Current and Future Strategies for Treatment of Relapsed Neuroblastoma

Introduction

Despite recent advances in the treatment of high-risk neuroblastoma (HR-NBL), such as the addition of immunotherapy to upfront intensive multimodal therapy , and consequent incremental improvements in outcome, only approximately half of patients will achieve a long-term cure ( Fig. 15.1 ). Approximately 10%–20% of HR-NBL patients, particularly adolescents whose tumors more commonly have loss-of-function mutations in the alpha thalassemia/mental retardation syndrome X-linked ( ATRX ) gene , have primary refractory disease and will not achieve remission. Of the remainder, 50%–60% will complete initial therapy only subsequently to relapse. Historically, outcomes for patients with relapsed HR-NBL have been very poor, with the European Neuroblastoma Study Group reporting 5-yr post-relapse overall survival (PROS) of only 5.6% for patients diagnosed 1982–1992 . Data from the International Neuroblastoma Risk Group (INRG) database similarly revealed 5-yr PROS of 8% for patients with metastatic disease at original diagnosis and only 4% for those with MYCN-amplified stage 4 disease . More recent data for patients with recurrent HR-NBL treated on a range of early phase trials suggests a potential for improved outcomes, with 20% 4-yr PROS and 6% 4-yr post-relapse progression-free survival . More encouraging outcomes have been reported for the selected subset of patients with recurrent HR-NBL in whom it is possible to achieve a second complete response (CR) or very good partial response (VGPR) and then consolidate with anti-GD2-based immunotherapy. With this approach, 33% 5-yr progression free survival (PFS) and 48% 5-yr overall survival (OS) have been reported .

By far the majority of patients with relapsed neuroblastoma will have had high-risk disease at initial diagnosis and the remainder of this chapter is focused on the treatment strategies for these patients. For the remainder who initially had low or intermediate-risk treatment, therapeutic approaches at relapse are in general more straightforward, and standard chemotherapy approaches can be used depending on individual circumstances. For example, patients who initially had localized disease treated with primary resection alone may be cured with minimal chemotherapy and reresection in the event of localized recurrence . Patients who were initially treated for intermediate risk disease will usually be treated with reinduction chemotherapy with the aim of consolidating the response with high-dose chemotherapy and stem cell rescue, followed by immunotherapy; essentially a similar approach to upfront high-risk therapy.

Management decisions for patients with relapsed HR-NBL are potentially complex as a result of several important considerations. First, at the patient level, quality of life, burden of therapy, disease-related symptoms, and toxicities from prior therapy, all need to be taken into account; particularly since there are as yet no proven curative therapies for relapsed HR-NBL. Second, although a wide range of therapeutic approaches has been developed for relapsed HR-NBL, only rarely have these been tested in a randomized fashion, and therefore, the benefits of one approach over another, or the relevance of a particular strategy based on clinical or biological features, is typically unknown. Third, in addition to cytotoxic chemotherapy, there is a growing interest in precision medicine approaches raising important questions about rebiopsy of disease at relapse (even when the diagnosis itself is not in doubt) to obtain tissue for genetic analyses, and about how targeted agents are incorporated into the overall therapeutic approach.

Standard Chemotherapy Approaches

For the majority of patients with relapsed HR-NBL, initial treatment will comprise reinduction chemotherapy typically based around combinations of topotecan or irinotecan, with temozolomide or cyclophosphamide. The relative efficacy of these combinations (summarized in Table 15.1 ) is difficult to ascertain since the majority of published studies have been single-arm Phase II trials, with no comparison of treatment strategies, endpoints of response rates (rather than survival) and heterogeneous populations in terms of the extent of disease at relapse (such as measurable soft-tissue lesions vs. evaluable metaiodobenzylguanidine (MIBG)-avid skeletal disease or bone marrow only disease). The combination of topotecan and cyclophosphamide (TopoCy) is generally well-tolerated and effective in up to ∼1/3 of patients; at least in terms of disease response. Although now incorporated into the upfront schedule of the standard HR-NBL approach in North America (Children's Oncology Group), evidence suggests that retreatment (at a lower dose than used in initial induction) can lead to durable responses in the relapse setting . Higher doses of TopoCy combined with vincristine have also shown significant responses in the setting of relapsed HR-NBL (overall response rate, ORR, 52%), as well as for patients with primary refractory HR-NBL (ORR 19%) .

Other frequently used chemotherapy approaches incorporate temozolomide, either as monotherapy or in combination with irinotecan or topotecan . Response rates of 8%–24% have been reported, but the absence of randomized trial data makes it difficult to justify the choice of one or other regimen solely on the grounds of predicted disease response. The current European BEACON-Neuroblastoma study ( NCT02308527 ) is testing these three regimens in a randomized Phase II trial that will also test the potential benefit of adding bevacizumab to backbone chemotherapy.

More intensive and potentially toxic chemotherapy regimens have also been utilized in the relapse setting, with Memorial Sloan Kettering Cancer Center (MSKCC) pioneering the use of high-dose ifosfamide, carboplatin, and etoposide (HD-ICE). Objective responses were reported in nine of 17 (53%) of patients with newly relapsed disease, although response rates for patients with primary refractory or progressive disease were lower . HD-ICE is associated with significant myelotoxicity, requiring preemptive stem cell rescue in patients predicted to have a poor hematological reserve. Importantly, HR-NBL patients treated at MSKCC typically do not receive high-dose chemotherapy (HDC) and stem cell rescue as part of their upfront therapy ; therefore the utility and risk/benefit balance of the HD-ICE approach for patients who have previously received HDC remains uncertain.

An alternative chemotherapy approach is to move from intensification towards low-dose metronomic chemotherapy approaches designed to elicit antitumor effects through antiangiogenesis. The combination of celecoxib, cyclophosphamide, vinblastine, and etoposide has been reported to have equivalent outcomes to standard chemotherapy for patients with relapsed HR-NBL , although this was a nonrandomized, retrospective study.

Novel Chemotherapy Combinations

In addition to the chemotherapy trials described above many trials are now incorporating chemotherapy backbones together with targeted therapies (see below, Precision Medicine Approaches section), immunotherapies or other noncytotoxic agents (e.g., angiogenesis inhibitors). These approaches vary from single-arm studies incorporating multiple agents to randomized trials including pick-the-winner and drop-the-loser designs to compare two or more novel agents on a backbone chemotherapy regimen. The RIST trial ( NCT01467986 ) is based on a regimen used for glioblastoma and includes 5 days of metronomic rapamycin (R) and dasatinib (Sprycel) followed by 5 days of irinotecan and temozolomide (I/T) . The early results demonstrated that of the first 21 patients 90% had an initial response based on imaging criteria with CR (57%), PR (14%) and SD (19%). The median progression-free survival (PFS) was 90 weeks. The ITCC has an ongoing randomized trial (BEACON Neuroblastoma ) (EudraCT 2012-000072-42) which will enroll 160 patients with relapsed and refractory neuroblastoma. The current version of the protocol includes six arms testing three backbone chemotherapy combinations (temozolomide alone, temozolomide plus irinotecan, and temozolomide plus topotecan) each with or without the addition of bevacizumab (Avastin).

Targeted Molecular Radiotherapy

As a result of their origination from sympathetic neuronal precursors, approximately 90% of neuroblastoma tumors express norepinephrine receptors (NET) and therefore readily take up the norepinephrine analog, meta-iodobenzylguanadine (MIBG) . In its radiolabeled form, this molecule (typically ¹²³ I-MIBG) is frequently used for evaluation of neuroblastoma metastatic sites and as ¹³¹ I-MIBG, which produces short-range β emission, has also been successfully used for targeted radiotherapy . Doses up to 12 mCi/kg can be administered without excessive myelotoxicity, but enhanced efficacy is seen with higher doses of 18 mCi/kg, following which autologous stem cell rescue is typically required. In the largest multicenter study to date, 164 patients with refractory, relapsed or progressive HR-NBL were treated with 12 or 18 mCi/kg ¹³¹ I-MIBG (depending on stem cell availability) . The overall response rate was 36%, with another 34% having stable disease (median 6.2 months), making ¹³¹ I-MIBG one of the most active agents in relapsed HR-NBL. Outcomes were significantly better in patients who had received fewer prior treatment regimens, in patients older than 12 years and in those with disease limited to bone and bone marrow, or to soft tissue (rather than combined). This overall response rate to ¹³¹ I-MIBG therapy of approximately a third appears to be robust and has been confirmed in a meta-analysis incorporating data from 782 patients across 27 trials .

In addition to single-agent ¹³¹ I-MIBG, combinations with potential radiosensitizers such as topotecan , irinotecan/vincristine and arsenic trioxide have been explored. The combination of ¹³¹ I-MIBG and histone deacetylase inhibitor vorinostat, which potentially acts both as a radiosensitizer and to upregulate NET expression, has also been successfully tested . As yet, however, although these combinations are feasible, it is not known whether they provide enhanced efficacy over single-agent ¹³¹ I-MIBG. A randomized trial of ¹³¹ I-MIBG alone, versus combinations with vorinostat or irinotecan/vincristine, is currently ongoing (NCT02035137) and may help to clarify this issue.

The optimal disease setting for the use of ¹³¹ I-MIBG therapy also is yet to be determined, although it is probably most effective in the setting of relatively indolent disease (as typically seen in adolescent or young adult patients) or for consolidation of response after initial reinduction chemotherapy. Major challenges exist in relation to the practicalities and radioprotection requirements of administering radiopharmaceuticals to young children particularly if they require sedation , although more facilities are being built and in some instances, regulations may permit ¹³¹ I-MIBG administration without specialized lead-lined rooms . In the future, ¹³¹ I-MIBG may become incorporated into upfront therapy for HR-NBL, as already the case in the Netherlands . It is currently being tested in a randomized fashion in the current Children's Oncology Group HR-NBL trial (ANBL1531, NCT03126916 ).

Beyond ¹³¹ I-MIBG, other forms of targeted radiotherapy for neuroblastoma are also being explored. MSKCC has developed a successful strategy for the treatment of central nervous system (CNS) relapsed neuroblastoma that incorporates compartmental radioimmunotherapy (cRIT) in the form of intrathecal ¹³¹ I–8H9, a monoclonal antibody against B7–H3 . Recent data indicate the potential for long-term survival after CNS relapse, with 45% alive at 3 years and 29% at 5 years . Another targeted radioisotope being explored for systemic disease is ¹⁷⁷ Lu-DOTATATE, based on the somatostatin analog octreotate and taken up by somatostatin receptor 2 (SSR2) expressed by many neuroblastomas .

Precision Medicine Approaches

Precision medicine approaches in which the choice of particular treatments and eligibility for trials are based on the specific molecular or genetic characteristics of a tumor are being used increasingly in the pediatric setting , including for patients with relapsed neuroblastoma. Precision medicine relies on targeting genetic and molecular alterations in patient tumors that have properties suggesting they are drivers and thus, responsible for providing selective growth advantages for the tumor, but not normal, cells. In contrast to more standard clinical trials, assignment (and the associated eligibility requirements) to a particular targeted therapy is based on the demonstration of molecular or genetic alterations in the tumor (often at the time of relapse) and are typically less dependent on specific histologic diagnosis. The most common alterations that represent tractable therapeutic targets are genomic alterations of kinases, including fusions, amplifications/overexpression, and activating missense mutations. Large studies in which pediatric patient tumors were profiled using next-generation sequencing (NGS)-based approaches (including neuroblastoma) suggest that 30%–45% of tumors may have potentially “actionable” alterations ; however, the definition of “actionable” varies among studies and access to targeted agents for pediatric patients may also be limited. A recent report from the Children's Oncology Group (COG) reviewed the evidence for target-agent pairs in pediatric tumors and Moreno and colleagues specifically reviewed the evidence and prioritization for current and future targets in relapsed neuroblastoma . However, it is important to recognize that given the rapid advances these definitions of “actionability” for particular genomic alterations and the agents predicted to target them are constantly evolving.

Historically, eligibility for the majority of early phase trials for relapsed neuroblastoma has been determined by the presence of relapsed or refractory neuroblastoma, usually defined as lesions that are detected on cross-sectional imaging and/or MIBG scans in a patient with a previous history of histologically proven neuroblastoma. However, patients with relapsed neuroblastoma show heterogeneous clinical behavior likely due to differences in genetic and molecular profiles. Recent next generation sequencing (NGS) approaches demonstrate that at diagnosis, and increasingly at relapse, neuroblastoma tumors harbor heterogeneous genomic alterations . The most common copy number alteration at diagnosis is the amplification of the MYCN oncogene (20%), and the most frequent mutations are missense alterations of the receptor tyrosine kinase anaplastic lymphoma kinase or ALK oncogene (8%–10%). In addition, there is emerging evidence that at the time of relapse there are additional changes that often include a higher number of mutations (tumor mutation burden) as well as clonal selection leading to increased allele frequency for certain mutations such as those involving ALK . The most common missense mutations at recurrence involve alterations of the RAS/MAPK pathway, including but not limited to RAS (NRAS, HRAS, KRAS), PTPN11, NF1 and activating ALK variants. These findings and similar results in other tumor types have provided support for obtaining new tissue for NGS studies at the time of recurrence. Most of the studies to date have focused on DNA sequencing, but early RNA and whole genome sequencing efforts have identified other potentially targetable alterations including fusions of TERT .

There are currently two types of precision medicine clinical trials for which patients with relapsed neuroblastoma are eligible: umbrella and basket trials. The general concepts for these trials were recently reviewed by Woodcock and Lavange . Umbrella trials evaluate different regimens for groups of molecularly defined tumors. Patients with a particular tumor type are screened for genomic or other molecular biomarkers that determine eligibility to a specific arm or strata with a targeted agent (or combination of agents). These trials have been used commonly in breast and lung tumors. Currently, there is only one umbrella trial for patients with neuroblastoma: NEPENTHE (Next generation personalized neuroblastoma treatment, NCT02780128 ). Patients with recurrent neuroblastoma undergo a biopsy, and the genomic profiling results are used to assign patients to one of the two arms based on the presence of specific alterations (ALK, RAS/MAPK) or a third arm for patients who do not have one of the eligible biomarker findings. In contrast to umbrella trials, basket studies generally involve patients with multiple histologic diseases. Tumors are profiled, and the patients enrolled based on specific genomic alterations predefined as the target biomarkers. Patients are then assigned to specific treatment arms or sub-protocols based on which targets are identified. Examples of basket trials in pediatrics for which neuroblastoma patients are eligible include in the US the Children's Oncology Group (COG) MATCH (NCT0315562) and in Europe the eSMART (reviewed in Ref. ) and INFORM trials. In addition to these basket and umbrella trials which provide profiling and drug access there are many local institutional and consortium trials as well as commercial laboratories that provide access to genomic profiling ranging from specific gene panels that identify alterations in a limited number of genes to more large-scale whole genome and RNA sequencing-based technologies that may detect novel alterations not previously identified in neuroblastoma. Findings from these profiles can then be used to determine eligibility for neuroblastoma patients to early phase trials of targeted therapies and/or off-label use of agents available for other diagnoses.

In the remaining parts of this section, we discuss drugs and trials that target genomic and molecular alterations detected in neuroblastoma. We will focus on pharmacologic agents (and their respective targets) in early phase clinical trials and others in development with the highest level of preclinical evidence.

Genetic and Molecular Targeted Agents