Neuroblastoma stands out among childhood solid cancers because of its relative frequency, its intriguing biological features, and its formidable therapeutic challenges. It derives from neural crest progenitor cells that normally give rise to the sympathetic nervous system. About 500 new cases are diagnosed yearly in the United States, accounting for 8% of pediatric malignancies.1 It is the most common extracranial solid cancer of childhood1 and is the most commonly diagnosed neoplasm in the first month and year of life.2· 5 The median age at diagnosis is 2 years but neuroblastomas are seen during the teenage years and beyond.3 The estimated incidence of 1 out of 10,000 births may be inexact because the well known phenomena of spontaneous remission or maturation into benign forms4 may allow cases to escape detection.
An additional confounding element as regards frequency is neuroblastoma in situ. This lesion consists of the incidental finding at autopsy of neuroblastic nodules in the adrenal glands of up to 1 out of 40 infants aged 3 months or less who died for reasons other than cancer.5 Although histologically the neuroblasts in these nodules appear malignant, confirmatory evidence - eg, cytogenetic abnormalities - is lacking, and the nodules may actually represent normal embryological structures.
Stages in the normal differentiation of neural crest cells are reflected in the respective pathologic features of neuroblastoma, ganglioneuroblastoma, and ganglioneuroma. Primitive neuroblastomas consist of closely packed small cells with scant cytoplasm. Although the tumor cells resemble the early migrating neural crest cells of embryogenesis, they may show early differentiation by grouping into circular structures, or rosettes, with central cores of nerve fibers. When these features are absent, light microscopy may be unable to distinguish between neuroblastoma and other small round cell neoplasms such as rhabdomyosarcoma, Ewing's sarcoma, and peripheral neuroepithelioma. Clinical and biochemical findings usually indicate neuroblastoma but additional tests are sometimes needed.6 Electron microscopy can identify neurosecretory granules and other features characteristic of neuroblasts. Immunohistological detection of neuronspecific enolase - a cytoplasmic glycolytic enzyme - confirms neural crest derivation; cytogenetic studies may be needed, however, to rule out other primitive neuroectodermal tumors.7 Panels of antibodies directed at tumor-associated antigens are also helpful in diagnosis.8
Ganglioneuroblastomas contain more differentiated neuroblasts, with frequent rosettes, plus larger ganglion cells, or they may exhibit areas of widely differing histologies, ie, nest of primitive neuroblasts admixed with areas of highly mature neural tissue. In contrast to ganglioneuroblastomas, which possess metastatic potential, ganglioneuromas are composed of large, differentiated, non-metastasizing ganglion cells. These well-encapsulated tumors may cause symptoms by impinging on adjacent tissues or by producing biologically active substances.9·10
Neuroblastomas have been reported to evolve into ganglioneuroblastomas and ganglioneuromas. The transformation from a primitive, highly malignant phenotype into a benign ganglion cell may be spontaneous or may follow chemo- or radiotherapy.9 Most often, the process is incomplete; cells with malignant potential persist and eventually cause tumor progression and death.
The young age of most patients with neuroblastoma and the nondividing state of cells in the mature autonomic nervous system suggest that causative events occur early in life or affect parental germ cells. Familial aggregations of neuroblastoma raise the issue of inherited susceptibilities.11·12 Epidemiologic data,13 plus reports of neuroblastoma in patients with the fetal hydantoin and fetal alcohol syndromes,14·15 implicate prenatal teratogenic-oncogenic insults in some cases. However, no excessive association with idiopathic developmental defects, other than a possible increased frequency of skull and brain anomalies,16 has been noted. The synchronous occurrence of neuroblastoma and other neurocristopathies, or entities of neural crest origin, such as neurofibromatosis17 or aganglionic megacolon,18 while intriguing in terms of suggesting a common genetic origin, are so rare that a random association cannot be excluded. To date, the same conclusion applies to the rare association of neuroblastoma with other genetic disorders such as cystic fibrosis19 or the Beckwith- Wiedemann syndrome.20 In sum, there is no consistent linkage of neuroblastoma with a single etiologic agent or genetic disease.
The presence of nonrandom deletions of the short arm of chomosome 1 in 70% of neuroblastoma tumors and cell lines21 has stimulated interest in a suppressor gene mechanism underlying neuroblastoma. This schema - which is consistent with Knudson's "twohit" hypothesis,11·22 a popular etiologic model for embryonal neoplasms - postulates that the deletion or mutation of one suppressor allele, when followed by the inactivation or loss of the homologous gene on the sister chromosome, leads to malignant transformation. In sporadic cases, both mutations are postzygotic events. In familial cases, the first hit is an inherited germ-line mutation; this may predispose to an early onset and to the development of multiple primaripc 11, 12
Chromosome Ip deletions may not be primary events in the genesis of neuroblastoma since they have not been detected in localized tumors23 and, in advanced disease, loss of discrete segments of both Ip chromosomes has not been found. More sensitive techniques may detect minute chromosome Ip defects. Confirmed or suspected neuroblastomas have been described in patients with constitutional genetic dysmorphic syndromes, and in otherwise normal patients with constitutional chromosomal rearrangements.24 However, in contrast to the inherited deletions of chromosomes lip in Wilms' tumor and 13q in retinoblastoma, two other embryonal neoplasms, no characteristic pattern of constitutional chromosomal abnormalities has emerged that might give insight into a genetic basis for neuroblastoma. It is possible that a constitutional deletion large enough to include a gene predisposing to this tumor and large enough to be seen microscopically results in early fetal death.
Chromosomal homogeneously staining regions (HSR) and double-minute chromosomes (DM) - cytogenetic hallmarks of neuroblastomas25·26 - represent multiple repeating, ie, amplified, DNA segments which stain uniformly with trypsin-giemsa, in contrast to the alternating dark and light bands normally seen in metaphase chromosomes. HSRs may break down and give rise to DMs which in turn reintegrate into chromosomes in an apparently nonspecific distribution, to re-form HSRs. Because of homology to the c-rrvyc oncogene, the amplified DNA segment in neuroblastoma cells is designated N-rrryc; the normal single copy is on chromosome 2. The function of the M-myc gene product is unknown, but the association of M-myc amplification with tumor progression suggests that increased expression influences the biological behavior of the tumor cell (see section on Prognostic Factors). A role, if any, in tumor initiation for N-myc or other oncogenes, such as N-ras, is unclear.26
Most symptoms and signs of neuroblastoma are attributable to local problems from primary tumors or metastases. However, some findings, such as diarrhea, hypertension, ataxia, and ocular abnormalities, are not directly related to mass or metastatic effect. Unsuspected tumors may be detected on routine physical examination, by screening for urinary catecholamines,27 or when investigative procedures such as xrays are performed for other reasons, eg, scoliosis. Because of its protean features, neuroblastoma has been mistaken for arthritic disease, infection, leukemia, intestinal disease, battered-child syndrome, and primary neurologic disorders; delays in diagnosis have been the consequence.
Neuroblastoma arises most often in the retroperitoneum (adrenal medulla more often than the paraspinal ganglia); less common sites of origin are the posterior mediastinum (15%), the pelvis (5%), and rhe neck (<5%). Multiple primaries can occur and are considered a manifestation of an inherited predisposition to neuroblastoma."·12 Up to 10% oí neuroblastomas have no identifiable site ot origin, but most, because oí location deep within the body or because of rapid growth rate, are very large at diagnosis. Metastases are found in 70% of newly diagnosed patients and involve cortical bone, bone marrow, lymph nodes, liver, and subcutaneous tissue, but spread to the lung or brain parenchyma is rare. The tendency of neuroblastoma to affect bones is particularly evident in the frequency of skull metastases; these may form visible masses or may expand inward, causing raised intracranial pressure. Sphenoid bone involvement or tumor invasion of the retrobulbar soft tissues produces the distinctive orbital proptosis and ecchymoses of this malignancy. Extensive hepatic involvement and skin metastases are characteristic ot neonatal neuroblastoma.
Neuroblastomas arising from para-spinal ganglia have a tendency to grow through the intervertebral foramina and compress the spinal cord.28 These "dumbbell" or "hourglass" tumors are especially common with neuroblastomas in the posterior mediastinum, but they may occur with paravertebral primaries in the abdomen, pelvis, or neck. Patients may be asymptomatic or may experience localized back pain, limb weakness, sensory deficits, incontinence, or rapid onset of paraplegia.
Upper thoracic or cervical neuroblastomas may cause Horner's syndrome (miosis, ptosis, enophthalmos, anhidrosis) or heterochromia iridis.29 Elevated levels oí catecholamines may cause hypertension.9·10 However, marked paroxysmal effects from catecholamine release, as seen in pheochromocytomas, are uncommon in neuroblastoma, possibly because defective storage results in the rapid inactivation of these substances. Chronic watery diarrhea can occur with ganglioneuroblastomas or ganglioneuromas. Vasoactive intestinal peptide is elevated in the blood of affected patients and can be detected in differentiating and mature ganglion cells but not in primitive neuroblasts. î0 The diarrhea resolves following tumor resection.
Myoclonic encephalopathy is another remote effect of relatively mature neuroblastomas. Clinical findings include ataxia, opsoclonia (multidirectional darting eye movements), and myoclonia. 51 Infants or children with unexplained cerebellar signs should be evaluated for occult neuroblastoma. Speculation regarding pathophysiology centers on an autoimmune phenomenon with cross-reactivity between tumor and cerebellar antigens. Treatment with corticosteroids or corticotropin (ACTH) may control the symptoms. Despite complete tumor removal, many surviving patients have neurological problems, including psychomotor delay, extrapyramidal deficits, and further episodes of acute cerebellar dysfunction. 32, 33
Essential pre-surgical and pre-treatment studies in patients with suspected neuroblastoma include quantifying biochemical markers associated with this tumor (Table 1). These may aid in diagnosis and in subsequent monitoring of disease status. A team effort should be organized among surgeons, oncologists, pathologists, and researchers to assure procurement of tumor tissue for investigative studies. Each case should be studied to the maximum; an ultimate impact on long-term survival depends on a better understanding of the biology of neuroblastoma.
Determining the extent of disease is crucial for prognostication, for choosing appropriate therapy, and forjudging treatment efficacy, especially when disease clinically appears to be localized. Improvements in imaging and in detecting bone marrow involvement have increased the accuracy of staging and may render comparisons with earlier descriptions or discussions of disease extent difficult.
In addition to histopathologic analysis of primary or metastatic tumor, a characteristic clinical picture in conjunction with typical bone marrow and biochemical findings establishes the diagnosis. Radiographic or radionuclide studies alone are insufficient for this purpose. 34
Because of either overproduction by individual cells or rapid release due to defective storage within tumor cells, 70% to 95% of patients at diagnosis have increased urinary excretion of vanillylmandelic acid (VMA), the main metabolite of epinephrine and norepinephrine, and/or of homovanillic acid (HVA), the main metabolite of dopamine. 35 Levels of these catecholamines reflect disease status after treatment begins. Several factors account for differences in reported rates of positivity. Some laboratory methods have limited sensitivity, other methods give misleading results if foods such as vanilla-containing products are not excluded during a urine collection. Certain medicines may give falsely high catecholamine levels. Reliable results are obtained when dietary restrictions are carried out and catecholamine content is correlated with the amount of creatinine present in a 24hour urine collection.
Figure. Total body scintigraphy in a patient with metastatic neuroblastoma 42 hours after injection of the l3ll-3F8. the monoclonal antibody specific for the ganglioside GD2. A large primary tumor was found in the lower abdomen with involvement of the right pelvis and sacrum. The sharp cut-off in the superior border of the tumor coincided with the lower edge of the radiation port through which external beam radiation (2,000 cGy) was given two weeks earlier. Metastatic disease was found in the cranium, left neck, spine, right chest, and left humerus. Except for the cervical tumor, all the other metastatic sites were not suspected from previous evaluations, based on bone scan or clinical examination.
Serum neuron specific enolase (NSE) is elevated (above 15 ng/mL) at diagnosis in most patients with neuroblastoma. 5<s Levels correlate with extent of disease: mild increases are noted with localized tumors, and marked increases (to greater than 100 ng/mL) occur in over 50% ot patients with widespread disease (see section on Prognostic Factors). Rarely, high levels are seen with leukemia, Ewing's sarcoma, and Wilms tumor even though NSE is not detected by histochemistry in these neoplasms.
The disialoganglioside GD2 >s a surface glycolipid uniformly expressed by human neuroblastomas. y' Circulating Gj-,7 's elevated (above 10 pmol/mL) in patients with neuroblastoma, irrespective of stage; it decreases with response to therapy and rises with tumor recurrence. Patients with ganglioneuroma or ganglioneuroblastoma have normal levels, suggesting an association of shedding of Gn, with the undifferentiated phenotype. Since GD2 is present on other solid tumors which may also shed GD2. 'ts diagnostic utility may be limited.
Recommended Evaluations in Patients with Neuroblastoma
Bone marrow studies are of paramount importance in defining disease status. Detection of characteristic syncytial clumps is enhanced by aspirating and biopsying multiple sites. ,y The yield of positi vity rises further upon screening with monoclonal antibodies'9,40 since solitary neuroblasts in the marrow may appear as lymphocytes on routine staining. Diffuse marrow involvement may suggest acute leukemia41·42; if urinary catecholamines are not elevated, electron microscopy or monoclonal antibody studies can pinpoint the diagnosis.6
Computed tomography (CT) is superior to conventional radiography and sonography tor detecting primary or metastatic neuroblastoma and has improved upon other imaging techniques for indicating tumor operability.42 By virtue of its multiplanar imaging capacity, magnetic resonance imaging may eventually supplant CT.4Î X-rays show speckled calcifications in 50% of primary neuroblastomas; this finding is more common with CT Sonography is a useful non- invasive means for monitoring retroperitoneal disease but, unlike CT, it does not demonstrate bone disease or extension above the diaphragm. When staging a paravertebral primary, the intrathecal administration of contrast material allows assessment of tumor extension into the spinal canal, a major point of concern with these primaries even in the absence of neurologic signs.
Radionuclide bone scans may be more sensitive than radiographic skeletal surveys for evaluating skeletal involvement.44 Radiolabeled metaiodobenzylguanidine (MIBG) and monoclonal antibodies allow imaging of non-osseous, in addition to osseous, sites. MIBG structurally resembles catecholamine precursors and is specifically taken up by adrenergic tissue; greater uptake is correlated with higher catecholamine production.45 UJ13A and 3F8 are murine monoclonal antibodies that are highly sensitive and specific in detecting primary as well as metastatic neuroblastoma in humans (Figure).46·47 Murine monoclonal antibodies, however, are not well-suited for repeated imaging studies because they induce neutralizing antibodies in humans.
Important investigative studies will evolve with increased understanding of neuroblastoma. Currently, they include identification and characterization of tumor markers (eg, gangliosides)37; chromosomal mapping for gene deletions (eg, chromosome Ip)21 or amplifications (eg, N-m^c)48·49; assessing levels of oncogene expression7; sequential monitoring of resistance factors (eg, multi-drug resistance)50; and delineation of growth factors and receptors (eg, epidermal growth receptor).51 These studies require fresh frozen tissue specimens, or tumors explanted in vitro or passaged into athymic rodents. Opportunities to study tumor tissue, at diagnosis and later, should not be missed.
Age at diagnosis and extent or pattern of disease spread are long-established independent prognostic variables in patients with neuroblastoma. Recent observations accord prognostic significance to biological markers such as N-m^c copy number and serum ferritin. Other findings of lesser prognostic importance include the site of primary tumor and the presence of opso-myoclonia.
Infants with neuroblastoma have a better prognosis than older patients. This derives partly from the greater percentage of patients less than 1 year old who present with prognostically favorable forms of neuroblastoma; the pattern holds, however, for poor-risk stages of the disease as well.52·53 Conversely, advanced neuroblastoma diagnosed in patients over 6 years of age may follow a more indolent course, resulting in longer survival before death. 3,54
The widely used clinical staging system of Evans classifies patients on the basis of physical and surgical findings, imaging studies, and bone marrow status.52 Local tumors are divided by size into Stage I, II, or III; disseminated disease is divided into Stage IV or IV-S ("Special"). The IV-S category denotes a specific pattern of disease: an undetectable or small, usually adrenal, primary (Stage I or II), and involvement of the liver, skin, and/or bone marrow; skeletal lesions or an unresectable primary exclude a patient from the IV-S grouping. Patients with Stage I or Stage IV-S do well; those with Stage IV do poorly. The prognostic value of Evans's system is less clear for patients with Stage II or III disease, 10% to 50% of whom relapse.5558 New approaches for detecting metastatic disease may improve staging accuracy and thereby help resolve the problem. In addition, however, the system does not base stage on tumor resectability or regional lymph node involvement; yet localized, but inoperable neuroblastoma - informally labeled Stage IH-U ("Unresectable") - has a poor prognosis, 59 and several studies find the same for lymph node disease.56'58 A proposed modification takes this last finding into account. 34
The favorable prognosis associated with extraabdominal primaries52 and with opso-myoclonia51 is attributable to the high proportion of these patients who have localized, relatively mature tumors. Nevertheless, despite a paucity of supporting statistical data, it is generally accepted that, stage for stage, patients with tumors arising from sympathetic ganglia in the neck, thorax, pelvis, and, possibly, the abdomen as well, do better than patients with adrenal primaries.
N-myc amplification (more than 3 to 10 copies) portends refractory disease regardless of age or stage.48·49 This applies to clinically localized or Stage IV-S tumors containing multiple N-myc copies. N-m^c amplification is less clearly prognostic in Stage IV patients because more than 40% of those who develop progressive disease have a single N-m^c copy. Laboratory work suggests that N-myc expression may affect the clinical course of neuroblastoma. For example, retinoic acid-induced differentiation of neuroblastoma cell lines is preceded by decreases in N-m\c-specific mRNA content.60
Elevated serum ferritin (above 150 ng/mL) indicates a poor prognosis. Thus, most patients with Stage III and IV neuroblastoma have high levels at diagnosis,61 as do the few patients with Stage I, II and IV-S tumors who develop progressive disease. During remission, levels decline to the normal range. The relation between ferritin and prognosis remains unexplained, but the ferritin appears to originate in neuroblasts since, for example, circulating human ferritin is detected in nude mice xenografted with human neuroblastomas.
The histopathological classification of Shimada et al62 constitutes a prognostic system based on age at diagnosis and tumor morphology. Favorable prognostic features include a stroma-rich organizational pattern, a greater degree of tumor cell maturation, and a low mitosis and karyorrhexis index.
Chromosomal ploidy may have prognostic significance. Flow cytometry and cytogenetic studies demonstrate hyperdiploidy in Stage IV-S tumors and triploidy in localized, good-prognostic tumors, versus near diploidy, pseudo diploidy, or hypotetraploidy in poor-risk neuroblastomas.23·63 These findings could indicate that chromosomal instability of Stage I, II, and IV-S tumors is such that, after a finite number of divisions, senescence sets in: limited proliferative capacity might translate into a benign clinical course.
Marked elevations in serum neuron specific enolase (NSE) (above 100 ng/mL) are predictive of a poor outcome for Stage III patients under 2 years of age and for infants with Stage IV disease. 36 Most patients with Stage I1 II, or IV-S disease have levels under 100 ng/ mL. Tumor burden and tumor cell NSE content are not obvious determinants of serum levels since these can vary widely in patients with apparently similar amounts of tumor, and cytoplasmic NSE is variably expressed in neuroblastoma cell lines from Stage IV patients.
Catecholamine levels at diagnosis are not predictive of outcome although, in one study, a low VMA:HVA ratio correlated with shorter survival of Stage IV patients. 35 This finding was interpreted as suggesting that biochemically primitive neuroblastomas, ie, those lacking the dopamine beta-hydroxylase enzyme that converts dopamine to norepinephrine, may manifest more virulent behavior clinically.
In sum, more accurate prognostication results from correlating clinical stage and age with biological parameters. One retrospective analysis, for example, builds an accurate prognostic system on age and ferritin alone, with further discrimination afforded by considering stage and the Shimada histopathology classification.64
Recent developments have facilitated identification of patients who do well with no, minimal, or surgical treatment alone, and have improved the outlook for patients who require intensive therapy to arrest lethal tumor growth.
Localized Tumors without Lymph Node Involvement
Survival rates of 80% to 100% are achieved with surgery alone in the 10% to 15% of neuroblastoma patients who at diagnosis have grossly resectable, localized tumors without lymph node involvement.52·57 These results hold when tumor is entirely excised as well as when microscopic residual disease confined to the tumor bed remains after resection of a tumor adherent to, but readily detached from, adjacent structures such as vertebral bodies. Post-surgical radiation or low-to-moderate-dose chemotherapy has not affected the survival rate.55'58
Improved staging techniques, including MIBG scans, radiolabeled monoclonal antibodies (Figure), and extensive marrow evaluations, allow a more confident assignment of patients to this grouping. Moreover, biological parameters such as high serum ferritin64 and N-myc amplification48·49 can identify the few patients in this category who are ar risk for progressive disease and who therefore require more intensive therapy (see below).
Stage IV-S Disease
Five to ten percent of patients with neuroblastoma have IV-S disease, and up to 90% have survived with minimal treatment.33,65·66 Most deaths result from coagulopathies, treatment toxicities, or compression of vital organs by a rapidly enlarging liver. Therapeutic approaches have included low-dose chemotherapy (usually including cyclophosphamide and/or vincristine), low-to-moderate doses of radiation to reverse hepatomegaly, and marsupialization with silastic patches to relieve intraabdominal pressure.
Biological factors may identify the 10% to 20% of clinical Stage IV-S patients who will later develop lethal Stage IV disease. A high serum ferritin, 6^ N-myc amplification,48·49 or serum neuron-specific enolase above 100 ng/mL36 may favor the use of intensive systemic therapy. In the absence oí these biological findings and of dangerous organomegaly, infants with IV-S disease may do well without any treatment.
Inoperable (Stage IH-U ) or Metastatic (Stage IV) Disease
Neuroblastoma was labeled the therapeutically most resistant extra-cranial solid cancer of childhood when early efforts utilizing surgery, chemotherapy, and radiation did not alter the dismal outlook for the 70% of patients who have unresectable disease at diagnosis. In recent years, however, sophisticated use of these treatment modalities has prolonged the duration of survival of poor-risk patients; further benefit may accrue from the advent of newer modalities, such as targeted radio- or immunotherapy.
Chemotherapy remains the key to control of widespread disease. The role of initial surgery in this context is primarily to obtain tissue for diagnostic and investigative studies; the extent of resection does not influence long-term outcome.67 If chemotherapy causes disease regression, an attempt at surgical removal of residual masses - "second-look" surgery - may be worthwhile.68 If complete removal of viable tumor is not possible, however, heroic debulking procedures will not be beneficial and, by risking damage to vital structures, may jeopardize susbequent delivery of chemo- or radiotherapy.
Although neuroblastoma is radioresponsive, definitions of curative dosing are not available. Most studies69 examine prevention of recurrence in patients with local disease, but the importance of radiation is unclear since, as noted above, many of these patients are cured with surgery alone. For unresectable disease, 2000 to 3500 cGy can achieve local control33; reports of progression despite 4500 cGy,69 however, underscore the heterogeneity of neuroblastoma. Radiobiological laboratory studies favor the use of hyperfractionated radiation since this method is known to reduce toxicity to normal tissues and may not compromise the anti-neuroblastoma effects of the radiation.70
Induction of remission represents a key first step toward long-term disease-free survival. Standardized response criteria have been proposed to facilitate comparisons of treatment results.34·71 Complete remission indicates absence of all chemical and radiological evidence of neuroblastoma, plus no viable tumor at the primary site on surgical evaluation. Good partial remission indicates persistence of microscopic, viable neuroblastoma at the primary site.
Cell kinetic data and the dose-responsiveness of neuroblastoma noted clinically as well as in the laboratory have served as the basis for rationally designed induction chemotherapy protocols.72'74 Most regimens employ the principal active anti-neuroblastoma agents, ie, cyclophosphamide, vincristine, adriamycin, cisplatin, and tenoposide (VM-26) or its congener etoposide (VP-16). The simultaneous75 or rotating76 use of multiple drugs has also been applied in an attempt to avert the emergence of drug-resistant disease. Intensive protocols,75'78 often supplemented with local irradiation and surgery, have produced durable responses in patients who 1) have inoperable localized disease plus lymph node involvement33; 2) are categorized as Stage IV on the basis of distant lymph node involvement79; or 3) have Stage IV disease but are less than 1 year old at diagnosis. 53 Success has been elusive in the largest high-risk subgroup, ie, patients over 1 year of age at diagnosis of advanced disease. Aggressive approaches achieve complete or good partial remissions in 70% of these patients, which is an improvement upon the remission rate of less than 50% obtained with prior protocols.74·80 Response criteria indicate, however, that most responders do not have disappearance of all evidence of disease (Table 2).
The dose-responsiveness of neuroblastoma provides the rationale for the use of treatment intensification with bone marrow transplantation in this disease. The "massive therapy" approach has been applied for consolidating remission status, and for achieving remissions in patients with disease resistant to conventional doses of chemoradiotherapy. Alkylating agents (primarily melphalan, but also busulfan, thiotepa, and cyclophosphamide) have been the drugs most commonly used for tumor ablation prior to bone marrow transplantation. These drugs are used because of their manageable extra-medullary toxicity and because of the known sensitivity of neuroblastoma to this pharmacologic class. Early studies with high-dose melphalan alone induced responses if not cures in 15% to 70% of patients with measurable disease,81·82 and was associated with prolonged disease-free survival, with no further therapy, in up to 25% of patients in complete remission at the time the drug was given.76·82 High doses of alkylating agents are currently administered with other drugs and/or total body irradiation.83-87
Autologous and allogeneic bone marrow transplants have been used but most experience has been with the former primarily because of the lack of suitable donors. In terms of survival, a superiority of one method of marrow rescue over another has not been demonstrated.88 The same holds for the different approaches for ridding autologous marrow of contaminating neuroblasts, although in vitro models attest to depletion efficacy.89 The most widely used purging technique involves monoclonal antibodies linked to magnetic microspheres89·90; other methods in use include monoclonal antibodies/human complement91 and cyclophosphamide congeners.87
The ultimate test of treatment efficacy is duration of disease-free survival when patients are taken off all therapy. Although the bone marrow transplant strategy is not curative of gross residual or progressive disease, it does appear useful for consolidating remissions and prolonging relapse-free survival. Currently, one third of patients who have no detectable disease when transplanted remain in remission off all therapy for up to one to two years. Since late relapses can occur, follow-up is too short to know the final cure rate.
Bone marrow transplantation exacts a tremendous toll on patients, families, and medical staff. Moreover, it carries the risk of long-term physical sequelae such as second malignancies. Therefore, despite encouraging therapeutic results with the transplant approach, other means tor consolidating remissions are urgently needed. Targeted radiotherapy and immunologic manipulations are promising possibilities. The favorable experience with radiolabeled MiBG45 in relapsed patients may prompt its earlier use in poor-risk patients. Radiolabeled monoclonal antibodies, by virtue of tumor specificity, also have therapeutic potential ( Figure). 92,93 In preliminary trials using 131I-3F8, short-term anti-tumor responses were noted.94 Bone marrow suppression has been a major limitation of targeted radiotherapy. The use of autologous marrow support will likely allow the administration of higher, more effective doses of radiation.
Advanced Neuroblastoma: Induction Regimens in Patients Over 1 Year of Age
Attempts to direct host immune mechanisms against neuroblastoma are under investigation. 3F8, like other murine IgG5 antibodies, activates human complement and effects antibody-dependent cellar cytotoxicity (AEXDC) in vitro. Intravenous injections of 3F8 have elicited inflammatory reactions that may account for tumor regressions seen in a phase I clinical trial.95 In the presence of 3F8, neuroblastoma cells are efficiently killed by human lymphocytes and neutrophils. The use of recently available leukocyte-activating cytokines such as interleukin-2,96 granulocytemacrophage colony stimulating factor (GM-CSF),97 and tumor necrosis factor98 may augment the efficacy of tumor-specific monoclonal antibodies.
The identification of prognostic factors has greatly facilitated the rational choice of therapies in individual patients. Intensive chemotherapy, supplemented with radiation and surgery, has increased the remission rate of patients with widespread disease. The persistence of microscopic foci of malignant cells, however, remains a difficult hurdle for long-term diseasefree survival. Highly toxic myeloablative therapies have had at most a modest impact on the overall cure rate of poor-risk patients. The use of novel biological therapies has provided new information on the mechanisms and potentials of immune-mediated tumor cytotoxicity. Timely clinical trials are needed to test their role in adjuvant treatment of occult microscopic disease.
1. Young JL Jr, Miller RW: Incidence of malignant tumors in U.S. children. J Pediatr 1975; 86:254.
2. Bader KL, Miller RW: US cancer incidence and mortality in the first year of life. AmJ Do Child 1979; 133:157.
3. Kaye JA, Warhol MJ, Kretschmar C, et al: Neuroblastoma in adults: Three case reports and a review of the literature. Cancer 1986; 58:1149.
4. Evans AE, Gcrson J. Schaufer L: Spontaneous regression of neuroblastoma. Natl Ctmcer Inst Monogr 1976; 44:49.
5. Türkei SB, ltabashi HH: The natural history of neuroblasts cells in the fetal adrenal gland. Am ; Pathol 1974; 76:225.
6. Triche TJ, Askin FB: Neuroblastoma and the differential diagnosis of small-, round-, blue-cell tumors. Hum PathiA 1983; 14(7):569.
7. Thiele CJ. McKeon C, Triche TJ, et al: Differential protooncogene expression characterizes histopathologically indistinguishable tumors of the peripheral nervous system. ] Clin Invest 1987; 80:804.
8. Oppedal BR, Brandtzaeg P, Kemshead JT: Immunohistochemical differentiation of neuroblastomas from other small round cell neoplasms of childhood using a panel of mono- and polyclonal antibodies. Histopathology 1987; 11:363.
9. Greenfield LJ, Shelley WM: The spectrum of neurogenic tumors of the sympathetic nervous system: Maturation and adrenergic function. J Nad Cancer Institute 1965; 35:215.
10. Weinblatt ME, Heisel MA, Siegel SE: Hypertension in children with neurogenic tumors. Pediatrics 1983; 71:947.
11. Knudson A, Strong LC: Mutation and cancer: Neuroblastoma and pheochromocvtoma. Am ) Hum Genet 1972; 24:514
12. Kushner BH, Gilbert F, Helson L: Familial neuroblastoma: Case reports, literature review, and etiologic considerations. Cancer 1986; 57:1887.
13. Kramer S, Ward E, Meadows AT, et al: Medical and drug risk factors associated with neuroblastoma: A case-conttol study. ] Natl Cancer Institute 1987; 78:797.
14. Kinney H. Faix R, Brazy J: The fetal alcohol syndrome and neuroblastoma. Pediatrics 1980; 66:130.
15. Ehrenhard LT, Chaganti RSK: Cancer in the fetal hydantoin syndrome (letter). Lancet 1981; 2:97.
16. Sy WM. Edmonson JH: The developmental defects associated with neuroblastoma - Etiologic implications. Cancer 1968; 22:234.
17. Kushner BH1 Hajdu SI. Helson L: Synchronous neuroblastoma and von linghausen's disease: A review of the literature. ) Clin Oncol 1985; 3:117.
18. Carachi R, Auldist AW. Chow CW: Neuroblastoma and Hirschsprung's (neuroblastom und hirschs,>rungsche Krankheit). Z Kmderchir 1982; 35:24
19. Moss RB, Blessing-Moore J, Bender SW. et al: Cystic fibrosis and Pediatrics 1985; 76:814.
20. Emery LG. Shields M. Shah NR. et al: Neuroblastoma associated with Wiedemann syndrome. Cancer 1983; 52:176.
21. Gilbert F, Feder M, Balaban G, et al: Human neuroblastomas and abnormalities chromosomes I and 17. Cancer Res 1984; 44:5444
22. Knudson AG: Hereditary cancer, oncogenes, and antioncogenes. Cancer Res 1985;
23. Kaneko Y. Kanda N, Maseki N, et al: Different karyotypic pattern in early and advanced stage neuroblastomas. Cancer Res 1987; 47:311.
24. Sanger WG. Howe J, Fordyce R. et al: inherited partial trisomy #15 complicated neuroblastoma. Cancer Genet Omogenei 1984; 11:153.
25. Brodeur GM, Seeger RC: Gene amplification in human neuroblastomas: Basic mechanisms and clinical implications. Cancer Gene! Cytogenet 1986;
26. Israel MA: The evolution of clinical molecular genetics - Neuroblastoma as a model tumor. Am } Pediatr Hematol Oncol 1986: 8:163.
27. Woods WG, Tuchman M: Neuroblastoma: The case for screening infants in America. Pediatrics 1987; 79:869.
28. Hayes FA. Thompson El. Hvizdala E, et al: Chemotherapy as an alternative laminectomy and radiation in the management of epidural tumor. } Pediatr 1984;
29. Jaffe N, Cassady JR. Filler RM. et al: Heterochromia and Horner syndrome associated with cervical and mediastinal neuroblastoma. J Pediatr 1975;
30. Mendelsohn G, Eggleston JC, Olson JL, et al: Vasoactive intestinal peptide and its relationship tu ganglion çc\\ differentiation in neuroblastic tumors. Lab Invest 1979;
31 - Altman AJ, Baehner RL: Favorable prognosis for survival in children with coincident opso-myoclonus and neuroblastoma. Ciincer 1976;
32. Jones A, Groover R. Smithson W: Acute cerebellar encephalopathy (ACE): Its natural history and relationship to neuroblastoma (NB). Abstracted. Prue Am Soc Oncol 1984; 3:86.
33. Rosen EM. Cassady JR, Frantz CN: Neuroblastoma: The Joint Center for Therapy/Dana-Farbcf Cancer Institute/Children's Hospital experience. I Clin Oncol 1984;
34. Brodeur GM . Seeger RC, Barrett A, et al: International criteria for diagnosis, staging. and response to treatment in patients with neuroblastoma, in Evans AE, et al Advances in Neuroblastoma Research. New York, Alan R. Liss, 1987 (in press).
35. Laug WE. Siegel SE, Shaw KNF, et al: Initial urinary catecholamine concentrations and prognosis in neuroblastoma, Pediatrics 1978; 62:77.
36. Zeltler PM, Marangos PJ, Evans AE, et al: Serum neuron-specific enolase in children with neuroblastoma: Relationship to stage and disease course. Cancer 1986;
37. Wu ZL, Schwartz E, Seeger R, et al: Expression of GD2 ganglioside by untreated primary neuroblastomas. CanceT Res 1986; 46:440.
38. Franklin IM, Pritchard J: Detection of bone marrow invasion by neuroblastoma improved by sampling at two sites with both aspirates and trephine biopsies. I CIm Pathol 1983; 36:1215.
39. Moss TJ, Seeger RC, Kindler-Rohrbom, et al: lmmunohistologic detection phenotyping of neuroblastoma cells in bone marrow using cytoplasmic neuron specific enolase and cell surface antigens, in Evans AE, D'Angio GJ. Seeger RC (eds): Advances in Neuroblastoma Research. New York, Alan R. Liss, 1985. ?
40. Cheung N-KV. Von Hoff DD, Strandjord SE, et al: Detection of neuroblastoma cells in bone marrow using GD2 specific monoclonal antibodies. i Clin Oncol 1986; 4:363.
41. Quinn JJ, Airman AJ: The multiple hematologic manifestations of Am J Pediatr Hematol Oncol 1979; 1:201.
42. Stark DD, Moss AA, Brasch RC, et al: Neuroblastoma: Diagnostic imaging staging. Pediatr Radiol 1983; 148:101.
43. Dietrich RB. Kangarloo H, Lenarsky C, et al: Neuroblastoma: The role of MR imaging. AJR 1987;
44. Heisel MA, Miller JH, Reid BS1 et al: Radionuclide bone scan in neuroblastoma. Pediatrics 1983;
45. Feine U, Muller-Schauenburg W, Treunet J, et al: Metaiodobenzylguanidine (MlBG) labeled with 123I/"1! in neuroblastoma diagnosis and follow-up treatment with a review of the diagnostic results of the international workshop of pediatric held in Rome. September 1986. Med Pediatr Oncol 1987; 15:181.
46. Goldman A, Gordon I, Kemshead J: Immunolocalization of neuroblastoma using radiolabeled monoclonal antibody UJ13A. ) Pediatr 1984;
47. Miraldi FD, Nelson AD. Kraly C, et al: Diagnostic imaging of human neuroblastoma with radiolabeled antibody. Radiology 1986; 161:413.
48. Seeger RC. Brodeur GM, Sather H, et al: Association of multiple copiesof the oncogene with rapid progression of neuroblastomas. N Engl I Med 1985; 313:1111.
49. Tsuda T, Obara M, Hirano H, et al: Analysis of n-MYC amplification in relation disease stage and histologic types in human neuroblastomas. Cancer 1987; 60:820.
50. Meyers MB, Biedler J: Evidence for reverse transformation in multidrug-resistant human neuroblastoma cells, in Evans AE. et al: Aduinces m Neuroblastoma New York. Alan R. Liss, 1987 (in press).
51. Rettig WJ, Spengler BA, Garin Chesa P, et al: Cootdinate changes in neuronal phenotype and surface antigen expression in human neuroblastoma cell Cancer Res 1987; 47:1383.
52. Evans AE: Staging and treatment of neuroblastoma. Cancer 1980; 45:1799.
53 . Kretschmar CS, Frantz CN , Rosen EM, et al: Improved prognosis for infants with stage IV neuroblastoma. } CIm Oncol 1984; 2:799.
54. Finklestein JZ, Klemperer MR, Evans A. et al: Multiagent chemotherapy for with metastatic neuroblastoma: A report from Children's Cancer Study Group. Med Pediatr Oncol 1979; 6:179.
55. Evans AE, Brand W1 de Lorimier A, et al: Results in children with local and regional neuroblastoma managed with and without vincristine, cyclophosphamide, imidazolecarboxamide. Am ] Clin Oncol 1984: 6:3.
56. Ninane J. Pritchard J. Morris Jones PH, et al: Stage U neuroblastoma. Adverse prognostic significance of lymph node involvement. Arch Dis Child 1982; 57:438.
57. Hayes FA, Green A, Hustu HO, et al: Surgicopathologic staging of neuroblastoma: Prognostic significance of regional lymph node metastases. J Pediatr 1983; 102:59.
58. De Bernardi B, Rogers D, Carli M, et al: Localized neuroblastoma. Cancer 1987; 60:1066.
59. Ninane J, Pritchard J, Malpas JS. et ah Chemotherapy of advanced neuroblastoma: Does adriamycin contribute? Arch Dis Child 1980; 56:544.
60. Thiele C], Reynolds CP, lrael MA. et al: Decreased expression of n-MYC precedes retinóte acid-induced morphological differentiation of human neuroblastoma. Nature (London) 1985; 313:404.
61. Hann H-WL, Evans AE, Siegel SE, et al: Prognostic importance of serum ferritin in patients with stages III and IV neuroblastoma: The Children's Cancer Study Group experience. Cancer Res 1985; 45:2843.
62. Shimada H, Chatten J, Newton WA, et al: Histopathologic aprognostic factors in neurohlastic tumors: Definition of subtypes of gangltoneurohlastoma and an agelinked classification of neuroblastomas. ) Nad Cancer Institute 1984; 73:405.
63. Look AT, Hayes FA, Nitschkc R, et al: Cellular DNA content as a predictor of response to chemotherapy in infants with unresectable neuroblastoma. N Engl } Med 1984; 311:231.
64. Evans AE, D'Angio GJ, Propert K, et al: Prognostic factors in neuroblastoma. Cancer 1987; 59:1853.
65. Hann H-WL, Evans AE, Cohen IJ, et al: Biologic differences between neuroblastoma stages IV-S and IV. N Engl) Med 1981; 305:425.
66. Nickerson HJ, Nesbit ME, Gtosfeld JL, et al: Comparison of stage IV and IVS neuroblastoma in the first year of life. Med Pediatr Oncol 1985; 13:261.
67. Sitarz A, Finklestein J, Grosfeld J, et al: An evaluation of the role of sutgery in disseminated neuroblastoma: A report from the Children's Cancer Study Group. ) Pediatr Surg 1983; 18:147
68. Grosfeld JL. Ballantine VN. Baehner RL, et al: Experience with "second lookoperations in pediatric solid tumors. ) Pediatr Sarg 1979; 13:275.
69. Jacobson GM, Sause WT, O'Brien RT, et al: Dose response analysis of pediatric neuroblastoma to megavoltage radiation. Ami Clin Oncol 1984; 7:693.
70. Deacon JM, Wilson PA, Peckham MJ, et al: The radiobiology of human neuroblastoma. Radiother Oncol 1985; 3:201.
71. Philip T, Helson L, Bernard J-L, et al: Definition of response and remission in children over one year of age with advanced neuroblastoma. Pediatr Hematol Oncol 1987; 4:25.
72. Hayes FA, Green AA, Mauer AM: Correlation of cell kinetic and clinical response to chemotherapy in disseminated neuroblastoma. Cancer Res 1977; 37:3766.
73. Helson L: The nude mouse model for rational therapy, in Evans AE (ed): Advances in Neuroblastoma Research. New York, Raven Press, 1980, ? 309.
74. Green AA, Hayes FA, Hustu O: Sequential cyclophosphamide and doxorubicin for induction of complete !emission in children with disseminated neuroblastoma. Cancer 1981; 48:2310.
75. Ftantz CN, Gelber RD. Belli JA, et al: Aggressive treatment of neuroblastoma, in Raybaud C, Clement R, Lebreuil G (eds): Pediatric Oncology. Amsterdam, Excerpta Medica. 1982, ? 175.
76. Shafford EA, Rogers DW, Pritchard J: Advanced neuroblastoma: Improved response rate using a multiagent regimen (OPEC) including sequential cisplatin and VM-26. ) Can Oncol 1984; 2:742.
77. Kushner BH, Helson L: Coordinated use of sequentially-escalated cyclophosphamide and eel I -cycle-spec i fie chemotherapy (the N4SE protocol) fot advanced neuroblastoma: Experience with 100 patients. ] CIm Oncol 1987; 5:1746.
78. Green AA, Hayes FA, Rao B. et al: Disease control and toxicity of aggressive 4 drug therapy for children with disseminated neuroblastoma (DNb). Abstracted, Proc Am Soc Clin Oncol 1986; 5:210.
79. Rosen EM, Cassady JR, Frantz CN. et al: Stage IV-N: A favorable subset of children with metastatic neuroblastoma. Med Pediatr Oncol 1985; 13:194.
80. Finklestein JZ, Klemperer MR, Evans AE, et al: Multiagent chemotherapy fot children with metastatic neuroblastoma: A report from Children's Cancer Study Group. Med Pediatr Oncol 1979; 6:179.
81. Pritchard J, McElwain TJ, Graham-Bale J: High-dose melphalan with autologous marrow for treatment of advanced neuroblastoma. Bri Cancer 1986; 45:86.
82. Hartmann O, Kalifa C, Benhamou E, et al: Treatment of advanced neuroblastoma with high-dose melphalan and autologous bone marrow transplantation. Cancer Chemother Pharmacol 1986; 16:165.
83. August CS, Serota FT, Penelope AK: Treatment of advanced neuroblastoma with supralethai chemotherapy, radiation, and allogeneic or autologous marrow reconstitution, i CIm Oncol 1984; 2:609.
84 Philip T, Bernard JL, Zucker JM, et al: High-dose chemoradiotherapy with bone marrow transplantation as consolidation treatment in neuroblastoma: An unselected group of stage IV patients over 1 year of age. J CIm Oncol 1987; 5:266.
85 Seeger R, Lenarsky C, Moss T. et al: Bone marrow transplantation (BMT) fot poor prognosis neuroblastoma. Abstracted, Proc Am Soc CIm Oncol 1987; 6:221.
86. Hartmann O, Benhamou E, Beaujean FC, et al: Repeated high-dose chemotherapy followed by purged autologous bone marrow transplantation as consolidation therapy in metastatic neuroblastoma. J CUn Oncol 1987; 5:1205.
87. Kushner BH, Cheung N-KV, Mandel] LR, er al: Myeloablaiive combination chemotherapy (without total body irradiation) for neuroblastoma. Abstracted, Proc Am Soc Clin Oncol, to be published.
88. D'Angio GJ, August C, Elkins W, et al: Metastatic neuroblastoma managed by supralethai therapy and bone marrow reconstitution (BMRc). Results of a fourinstitution Children's Cancer Study Group pilot study, in Evans AE, D'Angio GJ1 Seeger RC (eds): Advances m Neuroblastoma Research. New York, Alan R. Liss. 1985, ? 557.
89. Seeger RC, Reynolds CP. Dang Dai: Depletion of neuroblastoma cells from bone marrow with monoclonal antibodies and magnetic immunobeads, in Evans AE. D'AngioGJ. Seeger RC (eds): Advances in NeuroWasloma Research. New York. Alan R. Liss, 1985, ? 443.
90. Treleaven JG, Ugelstad J. Philip T. et al: Removal of neuroblastoma cells from bone marrow with monoclonal antibodies conjugated to magnetic microspheres. Lancet
91. Stein J, Strandjord, Saarinen U: In vitro treatment of autologous bone marrow for neuroblastoma patients with ami GD, monoclonal antibody and human complement: A pilot study, in Evans AE, D'Angio GJ, Seeger RC: Advances in NeuruMastiima Research. New York, Alan R. Liss, (in press).
92. Kemshead JT. Goldman A, Jones D, et al: Therapeutic application of radiolabeled monoclonal antibody UJI 3A in children with disseminated neuroblastoma - A phase I study, in Evans AE, D'Angio GJ, Seeget RC; (eds): Advances in Neuroblastoma Research. New York. Alan R. Liss, 1985, ? 533.
93. Cheung N-KV, Landmeier B, Neely J. et al: Complete tumor ablation with iodine '"radiolabeled disialogangliosidc GD. specific monoclonal antibody against human neuroblastoma xenografted in nude mice. ) Natl Cancer Institute 1986; 77:739.
94. Cheung N-KV. Miraldi FD: Iodine I 31 labele.1 GD, monoclonal antibody in the diagnosis and therapy of human neuroblastoma, in Evans AE. D'Angio GJ, Seeger RX (eds); Advances in Neuroblastoma Research. New York. Alan R. Liss, (in press).
95. Cheung N-KV, Lazarus H, Miraldi FD. et al: üanglioside GD2 specific monoclonal antibody 3F8: A phase 1 study in patients with neuroblastoma and malignant melanoma. ) Clin Oncol 1987; 5:1430.
96. Munn D, Cheung N-KV: Intetleukin-2 (IL-2) enhancement of monoclonal antibodymediated cellular cytotoxicity (ALXJC) against human melanomas. Cancer Res 1987; 47:6600.
97. Lopez AF, Williamson J, Gamble JR, et al: Recombinant human granulocyte-macrophage colony-stimulating factor stimulates in vitro mature human neutrophil and eosinophil function, surface receptor expression, and survival. ) Clin Invest 1986; 78:1220.
98. Klcbanoff SJ. Vadas MA, Harlan JM, et al: Stimulation of neuttophils by tumor necrosis factor I Immunol 1986; 1 36:4220.
Recommended Evaluations in Patients with Neuroblastoma
Advanced Neuroblastoma: Induction Regimens in Patients Over 1 Year of Age