New Prognostic Indicators in Pediatric Adrenal Tumors Neuroblastoma and Adrenal Cortical Tumors, Can We Predict When These Will Behave Badly?
Jason A. Jarzembowski, MD, PHDa,b, *
KEYWORDS
· Neuroblastoma · Ganglioneuroblastoma · Ganglioneuroma · Adrenocortical adenoma
· Adrenocortical carcinoma . MYCN · ALK · TP53
Key points
· Peripheral neuroblastic tumors (pNTs) are categorized according to their stromal content, degree of differentiation, architecture, mitotic-karyorrhectic index, and age via the International Neuroblas- toma Pathology Committee (INPC) classification.
. The prognosis of patients with pNTs depends most heavily on INPC classification, age, International Neuroblastoma Risk Group stage, MYCN amplification status, ploidy, and loss of heterozygosity at 1p and 11q.
· Children with adrenocortical tumors have a better prognosis than adults even when so-called malig- nant pathologic features are present. Tumor size greater than 5 cm and weight greater than 100 g, higher clinical stage and venous invasion, and increased mitotic rate portend a worse prognosis.
· TP53 mutations are common in pediatric adrenocortical tumors and, although not prognostically use- ful, could suggest an underlying tumor predisposition syndrome.
ABSTRACT
P ediatric adrenal tumors are unique entities with specific diagnostic, prognostic, and therapeutic challenges. The adrenal medulla gives rise to peripheral neuroblastic tumors (pNTs), pathologically defined by their architec- ture, stromal content, degree of differentiation, and mitotic-karyorrhectic index. Successful risk stratification of pNTs uses patient age, stage, tu- mor histology, and molecular/genetic aberrations. The adrenal cortex gives rise to adrenocortical tu- mors (ACTs), which present diagnostic and prog- nostic challenges. Histologic features that signify
poor prognosis in adults can be meaningless in children, who have superior outcomes. The key clinical, pathologic, and molecular findings of pe- diatric ACTs have yet to be completely identified.
OVERVIEW
Pediatric adrenal tumors represent a spectrum of disease; therefore, it is critical to identify which tu- mors will behave badly and require early aggres- sive treatment, and which will be indolent, with simple resection effecting a cure. The quest for prognostic factors for pediatric adrenal tumors
a Department of Pathology, Medical College of Wisconsin, Milwaukee, WI, USA; b Pathology and Laboratory Medicine, Children’s Wisconsin, Milwaukee, WI, USA
* Department of Pathology, Children’s Wisconsin, MS #701, 9000 West Wisconsin Avenue, Milwaukee, WI 53226.
E-mail address: jjarzemb@mcw.edu
https://doi.org/10.1016/j.path.2020.08.002
has been a long and winding journey through a host of clinical, pathologic, and molecular genetic variables, and although there have been some vic- tories en route, the ultimate destination yet lies beyond the horizon.
NEUROBLASTIC TUMORS
INTRODUCTION
Peripheral neuroblastic tumors (pNTs): neuroblas- tomas, ganglioneuroblastomas, and ganglioneur- omas, are the most common extracranial solid tumors in children, with a frequency of roughly 7 to 10 per million and about 650 new cases in the United States annually.1,2 The relative frequency of these tumors has allowed extensive refinement of the diagnostic and prognostic methods used to guide therapy, and the advancement of this knowl- edge over the past several decades has led to bet- ter outcomes for these patients.3 The heterogeneity of the diverse spectrum of pNTs can be captured by a combination of clinical, path- ologic, and genetic factors, which in turn explains the heterogeneity of their behavior.
CLINICAL PRESENTATION
Most patients with a pNT present before the age of 5 years with a palpable abdominal mass; subse- quent imaging often reveals a retroperitoneal mass with calcifications.4 The anatomic distribu- tion of pNTs reflects their origin from neural crest cells and sympathetic nervous system constitu- ents, within the adrenal medulla and paraverte- brally in the thoracic, abdominal, or cervical regions. Other “classic” (but rare) presentations include opsoclonus-myoclonus, periorbital ecchy- moses, Horner syndrome, and intractable diar- rhea. Several constitutional genotypes confer a predisposition to pNTs, including ALK and PHOX2B mutations, and pNTs may also be asso- ciated with neurofibromatosis, Beckwith- Wiedemann syndrome, Hirschsprung disease, and Turner syndrome.5 Although the risk of tumor development is higher in these conditions, the resulting pNTs vary in terms of prognosis.
PATHOLOGIC DIAGNOSIS
pNTs are composed of both neuroblastic cells and Schwannian stroma, the types and amount of which are key to proper classification.6,7 These groupings, in turn, help predict tumor behavior and patient outcome.
The neuroblastic component of a pNT can show a wide range of differentiation from primitive small round blue cells to mature ganglion cells (Fig. 1).
Undifferentiated/poorly differentiated neuroblasts have small round to oval nuclei with fine, speckled, “salt-and-pepper” chromatin and scant ampho- philic cytoplasm. In contrast, mature ganglion cells have eccentrically located, large, round nuclei with a single large central nucleolus, vesicular chro- matin, and abundant amphophilic to eosinophilic cytoplasm. Neuroblasts are often polygonal and fit together with a “paving stone” appearance. By immunohistochemistry, these cells are positive for NB84, neuron-specific enolase, PGP9.5, PHOX2B, synaptophysin, and tyrosine hydroxy- lase.7,8 Neuroblasts are often embedded in vari- able amounts of fibrillary eosinophilic neuropil and may also form Homer Wright rosettes with central neuropil. The degree of neuroblastic differ- entiation for a pNT is categorized as undifferenti- ated (no neuropil or ganglionic features; diagnosis relies on ancillary studies), poorly differ- entiated (some neuropil and/or ganglionic features present), or differentiating (>5% of cells show ganglionic features), defined by the presence of neuropil and ganglionic features (see Fig. 3).
Schwann cells are long and spindled with small wiry or comma-shaped nuclei and modest amounts of clear to lightly eosinophilic cytoplasm. The cells are immunohistochemically positive for S100 protein. It is important to correctly distin- guish between Schwannian stroma and neuropil. Per the International Neuroblastoma Pathology Committee (INPC) classification, pNTs with less than 50% Schwannian stroma are called neuro- blastomas (see Fig. 1A-C), and those with more than 50% stroma are deemed ganglioneuroblas- toma, intermixed (see Fig. 1D), or ganglioneuroma (see Fig. 1E); the exception to this is the ganglio- neuroblastoma, nodular (GNBn), which is defined by its architecture rather than its exact composi- tion (see Fig. 1F; Fig. 2). Ganglioneuroma is distin- guished from ganglioneuroblastoma by its lack of immature neuroblasts and neuropil.
CLINICAL PREDICTORS OF PROGNOSIS
Age
It has long been clear that younger children with pNTs fare better than older ones, and most current treatment protocols take this into account. For example, the INPC classification uses 1-year, 18- month, and 5-year cutoff points in its algorithm, and the current International Neuroblastoma Risk Group (INRG) risk group stratification system uses 12- and 18-month cutoffs.6,9 Age appears to be a continuous variable in terms of its utility as a prognostic factor; although a breakpoint needs to be established somewhere, there is no significant difference in risk for a 546 day old and
ARTICLE IN PRESS
A
B
C
D
E
F
A
B
METRIC 1
2
3
a 548 day old with similar tumors. One study showed an 83.0% 5-year event-free survival for children less than 1 year of age compared with 67.9% for 12 to 18 month olds and 38.3% for chil- dren older than 18 months, and numerous other studies show similar findings. 10-12
The precise reason for the age dependence of prognosis in children with pNTs is unclear but may reflect different underlying biology in tumors of young versus older patients.13,14 The best evi- dence for this is the spontaneous regression of many tumors in infants, a phenomenon that seems to occur less frequently with age. In 1963, Beck- with and Perrin15 observed clustered neuroblasts with mitotic activity and invasive growth patterns within the adrenal medulla of infants at autopsy; these lesions were present at a much higher rate than the incidence of pNTs and were not seen in older children. They postulated that these “neu- roblastomas in situ” underwent involution or delayed maturation in most cases and only rarely progressed to clinically apparent neoplasia. This hypothesis is also consistent with the high detec- tion rate of elevated urinary catecholamines in Japanese infants during national screening pro- grams, most of whom did not actually have evi- dence of a pNT.16 Finally, the spontaneous regression of the tumors in infants with stage MS neuroblastoma would seem to be another example of this biologic behavior. Thus, pNTs in younger children may represent something more akin to delayed maturation/differentiation instead of neoplasia, thus accounting for the superior out- comes in this cohort.
Stage
Since 2009, the International Neuroblastoma Response Group Staging System (INRGSS) has been used for pNTs.9 This system replaced the In- ternational Neuroblastoma Staging System (INSS), which had some shortcomings around its complexity and its postsurgical basis. The INRGSS, on the other hand, is based on pre-surgi- cal/pre-treatment imaging studies, clinical presen- tation, and patient age and is summarized in Table 2 here: https://ascopubs.org/doi/full/10.1200/ JCO.2008.16.6876.
Patients with higher INRGSS stages have worse prognoses. In the initial paper proposing this sys- tem, patients with stage L1 had a 90% 5-year event-free survival compared with 78% for stage L2 patients; overall survival was likewise signifi- cantly higher with lower stage.9 For stage M INRGSS patients, 1 study found an event-free sur- vival rate of 54.8%.17 Stage MS disease is known to have special characteristics, including very young patient age, absence of adverse molecular features, and a high frequency of spontaneous regression.14 One study demonstrated that in simi- larly aged children, stage 4S (MS) patients had markedly better 5-year event-free and overall sur- vival than stage 4 patients (77% and 84% vs 64% and 69%).18
The utility of stage as a prognostic factor em- phasizes the importance of correct and thorough pathologic evaluation of bone marrow speci- mens when assessing for stage M disease. The INRG and others have published recommended
approaches to this workup, which include ensuring adequate lengths of trabecular bone in the biopsy and adequate cellularity in the aspi- rate, using at least 2 different antibodies for immunohistochemical detection, reporting the area or percentage of metastatic involvement, and using molecular methods to detect minimal residual disease. 19-24
PATHOLOGIC PREDICTORS OF PROGNOSIS Gross Examination
Many of the important gross findings in other can- cers, tumor size, resection margins, and vascular invasion, do not affect staging and are of little prognostic significance in pNTs.7 The primary objective in gross evaluation is detection of well- demarcated nodules within the tumor parenchyma that define a GNBn.7,25 These nodules may appear hemorrhagic or simply darker than the surrounding tumor and are usually well demarcated but unen- capsulated (see Fig. 2). GNBn may have a worse prognosis than a homogenous ganglioneuroblas- toma, intermixed, depending on the composition of the tumor nodule (as described in later discus- sion). Needle core biopsies of a GNBn may be deceiving, because either the ganglioneuromatous component or the neuroblastic component, or both may be sampled, thus affecting prognostication.26,27
Histologic Examination
pNTs are classified according to their architecture, stromal content, and degree of differentiation of the tumor cells as described above, but several additional factors also contribute to the assign- ment of “favorable” or “unfavorable” histology, age and mitotic-karyorrhectic index (MKI).6
MKI is determined by counting the number of mitotic or apoptotic figures seen in 5000 tumor cells. In theory, the MKI reflects the proliferative and apoptotic rates of the tumor; both are usually higher in aggressive neoplasms than in indolent ones. Determination of an accurate MKI is depen- dent on several factors.6,7 First, the MKI should be representative across the entire tumor, so multiple representative fields (not “hot spots”) must be counted on each slide. Second, areas of necrosis should be avoided. Third, a cell should only be considered apoptotic if it shows nuclear pykno- sis/karyorrhexis; cells with eosinophilic cytoplasm alone do not qualify.
pNTs are categorized as low MKI (<100/ 5000 cells or <2%), intermediate MKI (100-200/ 5000 cells or 2%-4%), or high MKI (>200/ 5000 cells or >4%). Attempts to circumvent the 5000-cell denominator have demonstrated
variable success; there is no perfect substitute for performing a true 5000-cell count. This require- ment does, however, limit one’s ability to obtain a reliable MKI on a small biopsy or a focus of bone marrow involvement.
This histologic information, amount of Schwan- nian stroma, degree of differentiation, MKI, along with the patient’s age, is the basis of INPC classifi- cation of pNTs (Fig. 3).6,25 The INPC classification is by itself a remarkably powerful predictor of prog- nosis. In the initial paper defining the system, pa- tients with favorable histology tumors had 85% overall survival, whereas those with unfavorable histology tumors had only 40% overall survival.6 In a more recent study validating the age cutoffs, for patients greater than 18 months old, the event-free and overall survival rates for favorable histology tumors were 90.6% and 95.0%, and for unfavorable histology tumors were 31.7% and 38.4%.28 There are concerns about the confound- ing presence of age twice within the clinical risk- stratification algorithm (once as part of the INPC system and once as a clinical variable), which have recently generated a newly proposed 4-cate- gory system based on age, degree of differentia- tion, and MKI.29 Although this approach is more statistically sound, it may oversimplify pathologic classification, and more data and experience will be needed to determine its utility.
MOLECULAR/GENETIC PREDICTORS OF PROGNOSIS
MYCN
The most significant molecular factor in pNT prog- nosis is MYCN, a protooncogene and transcription factor that drives cellular proliferation.3º MYCN is amplified in about 20% of pNTs, as defined by a 4-fold increase in MYCN signal over the centro- meric probe on a fluorescence in situ hybridization assay; this can occur via linear amplification within homogenously staining regions or as extrachro- mosomal double minutes.31 MYCN amplification correlates strongly with undifferentiated or poorly differentiated histology as well as with high MKI, and nuclear hypertrophy or “bull’s-eye” nucleoli.32 As such, MYCN-amplified tumors usually have un- favorable histology and a poor prognosis; 1 recent study showed that amplification was associated with a 19.6-fold higher risk for children under 18 months and a 3-fold higher risk for older chil- dren.33 However, a small subset of MYCN-ampli- fied tumors does not express the N-myc protein and has favorable histologic features and a good outcome; hence, morphology can trump genetics in these cases. 34
ARTICLE IN PRESS
Undifferentiated
Unfavorable histology
High MKI
Unfavorable histology
Poorly differentiated
Unfavorable histology
Age >= 18 months
Low or intermediate MKI
Age < 18 months
Favorable histology
Stroma <50%, no nodules
High MKI
Unfavorable histology
Age >= 18 months
Unfavorable histology
Differentiating
Intermediate MKI
Age < 18 months
Favorable histology
Age >= 5 years
Unfavorable histology
Low MKI
Age < 5 years
Favorable histology
Stroma >=50%, no nodules
Immature neuroblasts or neuropil present
Ganglioneuroblastoma, intermixed
Favorable histology
Ganglion cells, no neuropil
Ganglioneuroma
Favorable histology
Nodules present
Classify according to least favorable nodule
Varies
MYCC
Another subset of neuroblastomas has unfavor- able histology with nucleolar hypertrophy, undif- ferentiated neuroblasts, and high MKI, but does not show MYCN amplification or express N-myc protein.35 Instead, these tumors often express C- myc protein and have poor prognosis, similar to MYCN-amplified ones. Immunohistochemistry for C-myc is currently the best way to identify this subset of cases.
ALK
Approximately 20% of pNTs have amplification of ALK, a tyrosine kinase receptor gene, and another 5% to 10% have activating point mutations within
the kinase domain.36 Dysregulated ALK activity leads to increased cell proliferation and portends a worse prognosis.37 ALK acts synergistically with MYCN, such that upregulation of both is frequently seen in high-risk neuroblastomas. Iden- tification of ALK aberrations is important for prog- nostication as well as therapy; tumors with constitutive ALK activity may respond to specific tyrosine kinase inhibitors. 38
Telomeric Maintenance
Telomeres, the repetitive DNA sequences capping the ends of chromosomes, cannot be completely replicated in each mitotic cycle in normal cells; they continually shorten and eventually drive a cell into senescence. Some pNTs upregulate the
Prognostic Indicators in Pediatric Adrenal Tumors
alternate lengthening of telomeres (ALT) pathway, which uses homologous recombination to over- come this imposed mortality.39 Rearrangements of telomerase reverse transcriptase (TERT), the catalytic subunit of the telomerase complex, occur in 20% to 30% of high-risk neuroblastomas and are associated with aggressive tumor behavior.40 Alpha thalassemia/mental retardation syndrome X-linked (ATRX), which also participates in ALT, is preferentially overexpressed in pNTs of older children and is associated with a poor prognosis. 41
Structural Chromosomal Alterations
Several large-scale chromosomal aberrations have prognostic value in pNTs. Diploid or near-diploid neuroblastomas have a worse prognosis than hyperdiploid ones, and this is incorporated into cur- rent risk-stratification systems.42 However, many ganglioneuroblastomas and nearly all ganglioneuro- mas have diploid DNA content; this does not change their usual favorable prognosis.43 Segmental chro- mosomal gains and losses are common in pNTs, with deletion of 1p, deletion of 11q, and gain of 17q associated with aggressive behavior.42
Other Molecular/Genetic Alterations
The neurotrophin receptors (TrkA, B, and C [NTRK1-3]) are tyrosine kinases involved in pNT growth and differentiation.44 TrkA is highly expressed in favorable tumors; it binds nerve growth factor and stimulates neuroblast differenti- ation. TrkC binds neurotrophin-3 and its expres- sion mirrors TrkA. TrkB is overexpressed in MYCN-amplified tumors; its ligand, brain-derived neurotrophic factor, is produced by the tumor cells and drives an autocrine loop of proliferation.
One study showed that upregulation of ARID1A and ARID1B, which are involved in chromatin remodeling, was associated with a poor prog- nosis. 45 Recent attention has focused on the development of gene expression signatures that could be useful for identification of minimal resid- ual disease as well as for identifying biologic risk.46,47 Some proposed candidates have been shown to be independent predictors of survival.
SUMMARY
Pathologic evaluation plays a major role in the diagnosis, prognosis, and treatment selection for patients with pNTs. This information is then com- bined with the results of molecular testing to form the foundation for risk stratification. The 2 major algorithms include similar factors: Children’s Oncology Group uses age, INSS stage, histology, MYCN status, ploidy, and LOH at 1p and 11q to assign patients to one of 3 groups; the INRG
uses INRGSS stage, histology, MYCN status, ploidy, and loss of heterozygosity at 11q and pla- ces patients in one of 4 categories (Please see figure 2 here https://ascopubs.org/doi/10.1200/ JCO.2008.16.6785).48,49 The risk grouping helps establish a prognosis and treatment options for the patient. Although great strides have been made in the understanding of pNTs, a constant search is on for better prognostic factors that will help identify those patients in need of novel or more aggressive therapies.
CLINICS CARE POINTS
· Peripheral neuroblastic tumors are diagnosed according to their stromal content, degree of differentiation, and architecture.
. The International Neuroblastoma Pathology Committee classification assigns peripheral neuroblastic tumors favorable or unfavorable histology based on diagnosis, mitotic- karyorrhectic index, and age.
. Overall, patient prognosis depends most heavily on INPC classification, age, stage, MYCN amplification status, ploidy, and loss of heterozygosity at 1p and 11q.
· Other genetic aberrations, such as ALK muta- tions/amplification, TERT, ATRX, and TRK, may also be useful in prognostication.
ADRENOCORTICAL TUMORS
INTRODUCTION
Adrenocortical tumors (ACTs) are rare in children (0.1-0.4 per million depending on age) with a fe- male preponderance of 2- to 4-fold.50,51 Most pe- diatric ACTs are hormonally active (unlike their adult counterparts), stemming from the function of their cells of origin in the zona fasciculata (gluco- corticoids), zona reticularis (androgens), and the provisional/fetal cortex (dehydroepiandrosterone). Many of the genes that drive adrenal development are, unsurprisingly, dysregulated in pediatric ACTs (discussed later). Importantly, children with ACTs have markedly better prognoses than adults with the same clinicopathologic features. Based on common morphologic classification schemes for adult ACTs, the vast majority (usually >90%) of pe- diatric ACTs would be called adrenocortical carci- nomas (ACCs), but this is clearly incorrect based on the high rates of event-free and overall survival in these patients.52-57 These findings have led to speculation that pediatric ACTs arise from the developing adrenal gland, whereas adult ACTs may arise from mature cortical cells.54,57 This the- ory in turn raises the possibility that spontaneous
regression and/or maturation may account for the better clinical outcome of ACTs in children. How- ever, because adult ACTs represent most of these tumors, most studies have focused on older pa- tients. Thus, much about pediatric ACTs remains poorly understood, especially the distinction be- tween adrenocortical adenomas (ACAs) and ACCs, and prognostic factors.
CLINICAL PRESENTATION
Children with ACTs primarily exhibit virilization with or without Cushing syndrome; a minority have feminization or isolated Cushing syndrome.53,58 Aldosterone-secreting tumors (Conn syndrome) are rare in children, although hypertension can be a common symptom of an ACT from either aldosterone or cortisol production.58 Occasionally, patients have nonsecreting tumors and instead present with abdominal pain, fever, or a palpable mass. True “incidentalomas” are less common in children than adults perhaps because they are less likely to undergo imaging for other reasons.
PATHOLOGIC DIAGNOSIS
ACTs have similar gross and microscopic appear- ances in children and adults. These tumors are usually unilateral and encapsulated (or at least well circumscribed), and their relationship to the native adrenal gland is evident. Pediatric ACTs can range from 1 to 20 cm and 10 to 2500 g and are usually in the yellow-tan-brown color spec- trum, but rarely may be pigmented (“black” ade- nomas) because of abundant lipofuscin within the cells.53,54 Cystic change may be seen, and frank necrosis may raise suspicion of malignancy (see later discussion).
Histologically, the tumor cells typically resemble the cortical layer from which they arose and are characterized by cytologically bland round to polygonal cells with clear or lightly eosinophilic, vacuolated cytoplasm and bland nuclei with vesic- ular chromatin and single small nucleoli (Fig. 4). A wide variety of growth patterns have been described, including diffuse/solid, alveolar, tubular, fibrohyaline, and yolk sac-like. Mitotic ac- tivity can vary considerably and, along with cytoa- typia, may be prognostic factors (discussed later). Fibrosis, necrosis, hemorrhage, and calcification can all be seen to varying degrees. By immunohis- tochemistry, the tumor cells are usually positive for vimentin, inhibin, melan A, synaptophysin, and cal- retinin and variably reactive for cytokeratins; they are negative for S100 and chromogranin.59,60 Several distinct ACT variants have been described, including oncocytic, myxoid,
sarcomatoid, and pediatric; the significance of the first 3 variants is unknown.54,61
CLINICAL PREDICTORS OF PROGNOSIS
Age
As described above, with ACTs have a substan- tially better prognosis that adults with similar stage, histology, and other features. In addition, multiple studies have shown that younger children with ACTs had longer overall survival than older ones.57,62-64 Although the precise age cutoff var- ied by study, most were between 3 and 7 years old. This variation has been interpreted by some investigators as further evidence that “pediatric” and “adult” ACTs are biologically distinct tumors with different origins and behaviors.
Stage
Multiple groups have established that stage is an important prognostic factor for pediatric ACTs. In 1 study, outcome correlated with tumor stage us- ing a novel system (stage 1: complete excision/ tumor <200 cm3; stage 2: microscopic residual disease/tumor >200 cm3/abnormal hormone levels postoperatively; stage 3: gross residual tu- mor; stage 4: metastatic disease). Overall survival was greater than 90% for children with stage 1 ACTs, and almost 0% for those with stage 4 ACTs; children with stage 2 and 3 tumors had var- iable prognosis.58 A subsequent series using the current American Joint Committee on Cancer (AJCC) staging system (T1: tumor <5 cm without local invasion; T2: tumor >5 cm without local inva- sion; T3: local invasion; T4: involvement of adja- cent organs) similarly showed survival of all the T1 patients and none of the T4 patients. 65,66 Although the AJCC staging system potentially confounds tumor size and spread, both have been shown to be independent prognostic factors via multivariate analyses. Recurrence is also a negative prognostic factor and is associated with low overall survival despite additional surgeries to reestablish local control.58
PATHOLOGIC PREDICTORS OF PROGNOSIS Gross Examination
Most analyses of pediatric ACTs have identified size as a clear prognostic factor, and many groups have included it in their staging systems as described above.53,65,67,68 The definition of favor- able versus unfavorable size varies between studies: although 5 cm was a typical linear break- point, the weight cutoff used varied from 50 g to 500 g. Although increased size generally portends a worse prognosis, individual patients and tumors
ARTICLE IN PRESS
A
B
C
D
E
F
may display unexpected behavior. In the AFIP study, although their proposed cutoff of 400 g was statistically significant, there was a malignant 24-g tumor and a benign one that weighed more than 2 kg; therefore, size must be considered in the context of other clinicopathologic factors.53
Histologic Examination
One of the best known forays into morphologic prognostication of ACTs was a study of 43 adult ACTs by Weiss52 that considered 9 features: nu- clear grade (using the Fuhrman system for renal cell carcinoma), mitotic rate (per 50 high- powered fields), atypical mitoses, character of cytoplasm (degree of vacuolization or clearing), architecture of tumor cells, necrosis, invasion of venous structures, invasion of sinusoidal struc- tures, and invasion of tumor capsule. He found that the 3 most useful factors in designating an ACT as an ACC were high mitotic rate (>5 per 50 high-powered fields), the presence of atypical mitoses, and venous invasion. No single criterion could dichotomize these ACTs into benign and malignant, and he proposed a scoring system based on specific values for each of the 9 morphologic features (Table 1). A subsequent paper using some of the same ACC cases demonstrated that mitotic rate was the key
prognostic factor for distinguishing high- and low-risk ACCs; this paper was limited to ACTs called carcinomas by the above criteria, and only included adult patients.69 Thus, although it seemed like prognosis could be somewhat pre- dicted by morphology, the applicability to pediat- ric tumors was unknown.
A decade later, Wieneke and colleagues53 analyzed 83 pediatric ACTs in an attempt to address this gap. They applied the Weiss criteria for adult ACTs to their pediatric group and although many of the features were associated with poor prognosis in univariate analysis, multi- variate analysis revealed that only 3 predictors of malignant behavior remained: invasion of the vena cava, tumor necrosis, and high mitotic activ- ity. There was substantial overlap in scoring be- tween patients with benign-behaving tumors (0 and 7 adverse histologic features) and those with metastatic disease (1 and 9 features). The dilemma was succinctly stated by the investigators: “Only 31% of histologically malignant tumors behaved in a clinically malignant fashion.”53 The investiga- tors proposed a 3-tier classification with 0 to 2 points = benign, 3 points = intermediate risk, and 4 points or more = malignant. Another group subsequently applied the Wieneke criteria to a cohort of 13 pediatric ACTs and found that all 7
| Pathologic features predictive of malignant behavior in adrenocortical tumors Table 1 | ||||
|---|---|---|---|---|
| Hough et al,55 1979 | Weiss, 52 1984 | Wieneke et al,53 2003 | van Slooten et al,56 1985 | |
| Group studied | Adult & pediatric | Adult | Pediatric | Adult |
| Size | >10 cm | >10.5 cm | ||
| Mass | >100 g | >250 g | >400 g | >150 g |
| Vascular invasion | X ☒ | X ☒ | X ☒ | X ☒ |
| Capsular invasion | ☒ X | X ☒ | ||
| Nuclear atypia | X ☒ | X ☒ | X ☒ | |
| Mitoses | >1 per 10 hpf | >5 per 50 hpf or atypical mitoses | >15 per 20 hpf or atypical mitoses | >2 per 10 hpf |
| Necrosis | ☒ X | ☒ X | X ☒ | Plus other "regressive" changes |
| Fibrous bands | ☒ X | |||
| Diffuse growth pattern | ☒ | ☒ | ☒ X | |
| Other | Soft tissue invasion | |||
patients with tumors categorized as benign or in- termediate had excellent long-term survival, whereas 4 of 6 patients with tumors deemed ma- lignant behaved as such.70 A later paper from Das and colleagues71 showed similar findings in their pediatric ACT cohort. Thus, although the au- thors might be able to identify the clearly benign ACTs based on morphology, some morphologi- cally malignant tumors have pleasantly unex- pected benign clinical courses.
Immunohistochemical and Special Staining
Ki67
Ki67 (MIB-1), an indicator of cell proliferation, has been investigated as a prognostic marker in ACTs in part because of the utility of mitotic counts in the previously described morphologic classifica- tions for adult and pediatric ACTs (Fig. 5). Immuno- histochemical staining for Ki67 appears useful in predicting the behavior of adult and pediatric ACTs, with cutoff values of 5% to 15% depending on the study.63,72 One group found that the 3-year event-free survival for patients with Ki-67 index ≥15% was 48.5% compared with 96.2% for pa- tients with a Ki-67 index less than 15%.63
Other markers
Several studies have assessed the utility of reticulin staining, in addition to the Weiss score, for distin- guishing between ACA and ACC.73,74 Although nearly all ACC but only rare ACA had disruption of the reticulin network, this method has not been vali- dated for pediatric ACTs. High numbers of tumor- infiltrating CD8+ cytotoxic T lymphocytes correlated with better prognosis in 1 study of pediatric ACC.75 Although decreased expression of class II HLA
antigens has been shown to be a negative prog- nostic factor in adult ACTs, there are conflicting re- ports of its expression in pediatric tumors.76,77 These and other potential biomarkers will require further investigation before their true prognostic util- ity is known.
MOLECULAR/GENETIC PREDICTORS OF PROGNOSIS
TP53
The p.R337H mutation, a specific TP53 germline mutation, underlies a Li-Fraumeni-like cancer pre- disposition syndrome in which affected children have an increased risk of breast, brain, and soft tissue cancers in addition to ACT, which is often the first manifestation.78 A Children’s Oncology Group study found TP53 mutations in 50% of pe- diatric ACTs, some of which were p.R337H cases, and only 5% of which were canonical hot-spot mu- tations.79 Unfortunately, TP53 status does not appear to correlate with behavior in pediatric ACTs, despite being a proven negative prognostic factor in adult ACTs.80-82 One possible exception is the coexistence of TP53 and ATRX mutations, which appeared in 1 study to be associated with a dismal prognosis.83
IGF-II and IGF-IR
ACTs occur at greater frequency in patients with Beckwith-Wiedemann syndrome, which is associ- ated with epigenetic alterations at the 11p15 lo- cus. One of the key regulatory genes present in this region and regulated at least in part by imprinting is insulin-like growth factor II (IGF-II), which is expressed from the paternal allele. 84
A
B
&
IGF-II signals through its receptor, IGF-IR, to stim- ulate proliferation through the mitogen-activated protein kinase pathway and to inhibit apoptosis through the phosphoinositol-3-kinase pathway. Most pediatric ACTs (60%-100% of cases, depending on the study) have loss of heterozygos- ity at 11p15 and associated elevated IGF-II expression, as much as 18-fold higher than normal adrenal glands. 85-87 Furthermore, 1 study of pedi- atric and adult ACTs showed increased methyl- ation of IGF-II regulatory regions in ACC, suggesting upregulated transcription.88 Neither LOH at 11p15 nor IGF-II overexpression portends poor prognosis in pediatric ACTs, even though the latter is known to be associated with malignant behavior in adult ACTs.72,85,89,90 However, IGF-IR messenger RNA was 3- to -fold higher in pediatric ACC than ACA, and IGF-IR levels were an inde- pendent predictor of metastases.91,92 IGF-I-R levels were similar between adult ACC and ACA.
SF-1
Multiple genomic studies of pediatric ACTs have shown gains of all or part of chromosome 9q, especially 9q34.93,94 Steroidogenic factor 1 (SF- 1) is located near this region (9q33.3) and plays an essential role in adrenal gland development and maintenance. Knockout mice lacking SF-1 have adrenal hypoplasia, and SF-1 is required for contralateral compensatory adrenal growth following unilateral adrenalectomy.95 SF-1 gene amplification and nuclear expression were found in a large fraction of pediatric ACTs (47% and 56%, respectively) but in only a minority of adult ACTs (10% and 29%).96 SF-1 expression appeared to correlate with steroid production by the tumor.97 However, SF-1 levels were similar be- tween pediatric ACAs and ACCs; nonetheless, 1 study showed that expression correlated with prognosis in adult ACCs.96-98
Genomic Alterations
West and colleagues86 found increased /GF-II and decreased KCNQ1 and CDKN1C expression in pediatric ACTs compared with normal adrenal cor- tex. All 3 genes map to the 11p15 locus, with /GF-II normally expressed from the paternal allele and KCNQ1 and CDKN1C both expressed from the maternal allele; this suggests that imprinting and methylation may be important in tumorigenesis and echoes the IGF-II results described above. They also found decreased expression of class II HLA genes in ACCs versus ACAs, although as mentioned previously, this result has not panned out at the protein expression level in all studies.77
SNP analysis of pediatric ACTs (not differenti- ating between benign and malignant) showed frequent chromosomal abnormalities, including loss of 4q34, gain of 9q33-q34 and 19p, and loss of heterozygosity of chromosome 17 and 11p15; the involvement of 11p15 corresponds well with the above findings.99 ACTs with TP53 aberrations tended to have more chromosomal gains and los- ses than those with wild-type TP53. Another group performed network analysis on preexisting gene expression data, comparing pediatric ACCs to normal adrenal glands, and found upregulation of IGF-II as well as 4 novel hubs: CDK1, CCNB1, CDC20, and BUB1B.100 As noted in later discus- sion, although expression of ß -catenin itself has not been shown to be a prognostic factor, other members of the CCNB1 network may be so.
Aneuploidy was a common finding in pediatric ACTs but did not discriminate between ACA and ACCs.101 Loss of heterozygosity for 9p21 that was also associated with loss of p16 expression was seen in adult ACC, but not ACA. 102
Other Markers
Early data for some other proposed markers of malignancy and/or poor prognosis have recently been published. In 1 study of pediatric ACTs, increased cytoplasmic membrane expression of GLUT-1, a key component of the glycolytic pathway, correlated with a greater Weiss score as well as shorter disease-free and overall sur- vival.103 High expression of YAP-1, a member of the Wnt/B-catenin signaling pathway, was associ- ated with poor outcome, despite the fact that ß- catenin expression itself did not correlate with prognosis.104,105 SHH was upregulated in adult ACCs but downregulated in pediatric ones, but neither finding correlated with outcome.106 The microRNA biogenesis pathway was investigated as a potential marker for ACT behavior, but 2 groups found drastically different relationships be- tween DICER1 and TARBP2 levels and prognosis in adult ACTs; furthermore, adrenal tumors are not a consistent finding in the DICER1-pleuropul- monary blastoma familial tumor predisposition syndrome. 107-109 ACTs seen in the setting of McCune-Albright syndrome are usually associ- ated with GNAS-activating mutations, but most of these patients have hyperplasia and not adenomas. 110
Most of these studies were limited by, unsurpris- ingly, low case numbers and lack of a true defini- tion of malignancy (most used Wieneke criteria), but nonetheless represent the forefront of efforts to find new ways to classify and treat pediatric ACTs.
Prognostic Indicators in Pediatric Adrenal Tumors
SUMMARY
Can it be predicted when pediatric ACTs will behave badly? Not for certain, not for everyone, not yet. Some criteria, patient age, tumor size, tu- mor stage, and mitotic rate, can assist patholo- gists and their clinical colleagues in distinguishing obviously benign from obviously malignant cases. However, it is the borderline cases and the outlier cases that have historically posed the most trouble for physicians and patients alike. Better prognostic factors must be identified to ensure that patients are managed appropriately. The best opportunity for this lies within the realm of genetics, but much more work lies ahead to reach this important goal.
CLINICS CARE POINTS
· Children with adrenocortical tumors have a better prognosis than adults even when malig- nant pathologic features are present, and younger children fare better than older ones.
· Gross pathologic features, such as size greater than 5 cm and weight greater than 100 g, as well as higher clinical stage and venous invasion, portend a worse prognosis.
· Histopathologic features are generally inade- quate at predicting behavior, although mitotic rate (or Ki67 immunohistochemical staining) and microscopic evidence of invasive growth suggest malignancy.
· TP53 mutations are common in pediatric ACTs and, although not prognostically useful, could suggest an underlying tumor predispo- sition syndrome.
DISCLOSURE
None.
REFERENCES
1. Siegel DA, King J, Tai E, et al. Cancer incidence rates and trends among children and adolescents in the United States, 2001-2009. Pediatrics 2014; 134(4):e945-55.
2. Ward E, DeSantis C, Robbins A, et al. Childhood and adolescent cancer statistics, 2014. CA Cancer J Clin 2014;64(2):83-103.
3. Pinto NR, Applebaum MA, Volchenboum SL, et al. Advances in risk classification and treatment strate- gies for neuroblastoma. J Clin Oncol 2015;33(27): 3008-17.
4. Sharma R, Mer J, Lion A, et al. Clinical presenta- tion, evaluation, and management of neuroblas- toma. Pediatr Rev 2018;39(4):194-203.
5. Barr EK, Applebaum MA. Genetic predisposition to neuroblastoma. Children (Basel) 2018;5(9):119.
6. Shimada H, Ambros IM, Dehner LP, et al. The inter- national neuroblastoma pathology classification (the Shimada system). Cancer 1999;86(2):364-72.
7. Jarzembowski JA, Rudzinski E, Shimada H. Col- lege of American Pathologists: protocol for the ex- amination of specimens from patients with neuroblastoma. 2014.
8. Picarsic J, Reyes-Múgica M. Phenotype and immu- nophenotype of the most common pediatric tu- mors. Appl Immunohistochem Mol Morphol 2015; 23(5):313-26.
9. Monclair T, Brodeur GM, Ambros PF, et al, INRG Task Force. The International Neuroblastoma Risk Group (INRG) staging system: an INRG Task Force report. J Clin Oncol 2009;27(2):298-303.
10. Moroz V, Machin D, Faldum A, et al. Changes over three decades in outcome and the prognostic influ- ence of age-at-diagnosis in young patients with neuroblastoma: a report from the International Neu- roblastoma Risk Group Project. Eur J Cancer 2011; 47(4):561-71.
11. London WB, Castleberry RP, Matthay KK, et al. Ev- idence for an age cutoff greater than 365 days for neuroblastoma risk group stratification in the Chil- dren’s Oncology Group. J Clin Oncol 2005;23(27): 6459-65.
12. Schmidt ML, Lal A, Seeger RC, et al. Favorable prognosis for patients 12 to 18 months of age with stage 4 nonamplified MYCN neuroblastoma: a Children’s Cancer Group Study. J Clin Oncol 2005;23(27):6474-80.
13. Brodeur GM, Bagatell R. Mechanisms of neuro- blastoma regression. Nat Rev Clin Oncol 2014; 11(12):704-13.
14. Brodeur GM. Spontaneous regression of neuro- blastoma. Cell Tissue Res 2018;372(2):277-86.
15. Beckwith JB, Perrin EV. In situ neuroblastomas: a contribution to the natural history of neural crest tu- mors. Am J Pathol 1963;43(6): 1089-104.
16. Sawada T. Past and future of neuroblastoma screening in Japan. Am J Pediatr Hematol Oncol 1992;14(4):320-6.
17. Naranjo A, Irwin MS, Hogarty MD, et al. Statistical framework in support of a revised Children’s Oncology Group neuroblastoma risk classification system. JCO Clin Cancer Inform 2018;2:1-15.
18. Taggart DR, London WB, Schmidt ML, et al. Prognostic value of the stage 4S metastatic pattern and tumor biology in patients with meta- static neuroblastoma diagnosed between birth and 18 months of age. J Clin Oncol 2011; 29(33):4358-64.
19. Burchill SA, Beiske K, Shimada H, et al. Recom- mendations for the standardization of bone marrow disease assessment and reporting in children with
Jarzembowski
neuroblastoma on behalf of the International Neuro- blastoma Response Criteria Bone Marrow Working Group. Cancer 2017;123(7):1095-105.
20. Beiske K, Burchill SA, Cheung IY, et al. Consensus criteria for sensitive detection of minimal neuroblas- toma cells in bone marrow, blood and stem cell preparations by immunocytology and QRT-PCR: recommendations by the International Neuroblas- toma Risk Group Task Force. Br J Cancer 2009; 100(10):1627-37.
21. Parsons LN, Gheorghe G, Yan K, et al. Improving detection of metastatic neuroblastoma in bone marrow core biopsies: a proposed immunohisto- chemical approach. Pediatr Dev Pathol 2016; 19(3):230-6.
22. Parsons LN, Gheorghe G, Yan K, et al. An evidence-based recommendation for a standard- ized approach to detecting metastatic neuroblas- toma in staging bone marrow biopsies. Pediatr Dev Pathol 2017;20(1):38-43.
23. Hata JL, Correa H, Krishnan C, et al. Diagnostic utility of PHOX2B in primary and treated neuro- blastoma and in neuroblastoma metastatic to the bone marrow. Arch Pathol Lab Med 2015;139(4): 543-6.
24. Nagai J, Kigasawa H, Tomioka K, et al. Immunocy- tochemical detection of bone marrow-invasive neu- roblastoma cells. Eur J Haematol 1994;53(2):74-7.
25. Peuchmaur M, d’Amore ES, Joshi VV, et al. Revi- sion of the International Neuroblastoma Pathology Classification: confirmation of favorable and unfa- vorable prognostic subsets in ganglioneuroblas- toma, nodular. Cancer 2003;98(10):2274-81.
26. Hassan SF, Mathur S, Magliaro TJ, et al. Needle core vs open biopsy for diagnosis of intermediate- and high-risk neuroblastoma in children. J Pediatr Surg 2012;47(6):1261-6.
27. Deeney S, Stewart C, Treece AL, et al. Diagnostic utility of core needle biopsy versus open wedge bi- opsy for pediatric intraabdominal solid tumors: re- sults of a prospective clinical study. J Pediatr Surg 2017;52(12):2042-6.
28. Sano H, Bonadio J, Gerbing RB, et al. International neuroblastoma pathology classification adds inde- pendent prognostic information beyond the prog- nostic contribution of age. Eur J Cancer 2006; 42(8):1113-9.
29. Sokol E, Desai AV, Applebaum MA, et al. Age, diag- nostic category, tumor grade, and mitosis- karyorrhexis index are independently prognostic in neuroblastoma: an INRG project. J Clin Oncol 2020;38(17):1906-18.
30. Beltran H. The N-myc oncogene: maximizing its tar- gets, regulation, and therapeutic potential. Mol Cancer Res 2014;12(6):815-22.
31. Shiloh Y, Shipley J, Brodeur GM, et al. Differential amplification, assembly, and relocation of multiple
DNA sequences in human neuroblastomas and neuroblastoma cell lines. Proc Natl Acad Sci U S A 1985;82(11):3761-5.
32. Tornóczky T, Semjén D, Shimada H, et al. Pathology of peripheral neuroblastic tumors: significance of prominent nucleoli in undifferentiated/poorly differ- entiated neuroblastoma. Pathol Oncol Res 2007; 13(4):269-75.
33. Campbell K, Shyr D, Bagatell R, et al. Comprehen- sive evaluation of context dependence of the prog- nostic impact of MYCN amplification in neuroblastoma: a report from the International Neu- roblastoma Risk Group (INRG) project. Pediatr Blood Cancer 2019;66(8):e27819.
34. Suganuma R, Wang LL, Sano H, et al. Peripheral neu- roblastic tumors with genotype-phenotype discor- dance: a report from the Children’s Oncology Group and the International Neuroblastoma Pathology Com- mittee. Pediatr Blood Cancer 2013;60(3):363-70.
35. Wang LL, Teshiba R, Ikegaki N, et al. Augmented expression of MYC and/or MYCN protein defines highly aggressive MYC-driven neuroblastoma: a Children’s Oncology Group study. Br J Cancer 2015;113(1):57-63.
36. Mossé YP, Laudenslager M, Longo L, et al. Identifi- cation of ALK as a major familial neuroblastoma predisposition gene. Nature 2008;455(7215): 930-5.
37. Bresler SC, Weiser DA, Huwe PJ, et al. ALK muta- tions confer differential oncogenic activation and sensitivity to ALK inhibition therapy in neuroblas- toma. Cancer Cell 2014;26(5):682-94.
38. Mossé YP, Lim MS, Voss SD, et al. Safety and activ- ity of crizotinib for paediatric patients with refrac- tory solid tumours or anaplastic large-cell lymphoma: a Children’s Oncology Group phase 1 consortium study. Lancet Oncol 2013;14(6): 472-80.
39. Koneru B, Lopez G, Farooqi A, et al. Telomere maintenance mechanisms define clinical outcome in high-risk neuroblastoma. Cancer Res 2020; 80(12):2663-75.
40. Valentijn LJ, Koster J, Zwijnenburg DA, et al. TERT rearrangements are frequent in neuroblastoma and identify aggressive tumors. Nat Genet 2015;47(12): 1411-4.
41. Kurihara S, Hiyama E, Onitake Y, et al. Clinical fea- tures of ATRX or DAXX mutated neuroblastoma. J Pediatr Surg 2014;49(12):1835-8.
42. Ambros PF, Ambros IM, Brodeur GM, et al. Interna- tional consensus for neuroblastoma molecular di- agnostics: report from the International Neuroblastoma Risk Group (INRG) Biology Com- mittee. Br J Cancer 2009; 100(9): 1471-82.
43. Okamatsu C, London WB, Naranjo A, et al. Clinico- pathological characteristics of ganglioneuroma and ganglioneuroblastoma: a report from the
Prognostic Indicators in Pediatric Adrenal Tumors
CCG and COG. Pediatr Blood Cancer 2009;53(4): 563-9.
44. Brodeur GM, Minturn JE, Ho R, et al. Trk receptor expression and inhibition in neuroblastomas. Clin Cancer Res 2009; 15(10):3244-50.
45. Sausen M, Leary RJ, Jones S, et al. Integrated genomic analyses identify ARID1A and ARID1B al- terations in the childhood cancer neuroblastoma. Nat Genet 2013;45(1):12-7.
46. Vermeulen J, De Preter K, Naranjo A, et al. Predict- ing outcomes for children with neuroblastoma us- ing a multigene-expression signature: a retrospective SIOPEN/COG/GPOH study. Lancet Oncol 2009;10(7):663-71.
47. Oberthuer A, Hero B, Berthold F, et al. Prognostic impact of gene expression-based classification for neuroblastoma. J Clin Oncol 2010;28(21):3506-15.
48. Tolbert VP, Matthay KK. Neuroblastoma: clinical and biological approach to risk stratification and treatment. Cell Tissue Res 2018;372(2):195-209.
49. Cohn SL, Pearson AD, London WB, et al. The Inter- national Neuroblastoma Risk Group (INRG) classi- fication system: an INRG Task Force report. J Clin Oncol 2009;27(2):289-97.
50. Altekruse SF, Kosary CL, Krapcho M, et al, editors. SEER cancer statistics review, 1975-2007. Be- thesda (MD): National Cancer Institute; 2010. Avail- able at: https://seer.cancer.gov/csr/1975_2007/, based on November 2009 SEER data submission, posted to the SEER web site.
51. Hsing AW, Nam JM, Co Chien HT, et al. Risk factors for adrenal cancer: an exploratory study. Int J Can- cer 1996;65(4):432-6.
52. Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocor- tical tumors. Am J Surg Pathol 1984;8(3):163-9.
53. Wieneke JA, Thompson LD, Heffess CS. Adrenal cortical neoplasms in the pediatric population: a clinicopathologic and immunophenotypic analysis of 83 patients. Am J Surg Pathol 2003;27(7):867-81.
54. Dehner LP, Hill DA. Adrenal cortical neoplasms in children: why so many carcinomas and yet so many survivors? Pediatr Dev Pathol 2009;12(4): 284-91.
55. 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;72:390-9.
56. van Slooten H, Schaberg A, Smeenk D, et al. Morphologic characteristics of benign and malig- nant adrenocortical tumors. Cancer 1985;55: 766-73.
57. Michalkiewicz E, Sandrini R, Figueiredo B, et al. Clinical and outcome characteristics of children with adrenocortical tumors: a report from the Inter- national Pediatric Adrenocortical Tumor Registry. J Clin Oncol 2004;22(5):838-45.
58. Sandrini R, Ribeiro RC, DeLacerda L. Childhood adrenocortical tumors. J Clin Endocrinol Metab 1997;82(7):2027-31.
59. Erickson LA. Challenges in surgical pathology of adrenocortical tumours. Histopathology 2018; 72(1):82-96.
60. Mete O, Asa SL, Giordano TJ, et al. Immunohisto- chemical biomarkers of adrenal cortical neo- plasms. Endocr Pathol 2018;29(2):137-49.
61. Sung TY, Choi YM, Kim WG, et al. Myxoid and sar- comatoid variants of adrenocortical carcinoma: analysis of rare variants in single tertiary care cen- ter. J Korean Med Sci 2017;32(5):764-71.
62. Gulack BC, Rialon KL, Englum BR, et al. Factors associated with survival in pediatric adrenocortical carcinoma: an analysis of the National Cancer Data Base (NCDB). J Pediatr Surg 2016;51(1):172-7.
63. Pinto EM, Rodriguez-Galindo C, Pounds SB, et al. Identification of clinical and biologic correlates associated with outcome in children with adreno- cortical tumors without germline TP53 mutations: a St Jude Adrenocortical Tumor Registry and Chil- dren’s Oncology Group study. J Clin Oncol 2017; 35(35):3956-63.
64. Sabbaga CC, Avilla SG, Schulz C, et al. Adrenocor- tical carcinoma in children: clinical aspects and prognosis. J Pediatr Surg 1993;28(6):841-3.
65. Amin MB, Greene FL, Edge SB, et al. editors. American Joint Committee on Cancer. Adrenal Cortical. In: AJCC cancer staging manual. 8th edi- tion. New York: Springer; 2017. p. 911-8.
66. Tucci S Jr, Martins AC, Suaid HJ, et al. The impact of tumor stage on prognosis in children with adre- nocortical carcinoma. J Urol 2005;174(6):2338-42.
67. Bugg MF, Ribeiro RC, Roberson PK, et al. Correla- tion of pathologic features with clinical outcome in pediatric adrenocortical neoplasia. A study of a Brazilian population. Brazilian Group for Treatment of Childhood Adrenocortical Tumors. Am J Clin Pathol 1994;101(5):625-9.
68. Cagle PT, Hough AJ, Pysher TJ, et al. Comparison of adrenal cortical tumors in children and adults. Cancer 1986;57(11):2235-7.
69. Weiss LM, Medeiros LJ, Vickery ALJ. Pathologic features of prognostic significance in adrenocor- tical carcinoma. Am J Surg Pathol 1989; 13:202-6.
70. Chatterjee G, DasGupta S, Mukherjee G, et al. Use- fulness of Wieneke criteria in assessing morpho- logic characteristics of adrenocortical tumors in children. Pediatr Surg Int 2015;31(6):563-571.
71. Das S, Sengupta M, Islam N, et al. Wieneke criteria, Ki-67 index and p53 status to study pediatric adre- nocortical tumors: is there a correlation? J Pediatr Surg 2016;51(11):1795-800.
72. Soon PS, Gill AJ, Benn DE, et al. Microarray gene expression and immunohistochemistry analyses of adrenocortical tumors identify IGF2 and Ki-67 as
Jarzembowski
useful in differentiating carcinomas from ade- nomas. Endocr Relat Cancer 2009; 16(2):573-83.
73. Duregon E, Fassina A, Volante M, et al. The reticulin algorithm for adrenocortical tumor diagnosis: a multicentric validation study on 245 unpublished cases. Am J Surg Pathol 2013;37(9): 1433-40.
74. 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. Histopatholo- gy 2009;55(5):535-43.
75. Parise IZS, Parise GA, Noronha L, et al. The prog- nostic role of CD8+ T lymphocytes in childhood adrenocortical carcinomas compared to Ki-67, PD-1, PD-L1, and the weiss score. Cancers (Basel) 2019;11(11):1730.
76. Leite FA, Lira RC, Fedatto PF, et al. Low expression of HLA-DRA, HLA-DPA1, and HLA-DPB1 is associ- ated with poor prognosis in pediatric adrenocor- tical tumors (ACT). Pediatr Blood Cancer 2014; 61(11):1940-8.
77. Magro G, Esposito G, Cecchetto G, et al. Pediatric adrenocortical tumors: morphological diagnostic criteria and immunohistochemical expression of matrix metalloproteinase type 2 and human leucocyte-associated antigen (HLA) class II anti- gens. Results from the Italian Pediatric Rare Tumor (TREP) Study project. Hum Pathol 2012;43(1):31-9.
78. Achatz MI, Olivier M, Le Calvez F, et al. The TP53 mutation, R337H, is associated with Li-Fraumeni and Li-Fraumeni-like syndromes in Brazilian fam- ilies. Cancer Lett 2007;245(1-2):96-102.
79. Wasserman JD, Novokmet A, Eichler-Jonsson C, et al. Prevalence and functional consequence of TP53 mutations in pediatric adrenocortical carci- noma: a Children’s Oncology Group study. J Clin Oncol 2015;33(6):602-9.
80. Latronico AC, Pinto EM, Domenice S, et al. An in- herited mutation outside the highly conserved DNA-binding domain of the p53 tumor suppressor protein in children and adults with sporadic adre- nocortical tumors. J Clin Endocrinol Metab 2001; 86(10):4970-3.
81. Ragazzon B, Libé R, Gaujoux S, et al. Transcriptome analysis reveals that p53 and {beta)-catenin alter- ations occur in a group of aggressive adrenocortical cancers. Cancer Res 2010;70(21):8276-81.
82. Faria AM, Almeida MQ. Differences in the molecu- lar mechanisms of adrenocortical tumorigenesis between children and adults. Mol Cell Endocrinol 2012;351(1):52-7.
83. Pinto EM, Chen X, Easton J, et al. Genomic land- scape of paediatric adrenocortical tumours. Nat Commun 2015;6:6302.
84. Ribeiro TC, Latronico AC. Insulin-like growth factor system on adrenocortical tumorigenesis. Mol Cell Endocrinol 2012;351(1):96-100.
85. Wilkin F, Gagné N, Paquette J, et al. Pediatric adrenocortical tumors: molecular events leading to insulin-like growth factor II gene overexpres- sion. J Clin Endocrinol Metab 2000;85(5): 2048-56.
86. West AN, Neale GA, Pounds S, et al. Gene expres- sion profiling of childhood adrenocortical tumors. Cancer Res 2007;67(2):600-8.
87. Peixoto Lira RC, Fedatto PF, Marco Antonio DS, et al. IGF2 and IGF1R in pediatric adrenocortical tumors: roles in metastasis and steroidogenesis. Endocr Relat Cancer 2016;23(6):481-93.
88. Creemers SG, van Koetsveld PM, van Kemenade FJ, et al. Methylation of IGF2 regulatory regions to diagnose adrenocortical carcinomas. Endocr Relat Cancer 2016;23(9):727-37.
89. Gicquel C, Bertagna X, Schneid H, et al. Rear- rangements at the 11p15 locus and overexpression of insulin-like growth factor-II gene in sporadic adrenocortical tumors. J Clin Endocrinol Metab 1994;78(6):1444-53.
90. Rosati R, Cerrato F, Doghman M, et al. High fre- quency of loss of heterozygosity at 11p15 and IGF2 overexpression are not related to clinical outcome in childhood adrenocortical tumors posi- tive for the R337H TP53 mutation. Cancer Genet Cytogenet 2008; 186(1):19-24.
91. Almeida MQ, Fragoso MC, Lotfi CF, et al. Expres- sion of insulin-like growth factor-II and its receptor in pediatric and adult adrenocortical tumors. J Clin Endocrinol Metab 2008;93(9):3524-31.
92. Doghman M, El Wakil A, Cardinaud B, et al. Regu- lation of insulin-like growth factor-mammalian target of rapamycin signaling by microRNA in childhood adrenocortical tumors. Cancer Res 2010;70(11): 4666-75.
93. Figueiredo BC, Stratakis CA, Sandrini R, et al. Comparative genomic hybridization analysis of adrenocortical tumors of childhood. J Clin Endocri- nol Metab 1999;84(3):1116-21.
94. James LA, Kelsey AM, Birch JM, et al. Highly consistent genetic alterations in childhood adreno- cortical tumours detected by comparative genomic hybridization. Br J Cancer 1999;81(2):300-4.
95. Wong M, Ikeda Y, Luo X, et al. Steroidogenic factor 1 plays multiple roles in endocrine development and function. Recent Prog Horm Res 1997;52: 167-84.
96. Almeida MQ, Soares IC, Ribeiro TC, et al. Steroido- genic factor 1 overexpression and gene amplifica- tion are more frequent in adrenocortical tumors from children than from adults. J Clin Endocrinol Metab 2010;95(3):1458-62.
97. 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.
Prognostic Indicators in Pediatric Adrenal Tumors
98. Pianovski MA, Cavalli LR, Figueiredo BC, et al. SF- 1 overexpression in childhood adrenocortical tu- mours. Eur J Cancer 2006;42(8):1040-3.
99. Letouzé E, Rosati R, Komechen H, et al. SNP array profiling of childhood adrenocortical tumors re- veals distinct pathways of tumorigenesis and high- lights candidate driver genes. J Clin Endocrinol Metab 2012;97(7):E1284-93.
100. Kulshrestha A, Suman S, Ranjan R. Network anal- ysis reveals potential markers for pediatric adreno- cortical carcinoma. Onco Targets Ther 2016;9: 4569-81.
101. Zerbini C, Kozakewich HPW, Weinberg DS, et al. Adrenocortical neoplasms in childhood and adoles- cence: analysis of prognostic factors including DNA content. Endocr Pathol 1992;3(3):116-28.
102. Pilon C, Pistorello M, Moscon A, et al. Inactivation of the p16 tumor suppressor gene in adrenocortical tu- mors. J Clin Endocrinol Metab 1999;84(8):2776-9.
103. Pinheiro C, Granja S, Longatto-Filho A, et al. GLUT1 expression in pediatric adrenocortical tu- mors: a promising candidate to predict clinical behavior. Oncotarget 2017;8(38):63835-45.
104. Abduch RH, Carolina Bueno A, Leal LF, et al. Un- raveling the expression of the oncogene YAP1, a Wnt/beta-catenin target, in adrenocortical tumors and its association with poor outcome in pediatric patients. Oncotarget 2016;7(51): 84634-44.
105. 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.
106. Gomes DC, Leal LF, Mermejo LM, et al. Sonic hedgehog signaling is active in human adrenal cor- tex development and deregulated in adrenocor- tical tumors. J Clin Endocrinol Metab 2014;99(7): E1209-16.
107. de Sousa GR, Ribeiro TC, Faria AM, et al. Low DICER1 expression is associated with poor clinical outcome in adrenocortical carcinoma. Oncotarget 2015;6(26):22724-33.
108. Caramuta S, Lee L, Ozata DM, et al. Clinical and functional impact of TARBP2 over-expression in adrenocortical carcinoma. Endocr Relat Cancer 2013;20(4):551-64.
109. Schultz KA, Yang J, Doros L, et al. DICER1-pleu- ropulmonary blastoma familial tumor predisposi- tion syndrome: a unique constellation of neoplastic conditions. Pathol Case Rev 2014; 19(2):90-100.
110. Almeida MQ, Azevedo MF, Xekouki P, et al. Activa- tion of cyclic AMP signaling leads to different pathway alterations in lesions of the adrenal cortex caused by germline PRKAR1A defects versus those due to somatic GNAS mutations. J Clin Endo- crinol Metab 2012;97(4):E687-93.