenomas are composed of cells with a more heterogeneous appearance. In core biopsies of such cases, knowing that ade- nomas can appear cytologically heterogeneous is helpful. Finally, correlating the histologic features with clinical and radiologic findings is important.
ADRENOCORTICAL CARCINOMA
OVERVIEW
In contrast to adrenal cortical adenomas, adreno- cortical carcinomas (ACCs) are extremely rare, with an estimated annual incidence of 0.5 to 2.0 cases per million.5 The average age at presenta- tion is slightly under 50 years (though ACC demon- strates a bimodal distribution with a first peak in early childhood and a second higher peak in adults), and a female predominance is reported in most studies.6-9 Approximately half of ACCs are functional.6,8 Cortisol production is most com- mon, either alone or in combination with a sex ste- roid; less commonly, sex steroids are produced in isolation, and, although rare, aldosterone- producing ACCs do occur.6,8 In an analysis of Sur- veillance, Epidemiology, and End Results data, approximately 40% of patients had localized dis- ease, whereas more than half had locally invasive disease, lymph node involvement, or distant me- tastases.9 The overall 5-year survival for patients with ACC is 35% to 40%,6,8,9 with patients with localized disease demonstrating a significantly better outcome than those with distant metastatic disease at presentation (5-year survival 62% compared with 7%, respectively).º Most patients develop distant metastatic disease during follow- up, with lung, liver, and bone being the most frequent metastatic sites.6
GROSS FEATURES
ACCs are usually large, heavy tumors. The average size is 10 to 12 cm, and the average weight is approximately 300 g.6,7,9 Although a large size and weight are worrisome features for an adrenal cortical tumor, tumors smaller than 5 cm and less than 100 g can pursue an aggressive
clinical course.9,10 The tumors are yellow/tan to red/brown depending on the associated steroid production (cortisol-producing tumors are gener- ally yellow/tan whereas sex-steroid-producing ACCs are red/brown). Fibrous septa, necrosis, and hemorrhage are frequent findings, and rare cases of ACC are cystic.11 Many ACCs are clearly malignant on gross examination alone, based on a combination of tumor size, presence of necrosis, invasion into adjacent tissues, or gross involvement of associated lymph nodes (Fig. 3).
MICROSCOPIC FEATURES AND CRITERIA FOR MALIGNANCY
Histologically, ACCs can range from fairly uniform to highly heterogeneous. Many ACCs are predom- inantly composed of small cuboidal cells with eosinophilic cytoplasm with scattered markedly enlarged cells with bizarre nuclei (Fig. 4A, B). Necrosis and increased mitotic activity, including atypical mitoses, is common (see Fig. 4C, D). Tumors can be very hepatoid in appearance (a fact to keep in mind when evaluating core bi- opsy specimens of a “liver mass”) or can show a rhabdoid cytomorphology (which is significant when considering whether a biopsy might repre- sent a high-grade renal cell carcinoma with a rhabdoid component vs an ACC; the former is PAX8 positive and the latter is PAX8 negative; see later in this article for more detail). In some tu- mors clear cells predominate. In cases with this morphology, metastatic disease and pleomorphic liposarcoma should be considered in the differen- tial diagnosis (approximately a quarter of pleomor- phic liposarcomas have an epithelioid morphology resulting in a striking overlap with ACC). Although most tumors are histologically malignant throughout, some tumors have areas that may not be diagnostic of malignancy (Fig. 5). This is an important fact to keep in mind when evaluating core biopsy specimens of adrenal masses. Thus, although some core biopsy specimens may be diagnostic of malignancy, others are not. In core biopsies lacking histologic features of malignancy, it is prudent to indicate that although no features of malignancy are seen in this small core biopsy specimen, clinical and radiologic correlation is required.
The most widely used criteria for malignancy were proposed by Weiss12 3 decades ago based on the investigator’s assessment of 43 adrenocor- tical tumors. The 9 histologic features of the Weiss criteria are as follows: mitotic activity (>5 per 50 HPFs), presence of atypical mitoses, necrosis, high nuclear grade, venous invasion,
A
B
sinusoidal invasion, capsular invasion, diffuse growth accounting for more than one-third of the tumor, and clear cells comprising ≤25% of the tu- mor (Table 2). Mitotic activity should be assessed by evaluating the most proliferative area of 5 tumor slides (with 10 HPFs assessed on each slide); how- ever, if fewer than 5 tumor slides are available, then more fields should be counted per slide so that a total of 50 fields is achieved. Atypical mito- ses are defined as those with an abnormal
chromosomal distribution or an excessive number of mitotic spindles resulting in a multipolar appear- ance. Necrosis requires the involvement of a confluent area of cells (ie, apoptotic cells alone do not suffice). High nuclear grade is based on Fuhrman criteria, with Fuhrman grade 3 or 4 war- ranting a designation as high grade. In heteroge- neous tumors, nuclear grade is based on the highest degree of atypia within the tumor, even if that atypia is focal. Venous invasion implies
A
B
C
D
invasion of an endothelial-lined structure with a wall containing smooth muscle, whereas sinusoi- dal invasion reflects invasion of vessels that lack smooth muscle. In both cases, the tumor should appear as plugs of cells or polypoid projections of tumor. Because free-floating tumor in vessels
can be artifactual, this finding should be excluded when considering venous and sinusoidal invasion. Capsular invasion is defined as tumor invading into or entirely through the adrenal capsule with an associated stromal reaction. Diffuse architecture requires a sheeted-appearance, with all other
A
B
6
growth patterns (ie, nested, trabecular, corded) considered nondiffuse. And last, clear cells imply cells similar in appearance to the cells of zona fasciculata.
Each histologic feature receives 1 point when present (no weighting of features) resulting in a total score of 0 to 9. Although the score for malig- nancy was initially set at 4 or more, in a study Weiss published 5 years later, the threshold was lowered from 4 to 3 based on a patient with a tumor with a score of 3 who developed local recurrence and subsequently died due to disease.10 In this later study by Weiss, 76% of ACCs showed a mitotic count of more than 5 per 50 HPFs, 70% had
atypical mitoses, 90% had necrosis, 88% demon- strated a high Fuhrman nuclear grade, 50% had venous invasion, 57% had sinusoidal invasion, 57% had capsular penetration, 71% had a diffuse architecture comprising more than a third of the tu- mor, and 90% were composed of 25% or fewer cells with clear cytoplasm.10 All tumors with fewer than 3 criteria pursued a benign clinical course, whereas 91% of patients with a tumor score of 3 or more developed recurrent or metastatic dis- ease. The strongest predictor of clinical outcome was mitotic count. The investigators reported that patients with tumors with a mitotic count of more than 20 mitoses per 50 HPFs had a median survival
| Histologic Feature | Prevalence, % |
|---|---|
| >5 mitoses per 50 HPFs | 76 |
| Atypical mitoses | 70 |
| Necrosis | 90 |
| High nuclear grade | 88 |
| Venous invasion | 50 |
| Sinusoidal invasion | 57 |
| Capsular invasion | 57 |
| Diffuse growth >1/3rd of tumor | 71 |
| Clear cells <25% of the tumor | 90 |
| 3 or more criteria are required for a diagnosis of adrenocortical carcinoma. | |
Adapted from Weiss LM, Medeiros LJ, Vickery AL Jr. Pathologic features of prognostic significance in adreno- cortical carcinoma. Am J Surg Pathol 1989;13(3):202-6.
of 14 months, whereas those with tumors with 20 or fewer mitoses per 50 HPFs had a median survival of 58 months (P = . 02). Based on this finding, they proposed that tumors with more than 20 mitoses per 50 HPFs should be considered high grade, whereas those with 20 or fewer mitoses per 50 HPFs should be considered low grade. Subse- quently, others have advocated for this grading system based on the fact that mitotic count is not only predictive of survival (with a doubling of the mitotic rate resulting in a 4.7-fold increase in rela- tive risk of dying of ACC in the subsequent 5-year period), but also correlates with gene expression profiling.13 Although many pathologists do not grade ACCs, all reports should indicate mitotic count per 50 HPFs because it clearly is of prog- nostic significance, and in some settings, may guide treatment (see also Elfiky, Adrenal Cortical Carcinoma: A Clinician’s Perspective, Surgical Pa- thology Clinics, 2015, Volume 8, Issue 4).
Other systems have been proposed to distin- guish benign from malignant adrenal cortical tu- mors. The system proposed by Hough and colleagues 14 weighs pathologic findings, including tumor mass, diffuse growth, vascular invasion, tu- mor necrosis, fibrous bands, capsular invasion, mitotic index, and nuclear pleomorphism, and in addition takes into account clinical data, including urinary ketosteroids, response to adrenocortico- tropic hormone, presence (or absence) of Cushing syndrome or virilization, and weight loss. Because many of these clinical parameters are often not available to a pathologist, this system is not readily applied in most cases. In the system proposed by van Slooten and colleagues,15 histologic findings
are weighted. Regressive changes (necrosis, hem- orrhage, fibrosis, or calcification) receive 5.7 points, loss of normal structure 1.6 points, nuclear atypia 2.1 points, nuclear hyperchromasia 2.6 points, abnormal nucleoli 4.1 points, mitotic count of 2 per 10 HPFs 9 points, and capsular or vascular invasion 3.3 points. A score of more than 8 is asso- ciated with malignancy. Again, this system is not as easily used as the Weiss criteria. In 2002, Au- bert and colleagues 16 proposed a modification of the Weiss system by eliminating criteria more prone to low interobserver reproducibility. Instead of the 9 histologic features assessed in the original Weiss system, 5 criteria were kept. Diffuse archi- tecture, sinusoidal invasion, nuclear grade, and venous invasion were all eliminated because they had kappa values of less than 0.60. The following features were kept because all demonstrated sub- stantial interobserver reproducibility (kappa value >0.60): mitotic count greater than 5 per 50 HPFs, 25% or fewer clear cells, abnormal mitoses, ne- crosis, and capsular invasion. A mitotic count of more than 5 per 50 HPFs and 25% or fewer clear cells both receive 2 points, whereas abnormal mi- toses, necrosis, and capsular invasion all receive 1 point. A tumor can achieve a score of 0 to 7, and similar to the Weiss system, a score of 3 or more is considered indicative of malignancy. The validity of this modified Weiss system has subsequently been confirmed.17 In 2009, an Italian group, again citing issues of reproducibility for some parame- ters of the Weiss system, proposed the “reticulin algorithm.”18 This algorithm defines malignancy through an altered reticulin framework associated with 1 of 3 additional parameters: necrosis, high mitotic rate (>5 per 50 HPFs), and venous invasion. An altered reticulin network is defined as either a quantitative loss of the reticulin network that is seen in normal or adenomatous cortex or a quali- tative change. Tumors with a quantitative change demonstrate areas of tumor lacking the reticulin network altogether (Fig. 6), whereas cases with a qualitative change have an intact network composed of irregularly thickened or frayed fibers. A subsequent study demonstrated substantial interobserver reproducibility in assessing whether the reticulin network is normal or altered.19 Although the modified Weiss system and reticulin algorithm both seem valid and potentially repro- ducible, the original Weiss criteria are still used by most pathologists to assess the malignant po- tential of adrenocortical tumors.
VARIANT ADRENOCORTICAL TUMORS
There are 2 rare adrenocortical “variants” that deserve special mention because assessment of
A
B
their malignant potential is especially challenging. Oncocytic adrenocortical tumors are rare with only 147 cases described in the literature as of 2012.20 These tumors are usually incidentally discovered (20%-50% are functional, most commonly producing cortisol, although sex ste- roid and aldosterone production have also been reported), present at a mean age of 47 years (range 27-72), and demonstrate a female predom- inance (2.5:1.0).20,21 Oncocytic adrenocortical tu- mors are typically larger than their non-oncocytic counterparts with a mean size of 8 cm (range 2-20 cm).20 They are grossly well circumscribed and have a brown/mahogany cut surface. They
are composed entirely or nearly entirely of onco- cytic cells with abundant granular eosinophilic cytoplasm secondary to an accumulation of mitochondria within the cytoplasm. As Lin and colleagues22 discussed in their description of 7 on- cocytic adrenocortical neoplasms, these tumors by definition have 25% or fewer clear cells, and nearly always have a diffuse growth pattern, as well as high-grade nuclear atypia with frequent eosinophilic nuclear pseudo-inclusions (Fig. 7A, B), regardless of whether they are benign or malig- nant. As a result, most oncocytic adrenocortical tumors inherently have a Weiss score of 3. Due to this observation, in 2004 Bisceglia and
A
B
C
D
Box 1 Lin-Weiss-Bisceglia system for the assessment of malignant potential of oncocytic adrenocortical tumors
Major criteria More than 5 mitoses per 50 HPFs Atypical mitoses Venous invasion Minor criteria Tumor larger than 10 cm or greater than 200 g Necrosis Capsular invasion Sinusoidal invasion
Malignant: presence of any major criteria. Uncertain malignant potential: lack of any major criteria, but presence of any minor criteria. Benign: lack of both major and minor criteria.
colleagues23 recommended altered criteria to assess malignancy in pure (defined as tumors with >90% oncocytic cells) oncocytic adrenocor- tical neoplasms. The major criteria of this system (often referred to as the Lin-Weiss-Bisceglia sys- tem) include a mitotic count of greater than 5 per 50 HPFs, atypical mitoses, and venous invasion, whereas minor criteria include a size larger than 10 cm or a weight of more than 200 g, necrosis, capsular invasion, or sinusoidal invasion (Box 1). If any major criteria are present, the tumor should be considered malignant, if any minor criteria are present, the tumor should be considered of uncer- tain malignant potential, and if no criteria are pre- sent, then the tumor can be considered benign. These criteria are not to be used for adrenocortical tumors with fewer than 90% oncocytic cells. In cases with a lesser oncocytic component, use of Weiss criteria has been advocated. However, given the rarity of such cases, a note indicating some degree of uncertainty of malignant potential is probably warranted. Interestingly, the reticulin algorithm (described previously) has been shown to be helpful in both pure oncocytic adrenocortical tumors and those with a lesser oncocytic compo- nent.19 The prognosis of oncocytic ACCs appears to be somewhat better than that of non-oncocytic ACCs, and most, although not all, borderline tu- mors have pursued a benign clinical course. 19,21
Even more uncommon than oncocytic tumors are myxoid adrenocortical tumors. These tumors are often functional (most commonly producing cortisol, although sex steroid and aldosterone pro- duction have also been reported), present at a
mean age of roughly 50, and demonstrate a female predominance.24 The mean tumor size is approxi- mately 12 cm (range 2-20 cm), and on cut section, tumors with an abundant myxoid component appear translucent gray or white.24 Histologically, the myxoid component can comprise 5% to 95% of the tumor. Tumors that are largely myxoid are characterized by a uniform population of small regular cells (often with cells reminiscent to those of zona glomerulosa) with a corded, nested, pseu- doglandular, or microcystic architecture (see Fig. 7C, D). ACCs may also show focal myxoid change. In these cases, the cells resemble con- ventional ACC. Myxoid tumors are evaluated by Weiss criteria. Although benign myxoid adreno- cortical tumors with reasonable follow-up time have been reported, the amount of follow-up data are limited and there is one case in the litera- ture that pursued a malignant course despite a Weiss score of 1.24,25 Therefore, for cases with myxoid features that do not meet Weiss criteria for malignancy, it might be best to consider these tumors of uncertain malignant potential.
IMMUNOHISTOCHEMISTRY
The immunohistochemical profile of ACCs is sum- marized in Table 3. ACCs are positive for inhibin, synaptophysin, A103 (Melan-A), calretinin, and SF-1 (steroidogenic factor 1); negative to weakly positive for keratins (Cam5.2 shows the most expression); negative for PAX8; and virtually al- ways negative for chromogranin.26 Markers like inhibin and Melan-A are highly specific for ACC. Pheochromocytomas and virtually all carcinomas that might metastasize to the adrenal are negative for these 2 markers. However, the sensitivity of inhibin and Melan-A is 70% to 90%, with inhibin being more sensitive than Melan-A.26,27 As a result, SF-1 has become a very useful stain, as nu- clear SF-1 expression is both very specific and very sensitive for ACC, with a sensitivity of 95% (Fig. 8).27,28 This transcription factor is expressed in steroidogenic organs like the adrenal glands, testes, and ovaries, and plays a key role in the development of steroidogenic tissue and the regulation of steroid biosynthesis.27 Additionally, there is some evidence that tumors with a high level of SF-1 expression may be more clinically aggressive than tumors with low-level SF-1 expression.27 Chromogranin is often used when trying to differentiate an ACC from a pheochromo- cytoma because pheochromocytomas are posi- tive for both synaptophysin and chromogranin. The vast majority of ACCs are negative for chro- mogranin; however, rare cases of ACC may show some degree of chromogranin expression.26
| Pheochromocytoma | Adrenocortical Carcinoma | |
|---|---|---|
| Chromogranin | Positive | Negativeª |
| Synaptophysin | Positive | Positive |
| MelanA 103/Mart-1 | Negativeª | Positive |
| Calretinin | Negativeª | Positive |
| Inhibin | Negativeª | Positive |
| PAX8 | Negative | Negative |
| Keratins | Negativeª | Weak/Negative |
| SF-1 | Negative | Positive |
| GATA3 | Positive | Usually negativeb |
a <5% of cases are positive.
b ~10% of cases are positive.
Although these rare cases do not decrease the utility of chromogranin, they do indicate that immunohistochemical stains should be part of a broader panel and their results need to be inter- preted in the context of the morphology of the tumor.
There are also several immunohistochemical stains that have been reported either to aid in the assessment of malignant potential of adrenocor- tical tumors or have prognostic value in the setting of ACC. The Ki-67 proliferative index has been consistently shown to be higher in ACC than in ad- enomas. However, while the vast majority of ade- nomas have a Ki-67 proliferative index of less than 5%,29,30 not all ACCs will demonstrate a pro- liferative index of greater than 5%. This means that although a proliferative index above 5% war- rants concern, tumors with a proliferative index less than 5% cannot be assumed to be benign. As a consequence, many pathologists do not routinely report a Ki-67 proliferative index. But this may soon change, because a recent study by Beuschlein and colleagues7 has provided compel- ling evidence that Ki-67 is a strong predictor of outcome that could guide therapy in localized ACCs (see also Elfiky, Adrenal Cortical Carcinoma: A Clinician’s Perspective, Surgical Pathology Clinics, 2015, Volume 8, Issue 4). They found that Ki-67 indices of less than 10%, 10% to 19%, and 20% or more provided highly significant differ- ences for both recurrence-free survival (RFS) and overall survival (OS), translating into a median RFS and a median OS of 53.2 and 180.5 months for patients with tumors with a Ki-67 proliferative index of less than 10%, 31.6 and 113.5 months for patients with tumors with a Ki-67 proliferative in- dex of 10% to 19%, and 9.4 and 42.0 months for patients with tumors with a Ki-67 proliferative index
of 20% or more. Moreover, they demonstrated that the prognostic value of the Ki-67 proliferative index was maintained in multivariate analysis, and they confirmed the prognostic significance of the Ki-67 proliferative index in a second validation cohort. Additional markers that are preferentially ex- pressed in ACCs compared with adenomas include insulin-like growth factor-2 (IGF2), matrix metallo- protease type 2 (MMP2), p53, and aberrant cyto- plasmic/nuclear localization of B-catenin.29,31,32 Erickson and colleagues31 reported that IGF2 is ex- pressed at a higher frequency and at a higher level in ACCs compared with carcinomas, but their re- sults also indicated that IGF2 expression does not definitively reflect malignancy. Volante and col- leagues32 reported that approximately 75% of ACCs showed focal to diffuse MMP2 expression compared with only 2% of adenomas; thus, MMP2 expression could be used to support a ma- lignant diagnosis, but could not be used to exclude it. Similarly, although p53 overexpression is virtu- ally confined to carcinomas, it is seen in a relatively small percentage of ACCs.33 Finally, although diffuse aberrant cytoplasmic/nuclear B-catenin expression has been seen more in the setting of ACCs compared with adrenocortical adenomas, a subset of adenomas harbor B-catenin mutations, and therefore aberrant B-catenin expression is not definitive for malignancy.34
PHEOCHROMOCYTOMA
OVERVIEW
Pheochromocytomas arise from catecholamine- producing chromaffin cells of the adrenal medulla. They are rare tumors with an average age at diagnosis of 40 to 45 years and a roughly equal
A
B
sex distribution.35 The signs and symptoms of pheochromocytoma are secondary to catechol- amine excess and include hypertension (either sustained or paroxysmal), tachycardia, head- aches, palpitations, diaphoresis, chest pain, anxi- ety, and weight loss. Although most patients are hypertensive, roughly 10% are not. It is commonly thought that 10% of pheochromocytomas are ma- lignant; however, that percentage is closer to 5% with the 10% reflecting inclusion of extra-adrenal sympathetic paragangliomas (see also Arias-Stella and Williamson, Updates in Benign Lesions of the Genitourinary Tract, Surgical Pathology Clinics, 2015, Volume 8, Issue 4), which have a higher
rate of malignancy.35 For patients with malignant pheochromocytomas, the 5-year survival is approximately 50%.36
GROSS FEATURES
Pheochromocytomas are encapsulated or well- circumscribed tumors with a pink/tan cut surface (Fig. 9A). Hemorrhage is frequent, and larger tu- mors may also demonstrate cystic change (see Fig. 9B, C). It may be evident that the tumor is arising from the medulla; however, this can be difficult to appreciate when the tumor is large. Sporadic pheochromocytomas are almost always
A
B
C
solitary masses, whereas familial cases are more frequently bilateral and, in the setting of multiple endocrine neoplasia 2 (MEN) 2, may arise in a background of diffuse or nodular medullary hyperplasia.
MICROSCOPIC FEATURES
Pheochromocytomas are microscopically well demarcated from the adjacent adrenal tissue with some tumors demonstrating a thin fibrous capsule. Characteristically these tumors have “zellballen” (cell balls) morphology; that is, an alveolar architecture with nests of
neuroendocrine cells surrounded by delicate blood vessels and sustentacular cells (Fig. 10A). The zellballen architecture in some tumors is diffi- cult to appreciate with the architecture appearing more diffuse or solid (see Fig. 10B). The cytology of the tumor cells typically resembles non- neoplastic chromaffin cells: they are intermediate to large in size with abundant amphophilic to basophilic, granular cytoplasm (see Fig. 10). However, in some cases the cells may be smaller and may have clear cytoplasm (Fig. 11A), mimicking an adrenal cortical adenoma or a renal cell carcinoma. Other tumors may have a moder- ate amount of eosinophilic cytoplasm (see
A
B
A
B
Fig. 11B), resulting in a tumor that resembles ACC. Intracytoplasmic eosinophilic hyaline glob- ules are often present (these Periodic-acid- Schiff-positive globules are likely derived from the membrane components of secretory granules) (Fig. 12A). The tumor cell nuclei may be pleomorphic, hyperchromatic, or spindled (see Fig. 12B). Sustentacular cells within pheochromo- cytomas are generally inconspicuous on hematox- ylin and eosin, but can be highlighted by S100 immunohistochemical staining (Fig. 13A, B). These support cells, although actually non-neoplastic, can even populate metastatic deposits.
CRITERIA FOR MALIGNANCY
There are no universally accepted histologic find- ings to classify a pheochromocytoma as benign or malignant. Instead, malignancy is defined sim- ply by the presence of metastatic disease. To exclude multicentric disease, the metastatic site must be one that normally lacks chromaffin tissue (ie, lymph nodes, bone, liver, and lung). Surpris- ingly, both locally invasive growth and lymphovas- cular invasion are poor predictors of metastasis, as tumors that are confined to the adrenal may subsequently metastasize, and benign tumors
A
B
without known metastatic disease may show lym- phovascular invasion. In a 1990 study by Linnoila and colleagues,37 the investigators evaluated 120 pheochromocytomas and extra-adrenal sympa- thetic paragangliomas and found that more than 70% could be classified as benign or malignant on the basis of extra-adrenal location (with a signif- icantly higher percentage of extra-adrenal sympa- thetic paragangliomas being malignant than pheochromocytomas, 52% compared with 6%, respectively), coarse nodularity of the primary tu- mor, confluent necrosis, and absence of hyaline globules. Most (71%) malignant tumors had 2 to 3 of these characteristics, whereas 89% of benign
tumors had only 1 or none of these features. In 2002, Thompson36 proposed the “pheochromocy- toma of the adrenal gland scaled score” (PASS) to separate benign from malignant tumors. He evalu- ated 100 pheochromocytomas from the files of the Armed Forces Institute of Pathology, 50 of which were malignant based on his histologic criteria and 50 of which were benign. Neither size nor weight distinguished benign from malignant cases, and no one histologic parameter predicted malignancy. However, malignant pheochromocy- tomas more frequently demonstrated vascular invasion (1 point), capsular invasion (1 point), inva- sion into periadrenal adipose tissue (2 points),
A
B
C
D
focal or confluent necrosis (2 points), large nests (defined as 3-4 times the size of a zellballen nest) or confluent growth (2 points), high cellularity (2 points), a mitotic count greater than 3 per 10 HPFs (2 points), atypical mitoses (2 points), tumor cell spindling (2 points), cellular monotony (2 points), profound nuclear pleomorphism (1 point), and hyperchromasia (1 point). Tumors with a PASS score of 4 or more were deemed to have po- tential for biologically aggressive behavior, with 33 of 50 tumors in this category demonstrating a ma- lignant clinical course. Tumors with a score of less than 4 all demonstrated a benign clinical course over a 10-year follow-up period. Appropriately, the investigator indicated that a score of 4 or more could not be equated with malignancy, given that 34% of such tumors in his cohort were not clin- ically malignant; however, he indicated that these tumors are more at risk for pursuing a malignant clinical course and so deserve closer clinical follow-up. Thompson36 also found that malignant tumors had a decrease in the number of sustentac- ular cells as assessed by S100 immunohistochem- ical staining. This finding is consistent with other reports, and likely reflects the histologic observa- tion that malignant tumors tend to have a more diffuse architecture. After the PASS was published, Wu and colleagues38 evaluated the reproducibility of PASS among a group of 5 multi- institutional pathologists with at least 10 years of experience in endocrine pathology. They found that there was significant interobserver and intra- observer variation in assignment of PASS. The features that showed higher interobserver repro- ducibility were lymphovascular invasion, capsular penetration, invasion into extra-adrenal adipose tissue, atypical mitoses, and necrosis, whereas features that showed relatively lower interobserver reproducibility included nuclear hyperchromasia, nuclear pleomorphism, increased mitoses, tumor cell spindling, tumor cell monotony, and diffuse growth. Based on these findings, the investigators concluded that they could not recommend the use of PASS for clinical prognostication. Although PASS is not used by most pathologists, and while we still lack definitive histologic criteria of malig- nancy, it is prudent to mention histologic features that have been associated with a malignant outcome in a pathology report. For example, if a tu- mor demonstrates necrosis, increased mitotic ac- tivity, atypical mitoses, invasion into peri-adrenal adipose tissue, or vascular invasion, these features should be reported, with a note indicating that although there are no definite histologic criteria for malignancy in pheochromocytomas, the tumor demonstrates concerning histologic findings and close clinical follow-up is advised.
IMMUNOHISTOCHEMISTRY
The immunohistochemical profile of pheo- chromocytomas is summarized in Table 3. Pheo- chromocytomas exhibit immunopositivity for synaptophysin and chromogranin A. Keratins are almost always negative (which is useful when dis- tinguishing a pheochromocytoma/paraganglioma from a neuroendocrine carcinoma), although cases of pheochromocytoma with keratin positiv- ity have been reported.36 The sustentacular cells are positive for S-100 protein, and rarely the chro- maffin cells may exhibit some degree of S100 positivity. Pheochromocytomas are negative for SF1, inhibin, calretinin, and Melan-A. Recently the transcription factor GATA3 has been shown to be positive in pheochromocytomas and para- gangliomas (see Fig. 13C, D). Miettinen and colleagues39 reported GATA3 positivity in 82% of paragangliomas and 92% of pheochromocy- tomas. Along with keratins, it appears that GATA3 can be helpful in differentiating pheochro- mocytoma/paraganglioma from other neuroendo- crine tumors, such as small cell carcinoma, Merkel cell carcinoma, pulmonary carcinoid, and small intestine and pancreatic neuroendocrine tu- mors, as all these tumors lacked GATA3 expres- sion in the cohort of Miettinen and colleagues.39 It should be noted that GATA3 was positive in 11% of ACCs evaluated in this study; thus, it is not a good stain to use to differentiate pheochro- mocytoma from an ACC.
Hereditary Syndromes
For years it was thought that 10% of pheochromo- cytomas were hereditary; however, we now know that the percentage is roughly doubled if consid- ering pheochromocytomas alone and more than tripled if pheochromocytomas and paraganglio- mas are grouped together.35,40 Pheochromocy- tomas were known to be associated with MEN2A and 2B (with an underlying activating mutation in the RET proto-oncogene that results in activation of the encoded tyrosine kinase receptor), VHL dis- ease (with an underlying inactivating mutation in the tumor suppressor gene VHL), and NF1 (with an underlying mutation in the NF1 tumor suppres- sor gene) (Table 4). The risk of developing a pheo- chromocytoma in these syndromes is 50% for MEN2A and 2B, 10% to 26% for VHL disease, and 0.1% to 6% for NF1.41 Although MEN2 syn- dromes demonstrate the highest penetrance for the development of a pheochromocytoma, given that VHL disease is more common, VHL accounts for a higher percentage of hereditary pheochromo- cytomas. Although patients with VHL and NF1 develop pheochromocytomas more frequently
| Syndrome | Mutation | Proportion of All PCCs/PGLs, % | Penetrance of PCC/PGL, % | Frequency of Malignancy, % |
|---|---|---|---|---|
| MEN2 | RET | 5 | 50 | 3 |
| VHL | VHL | 9.0 | 10-26 | 3 |
| NF1 | NF1 | 3 | 0.1-6 | 9 |
| HPGL/PCC | SDHD | 7 | 86 | 3.5 |
| HPGL/PCC | SDHC | 0.5 | Unknown | 0 |
| HPGL/PCC | SDHB | 5.5 | 77 | 31 |
| Sporadic | None | 70 | Not applicable | 9 |
Abbreviations: HPGL/PCC, Hereditary paraganglioma/pheochromocytoma syndrome; MEN2, multiple endocrine neoplasia 2; SDH, succinate dehydrogenase.
Data from Welander J, Soderkvist P, Gimm O. Genetics and clinical characteristics of hereditary pheochromocytomas and paragangliomas. Endocr Relat Cancer 2011;18(6):R253-76.
than paragangliomas, paragangliomas are also associated with these syndromes. In contrast, patients with MEN2 virtually never develop para- gangliomas. Very rarely, pheochromocytomas can be seen in the setting of Carney triad, Carney-Stratakis syndrome, and multiple endo- crine neoplasia type 1 (MEN 1).41
In a landmark study by Neumann and col- leagues, 42 the investigators found that not 10%, but 24% of patients who presented with nonsyn- dromic pheochromocytomas (or paragangliomas) without a family history had germline mutations. This was in part due to more frequent germline mutations in RET and VHL than previously recog- nized, but also because succinate dehydroge- nase (SDH) mutations were now assessed. SDH is an enzyme complex localized to the inner mito- chondrial membrane that plays a role in cellular metabolism, participating in both the Krebs cycle and the electron transport chain.43 It is a hetero- tetrameric complex composed of 4 protein sub- units (SDHA, SDHB, SDHC, and SDHD). The SDH genes act as tumor suppressor genes: tu- mors demonstrate loss of heterozygosity in com- bination with germline inactivating mutations. In pheochromocytomas the rate of SDH deficiency is 4% to 5% (the rate is significantly higher in par- agangliomas, up to approximately 50% in some studies), with a roughly equal number of SDHB and SDHD mutations.35,44,45 SDHC mutations do not significantly contribute to the development of pheochromocytomas, and only rare pheochro- mocytomas harbor an SDHA mutation. 46,47 SDHAF2, a gene encoding a protein that is involved in stabilization of the SDH complex, germline mutations have been reported in head and neck paragangliomas, but not in pheochromocytomas. 48,49
Recently, 2 additional genes have been linked to the development of hereditary pheochromocy- tomas. TMEM127 encodes a transmembrane pro- tein that functions as a tumor suppressor. Yao and colleagues50 evaluated a cohort of 990 patients with pheochromocytomas or paragangliomas that were negative for RET, VHL, and SDH muta- tions and found germline TMEM127 mutations in 2% of cases. Only a quarter of these patients had a clear family history, suggesting a fairly low penetrance. Additionally, MAX, the MYC- associated factor X gene, mutations have been found to be responsible for approximately 1% of pheochromocytomas/paragangliomas in patients without evidence of other known mutations.51 MAX acts as a tumor suppressor gene, shows a paternal mode of inheritance, and is frequently associated with bilateral pheochromocy- tomas.51,52 Finally, it is worth noting that many sporadic tumors harbor somatic mutations of the genes responsible for the development of the familial syndromes discussed previously. Approximately 10% of pheochromocytomas/par- agangliomas harbor somatic VHL mutations, 10% harbor NF1 mutations, 5% have RET muta- tions, and 2.5% have MAX mutations.51,53,54
Although clinical history can significantly aid in identifying patients with pheochromocytomas that are most likely to have an underlying heredi- tary syndrome, clinical history may be deceiving. As the study by Neumann and colleagues42 demonstrated, 24% of patients with an underlying germline mutation lack a positive family history. This can be because of gaps in a patient’s knowl- edge of his or her relatives’ medical history, due to a new germline mutation in the patient that was not present in the parents, due to incom- plete penetrance of syndromes, or confounding
A
B
G
D
effects of a paternal mode of inheritance. Besides family history, other factors such as age, multifo- cality, and adrenal versus an extra-adrenal loca- tion can be used to stratify risk of an underlying hereditary syndrome. Neumann and colleagues42 found that younger age, multifocal tumors, and extra-adrenal location were significantly associ- ated with the presence of a mutation. However, among the patients who were positive for muta- tions, only 32% had multifocal tumors, and 35% presented after the age of 30 years, with 8% pre- senting after the age of 40. In a large Italian cohort, of the 278 patients who presented with unilateral pheochromocytomas, 21% were associated with germline mutations (9% VHL, 6% RET, 3% NF1, 2% SDHB, and 1% SDHD mutation). 35
Identifying patients with hereditary syndromes has significant clinical implications. For one, iden- tification of a syndrome prompts evaluation/ screening for other tumors associated with each syndrome, such as medullary thyroid carcinoma in the setting of MEN 2A or 2B, or renal cell carci- noma and hemangioblastoma in the setting of VHL disease. Additionally, identification of a hereditary syndrome can prompt genetic counseling and evaluation of family members at risk of disease. Finally, the underlying mutation may have prog- nostic significance (see Table 4). The rate of malig- nancy in pheochromocytomas in patients with germline RET or VHL germline mutations is signif- icantly lower than that of sporadic pheochromocy- tomas. The reported rate of malignancy for tumors with SDHB mutations ranges from 21% to 71%, which is significantly higher than that for tumors with SDHD mutations (approximately 3%) and SDHC mutations (virtually always benign). 35,41,42 Moreover, even among patients with malignant pheochromocytomas and paragangliomas, the presence of SDHB mutations confers a worse prognosis, with a significant and independent as- sociation with mortality (relative risk 2.7).55 The fact that SDHB mutations are prognostically signif- icant is especially important, given that there are no reliable histologic features that predict malig- nancy for paragangliomas and pheochromocy- tomas. Additionally, it appears that MAX mutations also may be associated with a higher risk of malignancy.52
Can tumors associated with a genetic syndrome be recognized in pathology? The answer is sometimes. As mentioned previously, pheochro- mocytomas in patients with MEN2A and 2B often arise in a background of diffuse or nodular medullary hyperplasia. There is also some indica- tion that pheochromocytomas occurring in the context of VHL have certain histologic features, such as a thick fibrous capsule, myxoid or
hyalinized stroma, more clear cells, and a lack of hyaline globules.56 However, VHL-associated tu- mors are not histologically distinct enough to allow for definitive identification based on histologic characteristics alone; thus, genetic testing is ulti- mately required to assess for a VHL mutation. Pheochromocytomas and paragangliomas arising in the setting of an SDH mutation may show worri- some features, such as increased mitotic activity; however, overall these tumors lack unique histo- logic features. Immunohistochemistry, on the other hand, is excellent at identifying tumors with an underlying SDH mutation. A mutation in any one of the SDH genes (SDHA, B, C, or D) not only results in a lack of SDH enzyme activity, it also causes destabilization of the SDH protein complex. Destabilization of the complex, as a result of a mutation in any of the 4 SDH genes, can be identified by immunohistochemical anal- ysis for SDHB. Although SDHB is ubiquitously ex- pressed (including in normal tissues of patients harboring a germline mutation in an SDH subunit gene), SDH-deficient tumors nearly always lack SDHB expression by immunohistochemistry (loss of SDHB staining is both very sensitive and specific for the detection of an underlying SDH germline mutation) (Fig. 14).57 For the rare pheo- chromocytomas with germline SDHA mutations, immunohistochemical staining for SDHA can be used specifically to detect the SDHA mutation. Tu- mors with SDHB, SDHC, or SDHD germline muta- tions will show loss of staining for SDHB but intact staining for SDHA (why the SDHA protein is not degraded in the presence of SDHB, SDHC, or SDHB mutations is not currently known). In contrast, tumors with SDHA germline mutations will show loss of immunohistochemical staining for both SDHB and SDHA. 46
REFERENCES
1. Cawood TJ, Hunt PJ, O’Shea D, et al. Recommen- ded evaluation of adrenal incidentalomas is costly, has high false-positive rates and confers a risk of fatal cancer that is similar to the risk of the adrenal lesion becoming malignant; time for a rethink? Eur J Endocrinol 2009; 161(4):513-27.
2. NIH state-of-the-science statement on management of the clinically inapparent adrenal mass (“incidenta- loma”). NIH Consensus State Sci Statements 2002; 19(2):1-25.
3. Gopan T, Remer E, Hamrahian AH. Evaluating and managing adrenal incidentalomas. Cleve Clin J Med 2006;73(6):561-8.
4. Reginelli A, Di Grezia G, Izzo A, et al. Imaging of ad- renal incidentaloma: our experience. Int J Surg 2014;12(Suppl 1):S126-31.
5. Dackiw AP, Lee JE, Gagel RF, et al. Adrenal cortical carcinoma. World J Surg 2001;25(7):914-26.
6. Ayala-Ramirez M, Jasim S, Feng L, et al. Adrenocor- tical carcinoma: clinical outcomes and prognosis of 330 patients at a tertiary care center. Eur J Endocri- nol 2013;169(6):891-9.
7. Beuschlein F, Weigel J, Saeger W, et al. Major prog- nostic role of Ki67 in localized adrenocortical carci- noma after complete resection. J Clin Endocrinol Metab 2015; 100(3):841-9.
8. Crucitti F, Bellantone R, Ferrante A, et al. The Italian Registry for Adrenal Cortical Carcinoma: analysis of a multiinstitutional series of 129 patients. The ACC Italian Registry Study Group. Surgery 1996;119(2): 161-70.
9. Paton BL, Novitsky YW, Zerey M, et al. Outcomes of adrenal cortical carcinoma in the United States. Sur- gery 2006;140(6):914-20 [discussion: 919-20].
10. Weiss LM, Medeiros LJ, Vickery AL Jr. Pathologic features of prognostic significance in adrenocortical carcinoma. Am J Surg Pathol 1989; 13(3):202-6.
11. Erickson LA, Lloyd RV, Hartman R, et al. Cystic ad- renal neoplasms. Cancer 2004;101(7): 1537-44.
12. Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am J Surg Pathol 1984;8(3): 163-9.
13. Giordano TJ. The argument for mitotic rate-based grading for the prognostication of adrenocortical carcinoma. Am J Surg Pathol 2011;35(4):471-3.
14. Hough AJ, Hollifield JW, Page DL, et al. Prognostic factors in adrenal cortical tumors. A mathematical analysis of clinical and morphologic data. Am J Clin Pathol 1979;72(3):390-9.
15. van Slooten H, Schaberg A, Smeenk D, et al. Morphologic characteristics of benign and malig- nant adrenocortical tumors. Cancer 1985;55(4): 766-73.
16. Aubert S, Wacrenier A, Leroy X, et al. Weiss system revisited: a clinicopathologic and immunohisto- chemical study of 49 adrenocortical tumors. Am J Surg Pathol 2002;26(12):1612-9.
17. van’t Sant HP, Bouvy ND, Kazemier G, et al. The prognostic value of two different histopathological scoring systems for adrenocortical carcinomas. His- topathology 2007;51(2):239-45.
18. Volante M, Bollito E, Sperone P, et al. Clinicopatho- logical study of a series of 92 adrenocortical carci- nomas: from a proposal of simplified diagnostic algorithm to prognostic stratification. Histopathology 2009;55(5):535-43.
19. Duregon E, Fassina A, Volante M, et al. The reticulin algorithm for adrenocortical tumor diagnosis: a mul- ticentric validation study on 245 unpublished cases. Am J Surg Pathol 2013;37(9):1433-40.
20. Mearini L, Del Sordo R, Costantini E, et al. Adrenal oncocytic neoplasm: a systematic review. Urol Int 2013;91(2):125-33.
21. Wong DD, Spagnolo DV, Bisceglia M, et al. Onco- cytic adrenocortical neoplasms-a clinicopathologic study of 13 new cases emphasizing the importance of their recognition. Hum Pathol 2011;42(4):489-99.
22. Lin BT, Bonsib SM, Mierau GW, et al. Oncocytic adrenocortical neoplasms: a report of seven cases and review of the literature. Am J Surg Pathol 1998;22(5):603-14.
23. Bisceglia M, Ludovico O, Di Mattia A, et al. Adreno- cortical oncocytic tumors: report of 10 cases and re- view of the literature. Int J Surg Pathol 2004;12(3): 231-43.
24. Papotti M, Volante M, Duregon E, et al. Adrenocor- tical tumors with myxoid features: a distinct morpho- logic and phenotypical variant exhibiting malignant behavior. Am J Surg Pathol 2010;34(7):973-83.
25. Brown FM, Gaffey TA, Wold LE, et al. Myxoid neo- plasms of the adrenal cortex: a rare histologic variant. Am J Surg Pathol 2000;24(3):396-401.
26. Weissferdt A, Phan A, Suster S, et al. Adrenocortical carcinoma: a comprehensive immunohistochemical study of 40 cases. Appl Immunohistochem Mol Mor- phol 2014;22(1):24-30.
27. Sbiera S, Schmull S, Assie G, et al. High diagnostic and prognostic value of steroidogenic factor-1 expression in adrenal tumors. J Clin Endocrinol Metab 2010;95(10):E161-71.
28. Duregon E, Volante M, Giorcelli J, et al. Diagnostic and prognostic role of steroidogenic factor 1 in adre- nocortical carcinoma: a validation study focusing on clinical and pathologic correlates. Hum Pathol 2013; 44(5):822-8.
29. Schmitt A, Saremaslani P, Schmid S, et al. IGFII and MIB1 immunohistochemistry is helpful for the differ- entiation of benign from malignant adrenocortical tu- mours. Histopathology 2006;49(3):298-307.
30. Vargas MP, Vargas HI, Kleiner DE, et al. Adrenocor- tical neoplasms: role of prognostic markers MIB-1, P53, and RB. Am J Surg Pathol 1997;21(5):556-62.
31. Erickson LA, Jin L, Sebo TJ, et al. Pathologic fea- tures and expression of insulin-like growth factor-2 in adrenocortical neoplasms. Endocr Pathol 2001; 12(4):429-35.
32. Volante M, Sperone P, Bollito E, et al. Matrix metallo- proteinase type 2 expression in malignant adreno- cortical tumors: diagnostic and prognostic significance in a series of 50 adrenocortical carci- nomas. Mod Pathol 2006; 19(12): 1563-9.
33. Reincke M, Karl M, Travis WH, et al. p53 mutations in human adrenocortical neoplasms: immunohisto- chemical and molecular studies. J Clin Endocrinol Metab 1994;78(3):790-4.
34. Tissier F, Cavard C, Groussin L, et al. Mutations of beta-catenin in adrenocortical tumors: activation of the Wnt signaling pathway is a frequent event in both benign and malignant adrenocortical tumors. Cancer Res 2005;65(17):7622-7.
35. Mannelli M, Castellano M, Schiavi F, et al. Clinically guided genetic screening in a large cohort of Italian patients with pheochromocytomas and/or functional or nonfunctional paragangliomas. J Clin Endocrinol Metab 2009;94(5):1541-7.
36. Thompson LD. Pheochromocytoma of the Adrenal gland Scaled Score (PASS) to separate benign from malignant neoplasms: a clinicopathologic and immunophenotypic study of 100 cases. Am J Surg Pathol 2002;26(5):551-66.
37. Linnoila RI, Keiser HR, Steinberg SM, et al. Histopa- thology of benign versus malignant sympathoadre- nal paragangliomas: clinicopathologic study of 120 cases including unusual histologic features. Hum Pathol 1990;21(11):1168-80.
38. Wu D, Tischler AS, Lloyd RV, et al. Observer varia- tion in the application of the Pheochromocytoma of the Adrenal Gland Scaled Score. Am J Surg Pathol 2009;33(4):599-608.
39. Miettinen M, McCue PA, Sarlomo-Rikala M, et al. GATA3: a multispecific but potentially useful marker in surgical pathology: a systematic analysis of 2500 epithelial and nonepithelial tumors. Am J Surg Pathol 2014;38(1):13-22.
40. Dluhy RG. Pheochromocytoma-death of an axiom. N Engl J Med 2002;346(19):1486-8.
41. Welander J, Soderkvist P, Gimm O. Genetics and clinical characteristics of hereditary pheochromocy- tomas and paragangliomas. Endocr Relat cancer 2011;18(6):R253-76.
42. Neumann HP, Bausch B, McWhinney SR, et al. Germ-line mutations in nonsyndromic pheochromo- cytoma. N Engl J Med 2002;346(19):1459-66.
43. Barletta JA, Hornick JL. Succinate dehydrogenase- deficient tumors: diagnostic advances and clinical implications. Adv Anat Pathol 2012; 19(4): 193-203.
44. Amar L, Bertherat J, Baudin E, et al. Genetic testing in pheochromocytoma or functional paraganglioma. J Clin Oncol 2005;23(34):8812-8.
45. Burnichon N, Rohmer V, Amar L, et al. The succinate dehydrogenase genetic testing in a large prospec- tive series of patients with paragangliomas. J Clin Endocrinol Metab 2009;94(8):2817-27.
46. Korpershoek E, Favier J, Gaal J, et al. SDHA immu- nohistochemistry detects germline SDHA gene mutations in apparently sporadic paragangliomas
and pheochromocytomas. J Clin Endocrinol Metab 2011;96(9):E1472-6.
47. Schiavi F, Boedeker CC, Bausch B, et al. Predictors and prevalence of paraganglioma syndrome associ- ated with mutations of the SDHC gene. JAMA 2005; 294(16):2057-63.
48. Bayley JP, Kunst HP, Cascon A, et al. SDHAF2 mutations in familial and sporadic paraganglioma and phaeochro- mocytoma. Lancet Oncol 2010;11(4):366-72.
49. Hao HX, Khalimonchuk O, Schraders M, et al. SDH5, a gene required for flavination of succinate dehydro- genase, is mutated in paraganglioma. Science 2009;325(5944):1139-42.
50. Yao L, Schiavi F, Cascon A, et al. Spectrum and prevalence of FP/TMEM127 gene mutations in pheo- chromocytomas and paragangliomas. JAMA 2010; 304(23):2611-9.
51. Burnichon N, Cascon A, Schiavi F, et al. MAX muta- tions cause hereditary and sporadic pheochromocy- toma and paraganglioma. Clin Cancer Res 2012; 18(10):2828-37.
52. Comino-Mendez I, Gracia-Aznarez FJ, Schiavi F, et al. Exome sequencing identifies MAX mutations as a cause of hereditary pheochromocytoma. Nat Genet 2011;43(7):663-7.
53. Burnichon N, Buffet A, Parfait B, et al. Somatic NF1 inactivation is a frequent event in sporadic pheochro- mocytoma. Hum Mol Genet 2012;21(26):5397-405.
54. Burnichon N, Vescovo L, Amar L, et al. Integrative genomic analysis reveals somatic mutations in pheochromocytoma and paraganglioma. Hum Mol Genet 2011;20(20):3974-85.
55. Amar L, Baudin E, Burnichon N, et al. Succinate dehy- drogenase B gene mutations predict survival in patients with malignant pheochromocytomas or paraganglio- mas. J Clin Endocrinol Metab 2007;92(10):3822-8.
56. Koch CA, Mauro D, Walther MM, et al. Pheochromo- cytoma in Von Hippel-Lindau disease: distinct histopathologic phenotype compared to pheochro- mocytoma in multiple endocrine neoplasia type 2. Endocr Pathol 2002;13(1):17-27.
57. van Nederveen FH, Gaal J, Favier J, et al. An immu- nohistochemical procedure to detect patients with paraganglioma and phaeochromocytoma with germline SDHB, SDHC, or SDHD gene mutations: a retrospective and prospective analysis. Lancet Oncol 2009;10(8):764-71.