Home > E. Pathology by systems > Endocrine system > Adrenals > pediatric adrenal cortical tumors

pediatric adrenal cortical tumors

Wednesday 24 November 2004

pediatric adrenal cortical neoplasms, pediatric adrenocortical tumors; pediatric adrenal cortical neoplasms (PACNs); PACTs

Adrenal cortical tumors include the adrenocortical adenoma (AA) and carcinoma (ACC).

Adrenal cortical tumors are rare in the pediatric age group.

Adrenal cortical tumors in children account for 50% to 70% of cases of Cushing syndrome, in contrast to only 20% of cases in adults.

Less often does an adrenal cortical tumor present with feminizing and masculinizing manifestations in children.

Conn syndrome, due to an aldosterone-producing ACN, is rare in childhood. Most tumors in Conn syndrome are benign and characterized by their lipid-rich clear cells.

Adrenal cortical neoplasms in pediatric patients (<20 years) are rare.

The clinical manifestations and biologic behavior of these lesions can be quite distinct from their histologically similar counterparts in the adult population, making pathologic criteria for distinguishing benign from malignant tumors equivocal.

Adrenal cortical neoplasms in the pediatric population are rare, with adrenal cortical carcinomas (ACC) accounting for the majority of the tumors.

It is estimated that approximately 25 cases of pediatric adrenal cortical neoplasms occur in the United States per year, of which approximately 75% are ACC.

Of greatest clinical importance, pediatric adrenal cortical neoplasms appear to behave differently than histologically similar tumors in the adult population. Pathologic criteria for malignancy in the adult population are well established in the literature.

Adrenal cortical tumors in children are both clinically and histologically different from those occurring in adults, even though there is overlap.

Some data indicate that although adrenal cortical neoplasms in childhood may not be as uniformly fatal as they are in adults, even with histologic evidence of malignancy, when they are clinically malignant, they are relatively aggressive, resulting in death within a relatively short period of time (<2 years).

However, 5- and 10-year disease-free survival rates of 63.9% and 59.0%, respectively, suggest that once a patient lives beyond 5 years, prognosis appears to be quite good.

Prognosis

The clinicopathological features of adrenal cortical tumors in children, their behavior and epidemiology contrast nominally with the same neoplasms in adults.

However, one of the consistent observations is that these neoplasms in children, though having several morphologic features associated with adrenal cortical carcinoma in adults, do not have the same unfavorable prognostic implications, especially in children under 6 years of age.

Because these atypical histological features are commonly found in adrenal cortical tumors in children, they are disproportionately interpreted as adrenal cortical carcinomas.

- Age

  • Age is an apparent prognostic factor as these patients were <3 yr old.
  • Because of the discrepancy between pathology and clinical outcome, adrenocortical tumors in this age group should be classified as neoplasms of unknown malignant potential.

The established criteria for distinguishing benign from malignant adrenal cortical neoplasms in the adult population have not been helpful in predicting the biologic behavior of such neoplasms in the pediatric population. (#12826878#)

In the distant past, most children with adrenal cortical tumors had a poor prognosis, perhaps because they would die following surgery of acute adrenal cortical insufficiency. (#12826878#)

Now, with the availability of pharmacologic replacement of cortisol, patients with histologically malignant-appearing tumors do well, often better than their adult counterparts. (#12826878#)

In further support of this claim, 69% patients with histologically malignant-appearing tumors in this clinical series had a benign clinical course without recurrence or death from disease. (#12826878#)

In contrast to the adult population, tumors in children more often are hormonally functional with clinical evidence of virilization, Cushing’s syndrome, mixed endocrine syndromes, and rarely feminization or Conn’s syndrome. (#12826878#)

Further, feminizing tumors in adults are fatal, a fate not shared in pediatric patients (the patient in this clinical series is alive without evidence of disease), although death from tumor in this population has been reported. (#12826878#)

Whereas a functional tumor may be indicative of poor outcome in adult patients, the same cannot be extrapolated to the pediatric population.

Although girls are affected more frequently than boys, gender did not prove to be a statistically significant predictor of an adverse patient outcome. (#12826878#)

- Age

  • There appears to be a biphasic age distribution: an “infantile” group (<5 years) and an “adolescent” group (>10 years). (#12826878#)
  • The patients with a poor clinical outcome tend to fall into the “adolescent” group, although age alone does not seem to independently predict clinical outcome. (#12826878#)

- Tumor weight

  • With increasing tumor size and weight, the likelihood of malignant behavior increases, to the point that some have suggested tumors >500 g are invariably malignant.
  • Although in this series, a tumor weight of >400 g was statistically predictive of poor patient outcome (p = 0.01), it did not prove to be an independent prognostic factor and so cannot be used as the sole criterion for malignancy. (#12826878#)
  • There were a few tumors that weighed >500 g and yet enjoyed a good clinical outcome (the largest tumor in this clinical series [2413 g] occurred in a patient with a clinically benign course), whereas tumors as small as 24 g yielded a malignant clinical outcome. (#12826878#)
  • A proposed cutoff of 400 g is higher than others, suggesting that higher weight is more predictive of patient outcome. (#12826878#)

- Tumor size

  • Furthermore, tumor size (as measured linearly in centimeters), directly related to weight, follows a similar predictive course. (#12826878#)
  • Tumors >10.5 cm had a statistically worse outcome than those of smaller size. (#12826878#)
  • Again, however, size alone was not an independent prognostic factor: a 20-cm tumor (along with 12 other tumors >10.5 cm) occurred in patients who had a good clinical outcome. (#12826878#)

- Extension into periadrenal soft tissues and/or adjacent organs and invasion into the vena cava

  • Features that have been found to be statistically predictive of patient outcome are extension into periadrenal soft tissues and/or adjacent organs and invasion into the vena cava.
  • No pediatric patient studies explicitly examined these features prognostically, although in adults these features place the tumor into an advanced T stage (T3 or T4).
  • In one analysis, these factors appear to be predictive of worse patient outcome (p = 0.01), with the presence of vena cava invasion independently predicting a poor patient outcome (p = 0.01). (#12826878#)
  • A separate staging system for childhood adrenocortical tumors has been proposed that differs from the adult population.
  • An increased stage does correlate with a poor clinical outcome.

There are a selection of histologic features that are associated with a more aggressive biologic behavior (#12826878#):
- capsular invasion,
- vascular invasion (separate from vena cava invasion),
- tumor necrosis,
- increased mitotic activity (>15/20 HPF),
- atypical mitotic figures.

Whereas necrosis and >15 mitotic figures per 20 HPF independently suggest a worse clinical outcome based on multivariate analysis, none of these features can be used solely as a predictor of malignancy: necrosis was present in 59% of the cases in group B, patients who had a good clinical outcome. (#12826878#)

Therefore, it is important to stress that no single histologic feature is diagnostic for malignancy.

Similar to malignancies of other organs, and especially of other endocrine organs, a constellation of histomorphologic features must be taken into consideration before a diagnosis of ACC can be made in pediatric patients.

Epidemiology

The incidence of cases classified as adrenal cortical carcinomas in children in the United States was 0.2 cases: 1,000,000 individuals less than 19 years of age.

ACC accounts for less than 0.5% of all pediatric malignancies but is the third most common carcinoma in children exceeded by thyroid papillary carcinoma and salivary gland carcinoma.

There is a bimodal age distribution of adrenal cortical tumors and they are more common in females in the pediatric population.

In a report of 256 cases from the International Pediatric Adre-nocortical Tumor Registry (IPATR), the male to female ratio was 1:1.6.

Two age distributions were noted by the IPATR, an infantile group with a peak incidence in the first year of life and an adolescent group with a peak inci-dence between 9 and 16 years.

A female predilection with a mean age at diagnosis of 4.6 years and nearly 50% of cases diagnosed in the first 4 years of life has been documented by others.

The experience of Dehner and Hill is similar. In a review of 39 cases, ACN presented in children between 7 days and 12 years of age with a mean age of 3 years and median of 2 years.

Seventy-six percent of children were less than 4 years. The male:female ratio was 1:2.5. Others have reported similar findings.

There are congenital examples of adrenal cortical tumors.

One of the highest incidences of purported adrenal cortical carcinomas in children (4.7 cases:1,000,000) has been identified in southern Brazil where a distinct germline p53 mutation has been found in the population, but whose other features are not those of classic Li-Fraumeni syndrome.

Predisposition

There are several syndromic associations with adrenal cortical tumors including BWS (hemihypertrophy, splanchnomegaly, mac-roglossia, and intraabdominal neoplasms), the Li-Fraumeni syndrome, and Carney complex.

Adrenal hyperplasia and adrenal cortical tumors have been reported in MEN1 syndrome, MAS, and neurofibromatosis 1.

Examples of adrenal cortical tumors have been seen in the setting of CAH.

Bilateral adrenal cortical tumors, often seen in syndromic-associated cases, and ectopic adrenal cortical tumors are uncommon.

One of the highest incidences of purported adrenal cortical carcinomas in children (4.7 cases:1,000,000) has been identified in southern Brazil where a distinct germline p53 mutation has been found in the population, but whose other features are not those of classic Li-Fraumeni syndrome.

Macroscopy

The most important initial step in the pathologic examination of an adrenal cortical tumors in childhood is weighing the tumor, since all other gross and histopathologic attributes of the tumor itself are in a sense secondary, if the tumor is confined to the gland and does not have evidence of metastatic spread to regional lymph nodes or to more distant sites such as the liver and lungs.

In some cases, it may be difficult to judge whether a circumscribed mass in the adrenal is part of multinodular hyperplasia of the cortex or even a PHEO.

Cortical neoplasms, not only in children but in adults, vary in size, weight, coloration, consistency (solid and/or cystic), presence or absence of hemorrhage, and presence or absence of necrosis.

A complete or incomplete fibrous capsule may be apparent at the periphery of the tumor, or the tumor appears to compress the adjacent parenchyma without a capsule.

As noted earlier, the size and weight, especially the latter, are closely correlated with the clinical outcome of an adrenal cortical tumor in a child.

The cut surface in the absence of hemorrhage and necrosis often has a pale to bright yellow to yellow-brown appearance which may or may not be uniform throughout because of cystic changes.

Uncommonly, the tumor may have a brownish-black appearance due to lipofuscin accumulation in the cytoplasm of tumor cells as in the case of the primary pigmented (micronodular) adrenocortical disease (PPAD).

A hemorrhagic mass may be difficult to differentiate from a neuroblastoma.

A number of pathologic studies in the literature have examined every conceivable microscopic feature of ACN for their predictive prognostic value, but most of the studies have consisted principally of tumors in adults where cytologic atypia, mitotic activity, zonal necrosis and transecting fibrous bands have predictive value in terms of outcome when several of these features are present in the adrenal cortical tumor.

Some of these same histological features are found with some frequency in adrenal cortical tumors in children yet lack any significant correlation with prognosis.

Microscopy

The classic pattern of an adrenal cortical tumor is a neoplasm which is composed of clear or pale polygonal cells with abundant lipid-rich cytoplasm or cells with more homogeneous eosinophilic cytoplasm, all arranged in short cords or trabecular profiles.

The nuclei are uniform and centrally positioned in the cell.

The adjacent cortex is often compressed and atrophic.

Other histological patterns include diffuse, formless sheets of relatively monotonous polygonal cells, alveolar pattern of loosely cohesive cells, glandular profiles, a yolk sac tumor-like pattern and delicate ribbons of cells in a hyaline myxoid stroma, either as an exclusive pattern or component of the adrenal cortical tumor.

Micoscopical prognostic parameters

Histologic parameters include:
- capsular invasion
- vascular invasion
- extra-adrenal soft tissue extension
- growth pattern
- cellularity
- tumoral necrosis (confluent necrosis)
- cytoplasmic eosinophilia
- nuclear pleomorphism
- nuclear-to-cytoplasmic ratio
- prominent nucleoli
- frequency of mitoses
- atypical mitotic figures
- bands of fibrosis
- calcifications
- Ki-67 positivity

Features associated with an increased probability of a malignant clinical behavior include:
- tumor weight (>400 g)
- tumor size (>10.5 cm)
- vena cava invasio
- capsular and/or vascular invasion
- extension into periadrenal soft tissue
- confluent necrosis
- severe nuclear atypia
- >15 mitotic figures/20 high power fields
- presence of atypical mitotic figures.

Vena cava invasion, necrosis, and increased mitotic activity (>15 mitotic figures/20 high power fields) independently suggest malignant clinical behavior in multivariate analysis.

- polymorphism, pleomorphism, nuclear atypia

  • If polymorphism is the theme for the various patterns in adrenal cortical tumors in children, then pleomorphism applies to the individual cellular features.
  • In some adrenal cortical tumors, monomorphism is an appropriate characterization in the presence of uniform tumor cells.
  • Especially prominent in cases of adrenal cortical tumors in children who are usually 4 years old or less is pleomorphism in terms of individual cell size as bordering on tumor giant cells with bizarre nuclear configurations and intense hyperchromatism; these latter cell types when present can constitute a minor or major component of a particular tumor.

- mitoses

  • By contrast, some tumors can have substantial mitotic activity yet are small (less 100 g) and confined to the gland.

- necrosis

  • necrosis, either as individual cells, or microfoci or macrofoci of necrosis the may have been appreciated in the gross examination.

- vascular invasion

  • Intratumoral microinvasion of blood vessels

- capsule breaching

  • Apparent microscopic breaching of the capsule are additional histological findings.

- Predisposition

  • BWS-associated ACCs are not aggressive neoplasms and that a congenital ACC is reported to have undergone spontaneous regression.

It would appear from the preceding paragraph that an adrenal cortical tumor with some of these features has attained the threshold for the pathologic diagnosis of ACC and for that reason many studies of adrenal cortical tumors in children are represented by a majority of cases with a diagnosis of ACC. However, the paradox is that the prognosis for ACCs in children, especially those under 5 years of age, are remarkably favorable with a 5-year event-free survival (EFS) of 70% to 80%.

Pediatric adrenocortical carcinoma

Adrenocortical carcinoma indisputably occurs in children and these tumors usually weigh in excess of 400 g and innodes with less common spread to the bone and brain.

Most children with “bona fide” ACCs are usually dead from tumor within 2 years of the diagnosis.

In the purest sense possible, recurrent or metastatic disease (lymph node and/or other organ site) was taken to be indicative of a malignant primary tumor.

Weight/stage risk groups

Three risk groups based on tumor localization and weight have been proposed as an alternative means of predicting the clinical behavior of a particular ACN in a child.

The three risk groups include:
- ACNs weighing less than 200 g and confined to the gland as “low risk” for malignant behavior and these tumors are interpreted as adenomas.
- The most problematic group consists of those adrenal cortical tumors that are confined to the gland but weigh between 200 and 400 g; these neoplasms are designated “atypical” adenomas with uncertain malignant potential.
- ACNs weighing in excess of 400 g with a high risk of malignant behavior and are commonly associated with the various histopathologic features of ACCs in adults.

Some of "atypical" tumors have invaded beyond the adrenal gland into adjacent tissues and/or have major vascular invasion, not simply microinvasion of vessels within the tumor itself; these tumors are clearly behaving in a malignant fashion.

Most of the “atypical” adenomas have a favorable outcome in our experience.

Histological risk groups

There are a selection of histologic features that are associated with a more aggressive biologic behavior (#12826878#):
- capsular invasion
- vascular invasion (separate from vena cava invasion)
- tumor necrosis
- increased mitotic activity (>15/20 HPF)
- presence of atypical mitotic figures.

No one pathologic feature can be used to accurately and independently predict patient outcome.

Several factors have been identified, which when identified in aggregate, appear to accurately predict a more aggressive biologic behavior.

By using a cutoff of four or more features, more aggressive clinical management and follow-up may yield an improvement in the overall patient outcome. (#12826878#)

Whereas necrosis and >15 mitotic figures per 20 HPF independently suggest a worse clinical outcome based on multivariate analysis, none of these features can be used solely as a predictor of malignancy: necrosis was present in 59% of the cases in group "atypical" or "intermediate outcome", patients who had a good clinical outcome.

Therefore, it is important to stress that no single histologic feature is diagnostic for malignancy.

A three-part separation has been proposes (#12826878#):
- up to two criteria, benign long term clinical outcome;
- three criteria: indeterminate for malignancy (intermediate, atypical, uncertain malignant potential);
- four or more criteria, portends a poor clinical outcome.

These breakpoints accurately classify 78% of all cases that behave in a clinically malignant fashion; 58% of all cases that behave in a clinically benign fashion; and 22% of patients are placed in an indeterminate category, of which 4% behave in a clinically malignant fashion. (#12826878#)

It is axiomatic, but perhaps needs emphasis, that the intermediate/indeterminate group of patients will need to have clinical follow-up similar to those that are frank carcinomas, at least until a 5-year disease-free survival period is achieved to exclude a more aggressive biologic behavior. (#12826878#)

Application of these criteria to adrenal cortical neoplasms in children in general practice is encouraged to further substantiate these initial findings. (#12826878#)

Immunochemistry

The immunophenotype of an ACN in a child is identical to its counterpart in the adult as the tumor cells are reactive with vimentin, melan-A, inhibin, and calretinin.

There are no immunophenotypic differences between an adenoma and carcinoma.

Adrenal cortical tumors, typically adenomas, may demonstrate immunoreactivity for cytokeratin; carcinomas are usually nonreactive.

Ploidy

Ploidy analysis has limited value in the discrimination of an adenoma from a carcinoma.

In one study of 50 ACNs in children, 21 of 29 patients (73%) with aneuploid tumors remained disease-free.

Differential diagnosis

- adrenal pheochromocytoma

  • The distinction between a cortical neoplasm and pheochromocytoma is not clear in every case, especially in a tumor with large, bizarre-appearing cells, granular basophilic cytoplasm, and a nested growth pattern.
  • The challenge is further heightened by the fact that the results of pertinent biochemical studies are usually not available to correlate with the pathological findings.
  • The tumor is more likely to be cystic and hemorrhagic, and the tumor cells are devoid of some of the tinctorial attributes that are useful in the differentiation of a cortical from a medullary neoplasm.
  • In these cases, immunohistochemistry is helpful with the differential diagnosis.
  • Synaptophysin and NSE are commonly immunoreactive in both pheochromocytomas and adrenal cortical tumors, whereas chromogranin is non-reactive in adrenocortical tumors, but is consistently expressed in pheochromocytomas and paragangliomas.
  • Cytokeratin is typically not found in either pheochromocytomas or para-gangliomas with rare exceptions, but they are often immu-noreactive for vimentin.
  • S-100 protein and HMB-45 staining is useful in the labeling of the sustentacular cells of pheochromocytomas.
  • It is necessary to acknowledge that the results of bcl-2, cytokeratin, and vimentin expression have not proven to discriminate between a cortical and medullary neoplasm in every case.

Types according to grade

- adrenocortical adenoma (ACAs)
- adrenocortical carcinoma (ACCs)

Types according to secretion

- non-secreting adrenocortical tumor
- androgens-secreting adrenocortical tumor
- aldosterone-secreting adrenocortical tumor
- cortisol-secreting adrenocortical tumor

LOH

- 17q22-24 LOH (PRKAR1A locus) (#14500362#)

Gene mutations

- somatic mutations in the PRKAR1A gene (#14500362#)

References

- Pediatric adrenocortical tumors: morphological diagnostic criteria and immunohistochemical expression of matrix metalloproteinase type 2 and human leucocyte-associated antigen (HLA) class II antigens. Results from the Italian Pediatric Rare Tumor (TREP) Study project. Magro G, Esposito G, Cecchetto G, Dall’Igna P, Marcato R, Gambini C, Boldrini R, Collini P, D’Onofrio V, Salfi N, d’Amore E, Ferrari A, Bisogno G, Alaggio R. Hum Pathol. 2012 Jan;43(1):31-9. PMID: #21820153#

- Adrenal cortical tumors in children: factors associated with poor outcome. Klein JD, Turner CG, Gray FL, Yu DC, Kozakewich HP, Perez-Atayde AR, Voss SD, Zurakowski D, Shamberger RC, Weldon CB. J Pediatr Surg. 2011 Jun;46(6):1201-7. PMID: #21683223#

- Adrenocortical neoplasms in young children: age as a prognostic factor. Ahmed AA. Ann Clin Lab Sci. 2009 Summer;39(3):277-82. PMID: #19667412#

- Adrenal cortical neoplasms in the pediatric population: a clinicopathologic and immunophenotypic analysis of 83 patients. Wieneke JA, Thompson LD, Heffess CS. Am J Surg Pathol. 2003 Jul;27(7):867-81. PMID: #12826878#

- Pediatric adrenocortical neoplasms: on the road to some clarity. Dehner LP. Am J Surg Pathol. 2003 Jul;27(7):1005-7. Review. No abstract available. PMID: #12826894#

Portfolio

  • Adrenocortical carcinoma
  • Aldosterone-secreting adrenocortical tumor
  • Aldosterone-secreting adrenocortical tumor
  • Aldosterone-secreting adrenocortical tumor