Neuroblastoma

ANDREW M. DAVIDOFF and JED G. NUCHTERN

Neuroblastoma is the most common solid extracranial malignancy of childhood and the most common
malignant tumor in infants.1 The overall incidence of neuroblastoma is 1 per 100,000 children in the
United States, thereby accounting for 7–10% of all malignancies diagnosed in patients younger than 15
years of age. Yet neuroblastoma is responsible for approximately 15% of all childhood cancer deaths.
Neuroblastoma is a heterogeneous disease. Tumors can spontaneously regress or mature, or display a
very aggressive, malignant phenotype. Because of these unique characteristics, neuroblastoma has been
of great interest to both clinicians and basic science researchers. Progress in molecular and cellular
biology in the past 40 years has contributed greatly to a better understanding of this disease.
Unfortunately, this progress has not significantly altered the clinical outcome for patients with high-risk
disease. Although the prognosis for these patients has improved somewhat in the past three decades,
the long-term outcome remains very poor.

The etiology of neuroblastoma is currently unknown, and no environmental factors have been
convincingly linked to its development. The disease generally occurs sporadically, but familial
neuroblastoma occurs in about 2% of the cases. The substantial biologic and clinical heterogeneity is
also observed in familial cases. The germline mutation associated with hereditary neuroblastoma has
been identified: activating mutations in the tyrosine kinase domain of the anaplastic lymphoma kinase
(ALK) oncogene on the short arm of chromosome 2 (2p23). These mutations can also be somatically
acquired, with the prevalence of ALK activation in sporadic neuroblastoma being approximately 8%.6
The treatment of neuroblastoma reflects the heterogeneity of the disease, requires an understanding of
the risk factors predictive of disease recurrence, and mandates a multidisciplinary approach. Although
resection may be the only therapy required for patients at low risk for disease recurrence, with some
low-risk tumors now simply being observed without any therapy, including surgical resection, the
surgeon provides but one component of the modern multimodal treatment of children with high-risk
disease. Oncologists, radiation therapists, and bone marrow transplantation (BMT) specialists are among
the other important members of the pediatric oncology team. The therapy for patients with
neuroblastoma, as for children with other malignancies, is generally driven by clinical research
protocols. Many of these protocols are sponsored by the Children’s Oncology Group (COG), the New
Approaches to Neuroblastoma Therapy (NANT) consortium, or the larger individual children’s hospitals.

Pathology

Neuroblastoma is an embryonal tumor of the sympathetic nervous system. These tumors arise during
fetal or early postnatal life from sympathetic cells (sympathogonia) derived from the neural crest.
Therefore, tumors can originate anywhere along the path that neural crest cells migrate, including the
adrenal medulla, paraspinal sympathetic ganglia, and sympathetic paraganglia such as the organ of
Zuckerkandl. The German pathologist Rudolph Virchow is generally credited with being the first to
describe the histologic appearance of what is now known as neuroblastoma in his 1864 article entitled,
“Hyperplasia of the Pineal and Suprarenal Glands.” The first to use the term neuroblastoma was James
Homer Wright in 1910, who described the classic appearance of rosettes of tumor cells around central
neural fibrils and noted the similarity to the morphology of the fetal adrenal gland. He also noted the
association between the common sites of tumor development and the pattern of migration of primitive
neural cells.

As one of the small, round, blue cell tumors of infancy and childhood, neuroblastoma, particularly when
undifferentiated, must be distinguished from other neoplasms in this group (Ewing sarcoma family of
tumors [ESFTs], non-Hodgkin lymphoma, and rhabdomyosarcoma). Neuroblastoma can be distinguished
histologically by the presence of neuritic processes (neuropil) and Homer Wright rosettes (neuroblasts
surrounding eosinophilic neuropil). Scattered ganglion cells or immature chromaffin cells can also be
seen. The histologic appearance of the tumor cells may vary from undifferentiated cells to fully mature
ganglion cells. In addition, neuroblastomas have variable degrees of Schwannian cell stroma, reactive
non-neoplastic tissue recruited by the tumor cells. This stroma is intermixed, to a greater or lesser
degree, as wavy bundles and sheets of spindle cells, and produces antiproliferative and differentiation-
inducing factors that are crucial to neuronal differentiation.

Neuroblastoma is characterized by several unique clinical behaviors, including the secretion of

catecholamine products and the potential to regress or mature, either spontaneously or in response to
treatment. Small nodules of primitive neuroblasts are routinely found in the developing adrenal gland,
even during the early postnatal period. Beckwith and Perrin described microscopic nodules that they
termed neuroblastoma in situ in the adrenals of infants undergoing autopsy following death from non-
malignancy related causes. The incidence of this finding was more than 200 fold greater than the clinical
incidence of neuroblastoma, which suggests that many neuroblastomas likely spontaneously regress or
mature into lesions that never become clinically apparent. The process of involution is well described

Fig. 65.1 Histologic appearance of neuroblastic tumors. (A) An undifferentiated neuroblastoma with high
MKI (10×). A clump of karyorrhectic tumor cells (white arrow) and a tumor cell undergoing mitosis (gray
arrow) are shown in the insert (60×). (B) A differentiating neuroblastoma, with low MKI (10×). A
primitive neuroblast (gray arrow) and a differentiating tumor cell (black arrow), with features of
differentiation in both the nucleus and cytoplasm, are shown in the insert (60×). Abundant neuropil is
also seen. (C) A stroma-rich ganglioneuroblastoma with infrequent neuroblasts intermixed within
abundant Schwannian stroma and ganglion cells (10×). (D) A stroma-rich ganglioneuroma. Ganglion cells
are seen (arrow) (10×). Infiltrating lymphoid cells are also seen, but no neuroblasts are present.
(Courtesy Jesse Jenkins, MD, and Christine Fuller, MD, St. Jude Children’s Research Hospital, Memphis
TN. Reprinted from Davidoff AM. Neuroblastoma. In: Oldham KT, Colombani PM, Foglia RP, et al.,
editors. Principles and Practice of Pediatric Surgery. Philadelphia: Lippincott, Williams & Wilkins; 2005.)
during embryonic life, especially in the developing central and peripheral nervous systems. Although
initially thought to be mediated by the immune system, the process of involution may be the result of
the withdrawal of neurotrophic maintenance factors such as nerve growth factor (NGF). Clinically
apparent neuroblastoma can also regress or spontaneously mature, but the mechanism remains
unknown.

HISTOPATHOLOGIC CLASSIFICATION

In 1984, Shimada et al. first developed an age-linked classification system of neuroblastic tumors based
on tumor morphology in which neuroblastomas were divided into two prognostic subgroups: favorable
histology and unfavorable histology. In 1999, the International Neuroblas-toma Pathology Classification
(INPC) was devised; it was then modified in 2003 and is an adaptation of the original Shimada system.
The INPC is based mainly on morphologic changes associated with the maturational sequence of
neuroblastic tumors. It remains an age-linked classification that depends on the differentiation grade of
the neuroblasts, the cellular turnover index (mitosis-karyorrhexis index [MKI]), and the presence or
absence of Schwannian stroma. The INPC classifies neuroblastic tumors into three morphologic
categories: neuroblastoma, ganglioneuroblastoma, and ganglioneuroma (Fig. 65.1).

Neuroblastomas are, by definition, Schwannian stroma poor (<50% of the tumor tissue) and can be
subtyped as undifferentiated, poorly differentiated, or differentiating. Undifferentiated tumors require
supplemental diagnostic methods such as immunohistochemistry, electron microscopy, or cytogenetics
to make the diagnosis of neuroblastoma. Moreover, neuropil is not present. In poorly differentiated
tumors, <5% of tumor cells have features of differentiation, and neuropil is present. Differentiating
tumors demonstrate >5% of tumor cells differentiating toward ganglion cells. Additional factors that
contribute to the prognostic distinction of stroma-poor neuroblastic tumors (neuroblastoma).

TABLE 65.1 Prognostic Evaluation of Neuroblastic Tumors According to the International Neuroblastoma
Pathology Classification (Shimada System)

International Neuroblastoma Pathology Classification. Histologic Characteristics. Prognostic

Group

MKI: mitosis-karyorrhexis index.

aRare subtype, especially diagnosed in this age group. Further investigation and analysis required.

bPrognostic grouping for these tumor categories is not related to patient age. Adapted from Shimada H,
Ambros IM, Dehner LP, et al. The International AIB Pathology Classification (the Shimada System).
Cancer. 1999;86:364–372.
as favorable or unfavorable subtypes include the MKI, which is defined as the number of tumor cells in
mitosis or karyorrhexis per 5000 neuroblastic cells (i.e., low MKI, L <100 cells; intermediate, 100–200
cells; high, >200 cells) and the patient’s age (<1.5 years, 1.5–5 years, >5 years) (Table 65.1). It has been
hypothesized that neuroblastic cells with maturational potential require a latent period before
demonstrating histologic evidence of differentiation. Therefore, there is a certain allowance for mitotic
and karyorrhectic activities of neuroblastic cells in tumors in infants and younger children.

The importance of this histopathologic classification was confirmed in a large, retrospective analysis
reported by Shimada et al. The INPC classification of tumor histology provided independent prognostic
information in which tumors of favorable histology had a 90.8% probability of 5-year event-free survival
[EFS] compared with 31.2% EFS for tumors of unfavorable histology. More recently, the INPC
classification was been shown to add independent prognostic information beyond the prognostic
contribution of age. Therefore, histopathology remains in the current multifactorial risk stratification for
certain patients with neuroblastoma. As the histopathologic pattern within a tumor can be
heterogeneous, it is recommended that representative sections from at least 1 cm3 of viable, non-
necrotic tissue be analyzed to determine histopathologic classification. The prognostic value of assessing
the histopathology of a neuroblastoma after chemotherapy or radiation therapy has not been validated.

Stroma-rich neuroblastic tumors are classified as either ganglioneuroblastomas or ganglioneuromas.

Ganglioneuroblastomas contain cells that are transitioning toward differentiation but are not completely
differentiated/mature. Also, less than 50% of the total volume is made up of neuroblastic cells.
Ganglioneuroblastomas can be further divided into intermixed and nodular subtypes, depending on the
distribution of the neuroblastic cells. The distinction is important because of the significantly worse
prognosis associated with the latter subtype, in which the neuroblastic clones that comprise grossly
distinct nodules appear to be responsible for the aggressive phenotype for this subtype.
Ganglioneuromas contain either maturing or mature cells and lack any neuroblastomatous component.
Most stroma-rich tumors (ganglioneuroblastoma, intermixed and ganglioneuroma, maturing subtype)
are classified as favorable by the INPC. However, the pathologic/prognostic classification of the
ganglioneuroblastoma, nodular subtype, is based on the morphologic evaluation of the
neuroblastomatous nodule(s), and can, therefore, be unfavorable. Tumors that fit the criteria for
ganglioneuroma, mature subtype, with abundant Schwannian stroma and fully mature ganglion cells, in
the absence of neuroblasts, are considered benign, and are generally not considered for enrollment in
protocols for neuroblastic tumors. Despite this, ganglioneuromas can be quite large and infiltrative, and
attempts at removal can be associated with significant complications. In addition, survival does not seem
to be influenced by extent of resection. Therefore, aggressive attempts at resection of ganglioneuromas
are not recommended.

Molecular Biology
Advances in molecular biology research in the past four decades have resulted in an increased
understanding of the genetic events in the pathogenesis and progression of many human malignancies,
including those in children. Neuroblastoma, in particular, has served as a model for a molecular
approach to treating patients with cancer, highlighting the utility of genetic analysis for diagnosis, risk
stratification, and treatment planning. Segmental chromosome aberrations (SCAs) play a role in
neuroblastomas, particularly those that result in the loss of tumor suppressors, gain of oncogenes or or
gene amplification; as well as activating or inactivating mutations of relevant genes or their regulatory
elements. The end result of alterations in these genetic elements, regardless of their specific
mechanisms, is the disruption of the normal balance between cell proliferation and cell death.

DNA CONTENT

Normal human cells contain two copies of each of 23 chromosomes; thus, a normal diploid cell has 46
chromosomes. The majority (55%) of primary neuroblastomas are triploid or “near-
triploid/hyperdiploid” and contain between 58 and 80 chromosomes; the remainder (45%) are either
near-diploid (35–57 chromosomes) or near-tetraploid (81–103 chromosomes). The DNA index of a
tumor is the ratio of the number of chromosomes present to a diploid number of chromosomes (i.e.,
46). Therefore, diploid cells have a DNA index of 1.0, whereas near-triploid cells have a DNA index
ranging from 1.26 to 1.76. Neuroblastomas that are near-diploid or near-tetraploid usually have
structural genetic abnormalities, most frequently chromosome 1p deletion and MYCN amplification.
Near-triploid or hyperdiploid tumors are characterized by almost three complete haploid sets of
chromosomes with few structural abnormalities. Importantly, patients with near-triploid tumors
typically have favorable clinical and biologic prognostic factors and excellent survival rates, as compared
with those patients who have near-diploid or near-tetraploid tumors. This association is most important
for infants with advanced disease as the prognostic significance of tumor ploidy appears to be lost in
patients older than 2 years. Currently, ploidy affects only the risk group assessment of very limited
subgroups of patients with neuroblastoma.

AMPLIFICATION OF MYCN

Investigation of the molecular biology of neuroblastoma began with the cytogenetic characterization of
tumor-derived cell lines. These studies showed the frequent presence of extrachromosomal double-
minute chromatin bodies (DMs) and chromosomally integrated homogeneously staining regions (HSRs)
characteristic of gene amplification (Fig. 65.2). Since that time, it has been shown that the amplified
region was derived from the distal short arm of chromosome 2 (2p24) and contained the MYCN proto-
oncogene. MYCN encodes a 64-kDa nuclear phosphoprotein that forms a transcriptional complex by
associating with other nuclear proteins expressed in the developing nervous system and other tissues.
Enforced expression of MYCN increases the rates of DNA synthesis and cell proliferation, and shortens
the G1 phase of the cell cycle. MYCN can also function as a classic dominant oncogene that cooperates
with activated ras to transform normal cells. Targeted expression of MYCN in transgenic mice results in
the development of neuroblastomas. This activity is potentiated when combined with mutations in ALK.

Overall, approximately 25% of primary neuroblastomas in children have MYCN amplification, with MYCN
amplification being present in 40% with advanced disease but only 5–10% with low-stage disease. The
copy number, which can range from 5- to 500-fold amplification, is usually consistent among primary
and metastatic sites and at different times during tumor evolution and treatment. This finding suggests
that MYCN amplification is an early event in the pathogenesis of neuroblastoma. Amplification of MYCN
is associated with advanced stages of disease, rapid tumor progression, and poor outcome; therefore, it
is a powerful prognostic indicator of tumor behavior. Amplification is most commonly assessed by
fluorescence in situ hybridization (FISH), and current therapeutic neuroblastoma protocols have
incorporated the presence or absence of MYCN amplification into their risk stratification schema.

Fig. 65.2 FISH analysis of a neuroblastoma. (A) Chromosomes in metaphase. The bright spots are double-
minute chromatin bodies. (B) The metaphase chromosomes are again seen. An intact interphase nucleus
is marked with an asterisk. The normal two copies of the MYCN gene are marked with solid arrows.
Homogeneously staining regions (HSRs) are also seen. One is seen in the interphase nucleus, and the
other is marked with a dotted arrow. (Courtesy Marc Valentine, St. Jude Children’s Research Hospital,
Memphis, TN.)

SEGMENTAL CHROMOSOME ABERRATIONS

Also noted on early karyotype analyses of neuroblastoma-derived cell lines were frequent deletions of
the short arm of chromosome 1. Deletions of genetic material in tumors suggest the presence (and
subsequent loss) of a tumor suppressor gene. Approximately 20–35% of primary neuroblastomas exhibit
1p deletion, as determined by FISH, with the smallest common region of loss located within region 1p36.
About 70% of advanced-stage neuroblastomas have 1p deletions. Molecular studies have shown that
there is a strong correlation between 1p deletion and MYCN amplification and other high-risk features
such as age older than 1 year and advanced-stage disease. One study demonstrated that 1p deletions
are independently associated with a worse outcome in patients with neuroblastoma.

Deletion of the long arm of chromosome 11 (11q) also appears to be common in neuroblastoma, being
present in about 40% of cases. Unbalanced deletion of 11q (loss with either retention or gain of 11p
material) is inversely related to MYCN amplification,33,34 yet is strongly associated with other high-risk
features. Attiyeh et al., on behalf of the COG, showed in a large cohort of patients that unbalanced
deletion of 11q and 1p36 was independently associated with a worse outcome in patients with
neuroblastoma. Therefore, the duration of treatment for children with intermediate-risk neuroblastoma
in a recent COG study was based, in part, on the 1p and 11q allelic status of the tumor.
MUTATIONS

Proto-oncogene activation can also occur by point mutation, as occurs with the tyrosine kinase receptor,
ALK, the gene for which resides on the short arm of chromosome 2 (2p23). Receptor tyrosine kinases
(RTKs) are high-affinity cell surface receptors for many growth factors, cytokines, and hormones. When
activated through ligand binding, these proteins mediate phosphorylation of tyrosine on target
molecules or substrates, resulting in intracellular signaling and, ultimately, the regulation of normal
cellular processes. Mutation of RTKs can lead to constitutive activation of the signaling pathway in the
absence of ligand. Activating mutations of ALK have been shown to be the germline abnormality
associated with hereditary neuroblastoma. These mutations can also be somatically acquired, as can
amplification of the gene. Activated ALK has proven to be a targetable abnormality in neuroblastoma,
with drugs such as crizotinib, an anti-ALK antibody, showing efficacy.

Further studies have identified loss-of-function mutations in the homeobox gene PHOXB2 on 4p13 that
are also associated with familial neuroblastoma, particularly when occurring together with Hirschsprung
disease and/or central hypoventilation.

More recently, inactivating mutations of ATRX, a transcriptional regulator that is part of a multiprotein
complex that plays a role in regulating chromatin remodeling, nucleosome assembly, and telomere
maintenance, have been found in neuroblastoma, particularly high-stage tumors in older patients. ATRX
mutations appear to be loss-of-function mutations associated with an absence of the ATRX protein in
the nucleus and with long telomeres. How these alterations lead to lengthened telomeres is uncertain.
These results may provide a molecular marker and potential therapeutic target for neuroblastoma
among adolescents and young adults. It may also delineate the subset of children with neuroblastoma
who have a chronic but progressive clinical course when receiving standard therapeutic approaches and
who may benefit from a different treatment strategy.

Clinical Presentation

Patients with neuroblastoma usually present with signs and symptoms that reflect the primary site and
extent of disease, although localized disease is often asymptomatic. As 75% of neuroblastoma occurs in
the abdominal cavity, an abdominal mass detected on physical examination is a common clinical feature,
as is the complaint of abdominal pain. Other primary sites of neuroblastoma include the posterior
mediastinum (20%), the cervical region (1%), and the pelvis (4%) (organ of Zuckerkandel) (Fig. 65.3).
Respiratory distress or dysphagia may be a reflection of a thoracic tumor. Altered defecation or
urination can be caused by mechanical compression from a pelvic tumor or by spinal cord compression
from a paraspinal tumor. Spinal cord compression may also manifest as an altered gait. A tumor in the
neck or upper thorax can produce Horner syndrome (ptosis, miosis, and anhydrosis), enophthalmos, and
heterochromia of the iris. Acute cerebellar ataxia has also been observed, characterized by the dancing-
eye syndrome, which includes opsoclonus, myoclonus, and chaotic nystagmus. Two-thirds of these cases
occur in infants with mediastinal primary tumors. Additional signs and symptoms that reflect excessive
catecholamine or vasoactive intestinal polypeptide (VIP) secretion include diarrhea, weight loss, and
hypertension.

More than 40% of patients have metastatic disease at diagnosis. These patients are often quite ill and
have systemic symptoms caused by widespread disease. Neuroblastoma in older patients has a pattern
of metastatic disease in which metastases to the bone marrow, lymph nodes, and bone predominate.
The frequency of involvement of distant sites is shown in Table 65.2. These metastases may manifest as
bone pain from cortical metastases or anemia from marrow infiltration. The brain, spinal cord, heart,
and lungs are rare sites of metastases, except with end-stage disease. Metastatic disease also may be
associated with darkened areas around the eyes, referred to as “raccoon eyes,” as a result of
retroorbital venous plexus spread (Fig. 65.4). This is an ominous physical sign, as is the presence of a
limp in children without a history of head or extremity trauma. The diagnosis of neuroblastoma is
generally made by histopathologic evaluation of the primary or metastatic tumor tissue, or by the
demonstration of tumor cells in the bone marrow together with elevated levels of urinary
catecholamines.

LABORATORY FINDINGS

Lactate Dehydrogenase

Despite its lack of specificity, serum lactate dehydrogenase (LDH) can have great prognostic significance.
High serum levels of LDH reflect high proliferative activity or a large tumor burden, and an LDH level
higher than 1500 IU/L appears to be associated with a poor prognosis. Thus, LDH can be used to monitor
disease activity or the response to therapy.

Ferritin

High levels of serum ferritin (>150 ng/mL) may also reflect a large tumor burden or rapid tumor
progression. Elevated serum ferritin is often seen in advanced-stage neuroblastomas and indicates a
poor prognosis. Levels often return to normal during clinical remission.

Fig. 65.3 Primary sites for neuroblastoma are depicted in this anatomic drawing. (Reprinted from
Davidoff AM. Neuroblastoma. In: Oldham KT, Colombani PM, Foglia RP, et al., editors. Principles and
Practice of Pediatric Surgery. Philadelphia: Lippincott Williams & Wilkins; 2005.)

Table 65.2 Sites of Metastases at Diagnosis for Patients With Evans Stage 4S and Stage 4
Adapted from Dubois SG, Kalika Y, Lukens JN, et al. Metastatic sites in stage IV and IVS neuroblastoma
correlate with age, tumor biology, and survival. Pediatr Hematol Oncol. 1999;21:181–189.

Catecholamine Metabolites

Neuroblastoma is characterized by the relatively unique capacity for secretion of catecholamine

products, the metabolites of which can be detected in the urine of more than 90% of patients with
neuroblastoma. Thus, a urine specimen is of clinical value in diagnosing neuroblastoma and determining
the response to therapy. Documentation of elevated urinary catecholamines is required if the diagnosis
of neuroblastoma is being made solely by the identification of neuroblasts in the bone marrow. Urinary
levels of these two catabolites can also be used as markers of tumor progression or relapse and serve as
a surrogate prognostic indicator. Random urine samples are preferable to 24-hour urine estimations for
younger children.

Fig. 65.4 Clinical evidence of metastatic neuroblastoma. “Raccoon eyes,” characteristic of metastatic
neuroblastoma in the posterior orbital venous plexus, are seen in a child with stage 4 disease.

DIAGNOSTIC IMAGING

Standard Radiographs

Chest radiography can be a useful tool for demonstrating the presence of a posterior mediastinal mass,
which in a child is usually a thoracic neuroblastoma. A Pediatric Oncology Group (POG) study
demonstrated that a mediastinal mass was discovered on incidental chest radiographs in almost half of
patients with thoracic neuroblastoma who had symptoms seemingly unrelated to their tumors.
Abdominal radiography is less often the modality by which a neuroblastoma is discovered. However, as
many as half of abdominal neuroblastomas are detectable as a mass with fine calcification.

Fig. 65.5 These MR images highlight several characteristics of high-risk neuroblastoma. (A) Bone
metastasis in femur (arrow). (B) Bone marrow metastases in the vertebral bodies. (C) Intraspinal tumor
extension (dotted arrow). Note displacement of spinal cord (solid arrow) from a large tumor (asterisk).
(D) Encasement of major intra-abdominal vessels (arrow points to the aorta and left renal artery).
Ultrasonography

Although ultrasonography (US) is the modality most often used during the initial assessment of a
suspected abdominal mass, its sensitivity and accuracy are less than that of computed tomography (CT)
or magnetic resonance imaging (MRI) for demonstrating a neuoblastoma. These latter modalities are
generally used after screening with US to assist in generating a differential diagnosis and for further
anatomic definition once the presence of a mass has been confirmed.

Computed Tomography

CT can demonstrate calcification in almost 85% of neuroblastomas, and intraspinal extension of the
tumor can be determined on contrast-enhanced CT. Overall, contrast-enhanced CT has been reported to
be 82% accurate in defining neuroblastoma extent, with the accuracy increasing to nearly 97% when
performed with a bone scan. Although some consider CT to have been supplanted by MRI, others still
consider it to be the image modality of choice for patients with neuroblastoma, especially when used in
conjunction with bone scintigraphy.

Magnetic Resonance Imaging

MRI is becoming the most useful and most sensitive imaging modality for the diagnosis and staging of
neuroblastoma. MRI appears to be more accurate than CT for detection of stage 4 disease. The
sensitivity of MRI is 83%, and that of CT is 43%, and the specificity of MRI is 97%, and that of CT is 88%.
Metastases to the bone and bone marrow, in particular, are better detected by MRI, as is intraspinal
tumor extension (Fig. 65.5). When considering skeletal metastases alone, MRI and bone scan have been
shown to be equivalent. Encasement of major vessels is better defined by MRI than CT, especially with
MR angiography (see Fig. 65.5D). MRI in the coronal plane is suitable for routine assessment of the
whole body from the neck to the pelvis. Evaluating the utility of whole-body MRI, perhaps performed in
conjunction with a functional imaging study such as positron-emission tomography (PET), is being
considered for future clinical staging studies. CT and MRI are not very accurate for staging localized
disease. However, the sensitivity of T1- and T2-weighted MRIs is 100% for detecting neuroblastomas in
infants identified by mass screening.

Fig. 65.6 Imaging of neuroblastoma with MIBG scintigraphy. (A) Scan obtained at presentation of a
patient with metastatic neuroblastoma. There is diffusely abnormal activity throughout much of the
skeleton including the proximal right humerus, both proximal and distal femurs, and the proximal right
tibia. There is also a focus of activity in the right upper retroperitoneum at the site of the primary tumor.
(B) Scan obtained of the same patient after completion of therapy shows no scintigraphic evidence of
MIBG-avid neuroblastoma.

Metaiodobenzylguanidine Imaging

Metaiodobenzylguanidine (MIBG) is transported to and stored in the chromaffin cells in the same way as
norepinephrine. The MIBG scintiscan is the preferred imaging study for evaluating the bone and bone
marrow involvement by neuroblastoma (Fig. 65.6), having largely replaced technetium-99m methylene
diphosphonate (99mTc-MDP) bone scans, which are generally inferior to MIBG in detecting skeletal or
extraskeletal involvement. In addition, monitoring MDP-avid neuroblastomas by bone scintigraphy often
results in false-positive imaging for months after tumor remission. Thus, 99mTc-MDP bone scanning is a
second choice if MIBG imaging is not available or does not visualize known disease. Iodine-131 (131I) or
iodine-123 (123I) can be used to label MIBG. 123I-MIBG supplies a reduced absorbed radiation dose and
superior spatial resolution. The reported sensitivity of MIBG in the detection of neuroblastomas with
metastases to the bone and bone marrow is 82%, and the specificity is 91%. Primary tumors and lymph
node metastases are also detectable. MIBG can demonstrate more sites of tumor involvement in bone
and bone marrow than either bone scintigraphy or standard radiography. However, false-negative MIBG
scans have been seen in patients in which the bone scintigraphy was positive.

BONE MARROW EXAMINATION

Marrow biopsy is a routine method for detecting bone marrow involvement. Both aspiration and
trephine biopsy should be performed, although the latter has better diagnostic value. To collect more
accurate information, taking specimens from multiple sites is recommended. Immunohistochemical
staining with antibodies such as anti-ganglioside GD2, S-100, neuron specific enolase (NSE), and ferritin
is also useful to help reduce the number of false-negative cases. Because biopsy is invasive and painful,
noninvasive alternatives are being evaluated. Studies have suggested the superiority of MRI and MIBG
scintigraphy over bone marrow biopsy in detecting bone marrow infiltration by neuroblastoma.
However, the specificity of these modalities requires further evaluation.

DIFFERENTIAL DIAGNOSIS

Making a correct diagnosis of neuroblastoma can be difficult, because patients present with such diverse
symptoms. For example, acute cerebellar ataxia with opsoclonusmyoclonus can be mistaken for a
primary neurologic disease. Widespread bone involvement may resemble non-neoplastic bone disease
such as osteomyelitis or rheumatoid arthritis, or be associated with systemic inflammatory changes.
Symptoms referable to VIP secretion such as diarrhea can be misinterpreted as symptoms of an enteric
infection or inflammatory bowel disease. Histologically, undifferentiated, small, round blue cell
neuroblastomas may be hard to distinguish from rhabdomyosarcoma, primitive neuroectodermal
tumors, ESFT, or non-Hodgkin lymphoma. Use of a panel of specific antibodies can facilitate histologic
differentiation.

TUMOR STAGING

International criteria for a common neuroblastoma staging system were first described in 1988 and
subsequently revised in 1993. The International Neuroblastoma Staging System (INSS) is a
surgicopathologic staging system that depends on the completeness of resection of the primary tumor,
assessment of ipsilateral and contralateral lymph nodes, and the relation of the primary tumor to the
midline (Table 65.3). Evaluation of the primary tumor and involvement of metastatic sites in the INSS
system depends largely on imaging studies (CT or MRI), although involvement of the bone marrow
continues to be an important component. MIBG scanning is also recommended as part of the initial
evaluation of new patients and, subsequently, for monitoring tumor response to therapy.

Risk-Directed Therapy Based on International Neuroblastoma Staging System

STAGING

As previously mentioned, one of the notable characteristics of neuroblastoma is the substantial

heterogeneity of the disease, which ranges from spontaneous regression or maturation, even without
therapy, to a highly malignant, aggressive phenotype that is poorly responsive to current intensive,
multimodal therapy. Increasing evidence indicates that the biologic and molecular features of
neuroblastoma are highly predictive of clinical behavior. Therefore, neuroblastoma has served as a
paradigm for phenotypic risk assessment and treatment assignment whereby those at high risk for
disease relapse are given intensive multimodal therapy. Those at low risk for relapse can have treatment
intensity diminished in an attempt to avoid therapy-associated toxicity while still achieving a high rate of
cure. The predictive value of these biologic factors is important not only for the oncologist when
considering appropriate chemotherapy, but also for the surgeon when considering the timing and extent
of an operative resection in a child with neuroblastoma.

Treatment of children with neuroblastoma is based on risk stratification that takes into account clinical
and biologic variables predictive of relapse. The most important clinical variables appear to be age at the
time of diagnosis and stage at diagnosis. The most powerful biologic factors at this time appear to be
MYCN status and the histopathologic classification.
In addition, other biologic and molecular variables continue to be evaluated and the allelic status at
chromosomes 1p36 and 11q23 has been used to dictate the duration of therapy for certain patients.
Taken together, these variables defined the COG risk stratification used for recent clinical trials (Table
65.4). On the basis of these clinical and biological variables, infants and children with neuroblastoma
have been categorized into three risk groups predictive of relapse: low, intermediate, and high risk. The
probability of prolonged disease-free survival for patients in each group is >95%, >90%, and <30%,
respectively.

Table 65. 3 International Neuroblastoma Staging System Criteria

Stage Definition

1 Localized tumor with complete gross excision, with or without microscopic residual disease;
representative ipsilateral lymph nodes negative for tumor microscopically (nodes attached to and
removed with the primary tumor may be positive)

2A Localized tumor with incomplete gross excision; representative ipsilateral nonadherent lymph nodes
negative for tumor microscopically

2B Localized tumor with or without complete gross excision, with ipsilateral nonadherent lymph nodes
positive for tumor. Enlarged contralateral lymph nodes must be negative microscopically

3 Unresectable unilateral tumor infiltrating across the midline,* with or without regional lymph node
involvement

or

Localized unilateral tumor with contralateral regional lymph node involvement

or

Midline tumor with bilateral extension by infiltration (unre-sectable) or by lymph node involvement

4 Any primary tumor with dissemination to distant lymph nodes, bone, bone marrow, liver, skin, or
other organs (except as defined for stage 4S)

4S Localized primary tumor (as defined for stage 1, 2A, or 2B), with dissemination limited to skin, liver,
and bone marrow† (limited to infants younger than 1 year old)

*The midline is defined as the vertebral column. Tumors originating on one side and crossing the midline
must infiltrate to or beyond the opposite side of the vertebral column.
†Marrow involvement in stage 4S should be minimal (i.e., <10% of total nucleated cells identified as
malignant on bone marrow biopsy or on marrow aspirate). More extensive marrow involvement would
be considered to be stage 4. The metaiodoberuylguanidine scan (if performed) should be negative in the
marrow.

LOW-RISK DISEASE

Neither adjuvant chemotherapy nor radiation therapy appears to be necessary for the management of
most patients with low-risk disease. Surgical resection is currently the only therapy given to these
patients. In some cases, resection is not even performed, and these patients are simply observed. If
these patients experience disease relapse, most can be salvaged with additional therapy.

The COG study P9641 was conducted from 1998–2006 to evaluate primary surgical therapy for
biologically defined low-risk neuroblastoma. The overall strategy of this study was to treat patients with
low-risk neuroblastoma with resection and supportive care only. Adjuvant therapy was given only when
less than 50% of the tumor was resected or when symptoms developed that were life- or
organthreatening. Excellent survival rates were achieved in asymptomatic, low-risk patients with stage
2A/B neuroblastoma after operation alone.65 However, children with stage 2B disease who were older
or had diploid or unfavorable histology tumors fared somewhat worse.

Small, localized neuroblastomas in young infants tend to regress spontaneously. Based on this
observation, the COG study, ANBL00P2, included an arm of expectant observation for patients with
these lesions to further define their natural history. This study was designed to prove the hypothesis
that close biochemical and sonographic observation could be safely applied in infants with small adrenal
masses. Resection was reserved for those rare cases in which there was evidence of continued growth.
To be eligible, infants with an adrenal mass had to be <6 months of age when the mass was first
identified; the mass must be <16 mL in volume, if solid, or <65 mL if at least 25% cystic; and disease
must be limited to the adrenal gland. The results from this study confirmed that expectant observation
of infants with small adrenal masses leads to excellent 3-year EFS (97.7 ± 2.3%) and 100% overall
survival while avoiding operative intervention in >80% of patients.

INTERMEDIATE-RISK DISEASE

The COG study A3961 was conducted from 1997–2005 to further refine therapy for patients with
intermediate-risk disease. The overriding aim of this study was to maintain or improve survival while
minimizing both acute and long-term morbidity in patients with intermediate-risk neuroblastoma.
Patients received four of the most active agents against neuroblastoma: cyclophosphamide,
doxorubicin, carboplatin, and etoposide, given for either four cycles (favorable biology) or eight cycles
(unfavorable biology); cycles were given every 3 weeks. Radiation therapy was not used unless there
was progressive disease or an unresectable primary tumor with unfavorable prognostic features at the
end of chemotherapy. The 3-year overall survival for the entire group was 96%. Survival was 98% for
those with favorable biologic features and 93% for those with unfavorable features.

Table 65.4 Children’s Oncology Group Risk Stratification for Children With Neuroblastoma

Risk Stratification. INSS Stage. Age. Biology

*If tumor contains chromosomal 1p loss of heterozygosity (LOH) or unbalanced 11q LOH or if data are
missing, treatment assignment is upgraded to next group.

The most recent COG study, ANBL0531, sought to further refine the minimal therapy needed to achieve
these excellent outcomes for patients with intermediate-risk neuroblastoma. As such, many patients, as
defined by favorable clinical and biologic factors, received a further reduction in therapy. However,
those patients in whom there was loss of heterozygosity (LOH, i.e., loss of one of two normally paired
chromosomal regions) at chromosome 1p or 11q (unbalanced) were not eligible for this dose reduction,
as these findings have been shown to be independently associated with decreased progression-free
survival in patients with low- and intermediate-risk disease.

The overall surgical goal in intermediate-risk patients is to perform the most complete tumor resection
possible, consistent with preservation of full organ and neurologic function. This may necessitate leaving
residual disease adherent to critical structures. If a primary tumor is judged by the surgeon to be
unresectable, a diagnostic biopsy is generally obtained and chemotherapy initiated. Delayed operation is
performed after the prescribed number of cycles, as dictated by the group assignment.

Radiation is administered only to symptomatic intermediate-risk patients when there is a risk of organ
impairment due to tumor bulk not responding to initial chemotherapy. This will most often be
encountered in patients with epidural disease and symptoms of spinal cord compression.

Patients >12 years of age with localized tumors have an indolent clinical course, but ultimately have an
unfavorable outcome, and are considered for more intensive therapy.

HIGH-RISK DISEASE

For patients with advanced neuroblastoma, chemotherapy has been the mainstay of multimodality
treatment. Neuroblastoma is generally a chemotherapy-sensitive tumor, and multiagent chemotherapy
is usually effective in achieving at least a partial response in older children with disseminated disease.
However, this approach rarely is curative. The vast majority of these patients ultimately succumb to
chemotherapy-resistant disease, despite the use of increasingly intensive chemotherapy.

The general approach to treating patients with high-risk neuroblastoma has included intensive induction
chemotherapy, myeloablative consolidation therapy with stem cell rescue, and targeted therapy for
minimal residual disease. Stem cell harvest is typically performed after the first two cycles of induction
therapy, and resection of the primary tumor and bulky metastatic sites is attempted after the fifth cycle
(Fig. 65.7).

The CCG-3891 protocol enrolled patients with high-risk neuroblastoma between 1991 and 1996 and was
designed to assess whether myeloablative therapy, in conjunction with autologous BMT, improved EFS
when compared with chemotherapy alone and whether subsequent treatment with 13-cis-retinoic acid
would further improve EFS. Retinoids are vitamin A derivatives of 13-cis-retinoic acid. Retinoids decrease
the proliferation and expression of MYCN of neuroblastoma cell lines in vitro and induce morphologic
differentiation. The results from this double-randomization study demonstrated that the 3-year EFS was
significantly better in patients who underwent BMT during the first randomization (34%) than in those
who did not (22%; P = 0.034). In the second randomization, those who received 13-cis-retinoic acid after
BMT experienced a significantly better 3-year EFS (46%) than those who did not receive the retinoid
(29%, P = 0.027). Thus, currently, all patients with high-risk neuroblastoma receive oral 13-cis-retinoic
acid twice daily for 2 weeks and are then off therapy for 2 weeks. This treatment is continued for six
cycles (6 months total). Some intermediate-risk patients also receive 13-cis-retinoic acid. Unfortunately,
the long-term survival advantage for these patients is becoming less apparent. Nevertheless, autologous
stem cell transplantation and 13-cis-retinoic acid are now part of most current high-risk neuroblastoma
protocols.

Neuroblastoma cells are sensitive to antibody-dependent cell-mediated cytotoxicity and complement-

dependent cytotoxicity. Immunotherapy using anti-ganglioside antibodies targeting GD2, the
predominant antigen in neuroblastoma cells, appears to be a promising approach for the treatment of
advanced neuroblastoma. As the induction of antibody-dependent cell-mediated cytotoxicity with anti-
ganglioside GD2 antibodies is enhanced by cytokines such as granulocyte-macrophage colony-
stimulating factor (GM-CSF) and interleukin-2 (IL-2). A COG phase III trial was conducted to determine
whether treatment with ch14.18 and cytokines (GM-CSF and IL-2) together with 13-cis-retinoic acid
improved EFS and overall survival after autologous BMT, as compared with treatment with 13-cis-
retinoic acid alone in patients with high-risk neuroblastoma. The study was stopped early because
immunotherapy was superior to standard therapy (2-year EFS: 66% vs 46%, P = 0.01 and 2-year overall
survival: 86% vs 75%, P = 0.02).

The most recent COG high-risk neuroblastoma study, ANBL0532, had as its primary goal to test whether
further intensification of myeloablative therapy would improve the cure rate. Randomization to either
one myeloablative consolidation with a carboplatin/etoposide/melphalan preparative regimen or two
myeloablative consolidations, in which the initial regimen included thiotepa and cyclophosphamide,
occurred at the completion of induction chemotherapy. Another aim of this study was to determine
whether additional radiation therapy delivered to gross residual disease improved local control. At 4–6
weeks after stem cell transplantation, radiation therapy was administered to the region of the primary
tumor site, including involved adjacent lymph nodes. The target volume was the area of residual
disease, which was determined radiographically, after induction chemotherapy but prior to delayed
surgical resection, with an additional 1.5-cm margin added even if a complete resection was ultimately
achieved. Sites of persistent active metastatic disease prior to stem cell transplantation were irradiated
at the same time and with the same dose as the primary site. Those patients whose primary site
achieved a complete response at the end of induction therapy received 21.6 Gy to the site of primary
locoregional disease, whereas areas with gross residual disease were treated with an additional boost of
14.4 Gy (36 Gy total).

The third primary aim of ANBL0532 was to test the use of a dose-intensive topotecan-containing
induction regimen, substituting two cycles of dose-intensive cyclophosphamide and topotecan for the
first two cycles of induction used on A3973. Secondary surgical objectives embedded in this protocol
included to (1) to determine whether resection completeness is predictive of local control rate or EFS,
(2) prospectively describe the complications related to efforts at local control, and (3) to describe the
neurologic outcomes in patients with paraspinal primary tumors. ANBL0532 was recently closed, and
the results are pending.

International Neuroblastoma Risk

Group Risk Stratification

Although INSS has been shown to have prognostic relevance, there have been some difficulties with its
widespread use. The expertise and aggressiveness of the surgeon influence tumor stage, lymph node
sampling is done erratically, and patients who are simply observed without surgery cannot be
appropriately staged. Therefore, a uniform, pretreatment staging system that could be used easily
throughout the world and subject to real-time central review was sought. Montclair et al., on behalf of
the International Neuroblastoma Risk Group (INRG), proposed a new staging system in 2009 based on
tumor imaging rather than the extent of surgical resection.79 In this staging system, localized tumors are
staged based on the absence (L1) or presence (L2) of one or more of 20 image-defined risk factors
(IDRFs). Previously, Cecchetto et al. reported that the presence of one or more of these image-defined
surgical risk factors was associated with a lower complete resection rate and a greater risk of surgery-
related complications when attempting an initial resection of a localized neuroblastoma.

The IDRFs are listed in Box 65.1 and generally reflect encasement of vital structures, primarily vessels
and nerves, as determined by diagnostic imaging studies. Absence of these factors had previously been
shown to be associated with safe, complete tumor resection. In a review of 661 patients in the INRG
database, Monclair et al. found that INRG staging had prognostic significance. Patients with stage L1
disease had significantly greater 5-year EFS than those with stage L2 disease (90% ± 3% vs 78% ± 4%, P =
0.001).
In an effort to establish an international consensus on pretreatment risk stratification, the INRG task
force developed the INRG Classification System based on an analysis of 8800 patients treated for
neuroblastoma between 1990 and 2002. They used survival tree regression analyses with EFS as the
primary endpoint to test the prognostic significance of 13 potentially prognostic factors. The analyses
determined that seven of these prognostic variables could define 16 different pretreatment risk groups
(Table 65.5). Metastatic tumors are defined as stage M. These risk groups could then be divided into
four categories based on expected 5-year event-free survival: very low (>85% EFS, 28.2% of patients),
low (>75 to ≤85% EFS, 26.8% of patients), intermediate (≥50 to 75% EFS, 9.0% of patients), and high
(<50% EFS, 36.1% of patients) risk. Of note, analysis of the prognostic importance of histopathology has
been confounded by the inclusion of age, itself an independent prognostic factor, in the past. Therefore,
in the INRG classification schema, tumor differentiation and MKI are separated for risk stratification.

Box 65.1 Image-Defined Risk Factors for Primary Resection of Localized Neuroblastoma

Neck

1. Tumor encasing major vessel(s) (e.g., carotid artery, vertebral artery, internal jugular vein)

2. Tumor extending to base of skull

3. Tumor compressing the trachea

4. Tumor encasing the brachial plexus

Thorax

1. Tumor encasing major vessel(s) (e.g., subclavian vessels, aorta, superior vena cava)

2. Tumor compressing the trachea or principal bronchi

3. Lower mediastinal tumor, infiltrating the costovertebral junction between T9 and T12 (may involve
the artery of Adamkie-wicz supplying the lower spinal cord)

Abdomen

1. Tumor infiltrating the porta hepatis and/or the hepatoduodenal ligament

2. Tumor encasing the origin of the celiac axis and/or the superior mesenteric artery

3. Tumor invading one or both renal pedicles

4. Tumor encasing the aorta and/or vena cava

5. Tumor encasing the iliac vessels

6. Pelvic tumor crossing the sciatic notch

Dumbbell tumors with symptoms of spinal cord compression: Any location

Infiltration of adjacent organs/structures: Diaphragm, kidney, liver, duodenopancreatic block, and
mesentery

Adapted from Cecchetto G, Mosseri V, DeBernardi B, et al. Surgical risk factors in primary surgery for
localized neuroblastoma: the LNESG1 study of the European International Society of Pediatric Oncology
NB Group. J Clin Oncol. 2005;23:8483–8439.

Table 65.5 International Neuroblastoma Risk Group (INRG) Pretreatment Classification