Bevacizumab, Irinotecan, or Topotecan Added to Temozolomide for Children With Relapsed and Refractory Neuroblastoma: Results of the ITCC-SIOPEN BEACON-Neuroblastoma Trial

PURPOSE Outcomes for children with relapsed and refractory high-risk neuroblastoma (RR-HRNB) remain dismal. The BEACON Neuroblastoma trial (EudraCT 2012-000072-42) evaluated three backbone chemotherapy regimens and the addition of the antiangiogenic agent bevacizumab (B). MATERIALS AND METHODS Patients age 1-21 years with RR-HRNB with adequate organ function and performance status were randomly assigned in a 3 × 2 factorial design to temozolomide (T), irinotecan-temozolomide (IT), or topotecan-temozolomide (TTo) with or without B. The primary end point was best overall response (complete or partial) rate (ORR) during the first six courses, by RECIST or International Neuroblastoma Response Criteria for patients with measurable or evaluable disease, respectively. Safety, progression-free survival (PFS), and overall survival (OS) time were secondary end points. RESULTS One hundred sixty patients with RR-HRNB were included. For B random assignment (n = 160), the ORR was 26% (95% CI, 17 to 37) with B and 18% (95% CI, 10 to 28) without B (risk ratio [RR], 1.52 [95% CI, 0.83 to 2.77]; P = .17). Adjusted hazard ratio for PFS and OS were 0.89 (95% CI, 0.63 to 1.27) and 1.01 (95% CI, 0.70 to 1.45), respectively. For irinotecan ([I]; n = 121) and topotecan (n = 60) random assignments, RRs for ORR were 0.94 and 1.22, respectively. A potential interaction between I and B was identified. For patients in the bevacizumab-irinotecan-temozolomide (BIT) arm, the ORR was 23% (95% CI, 10 to 42), and the 1-year PFS estimate was 0.67 (95% CI, 0.47 to 0.80). CONCLUSION The addition of B met protocol-defined success criteria for ORR and appeared to improve PFS. Within this phase II trial, BIT showed signals of antitumor activity with acceptable tolerability. Future trials will confirm these results in the chemoimmunotherapy era.

and, importantly, lack of clarity regarding comparative benefits, with only one trial being randomized. 86][17][18][19][20][21][22] B is among the most established of antiangiogenic drugs and has gained approval in many adult indications.
The BEACON-Neuroblastoma trial opened in 2013 with two main aims: to identify the best backbone chemotherapy regimen on which to add new targeted therapies (T, IT, or topotecan-temozolomide [TTo]) and to establish the role of the addition of B in children with refractory/relapsed neuroblastoma within a randomized phase II trial.

Patients
Eligible patients were age 1 to 21 years with confirmed diagnosis of high-risk neuroblastoma.Disease status was relapsed or refractory to frontline therapy.Refractory disease was defined as lack of adequate response to frontline therapy using European Neuroblastoma Research Network (SIOPEN) criteria used at the time of enrollment: persistence of three or more spots on metaiodobenzylguanidine (I 123 MIBG) scan (with or without persistence of bone marrow disease).Measurable (as per RECIST 1.1 23 ) or evaluable disease (uptake on I 123 MIBG scan) was required.Previous treatment with T, I, or B was not allowed.Patients with only bone marrow detectable disease were not eligible.
Eligible patients had adequate performance status, organ and bone marrow function, and appropriate wash out periods for previous therapies.Full eligibility criteria are provided in the trial protocol, available as Data Supplement (online only).All patients, parents, and/or legal guardians provided written informed consent.Assent was sought from minors where appropriate.

BEACON-Neuroblastoma
(ITCC-032; ClinicalTrials.govidentifier: NCT02308527; EudraCT 2012-000072-42) was an open-label, randomized phase II trial.The trial was started with a 2 3 2 factorial design, in which patients were randomly assigned to two chemotherapy regimens (T alone or IT), with or without B (bevacizumab-temozolomide [BT] and bevacizumab-irinotecan-temozolomide [BIT]).The allocation ratio was 1:1:1:1.In 2015, after release of results of the Innovative Therapies for Children with Cancer (ITCC) phase II trial of TTo, 11 the trial was amended to evaluate a third chemotherapy regimen (TTo), and the design was changed to a 3 3 2 factorial trial (allocation ratio 1:1:1:1:1:1).Subsequently, when the accrual target for the IT arm was reached, the trial reverted to a 2 3 2 factorial design.Trial arms and details of chemotherapy doses are depicted in Figure 1 and Data Supplement (Table S1).The following minimization factors were used to balance all random assignments: (1) disease status (early relapse [defined as <18 months from the time of diagnosis] and late relapse [≥18 months] and refractory) and (2) measurable versus evaluable disease (as per RECIST v1.1).Interactions between treatment arms (ie, I and B) were not anticipated but were to be explored at trial completion using heterogeneity tests.
Trial treatment was given for six cycles.Patients with acceptable toxicities and response equal to or better than stable disease after six cycles were allowed to continue trial treatment for 12 cycles.

Outcome Measures and Assessments
The primary outcome measure was initially overall response (complete or partial response) during the first six courses of trial therapy, evaluated locally at trial sites.Response was evaluated using RECIST 1.1 for patients with measurable disease. 23For those with only evaluable disease, response was evaluated using the new draft International Neuroblastoma Response Criteria including data on MIBG scores and bone marrow disease, which were subsequently published. 24ter completing planned recruitment for the first 106 patients and meeting the trial-defined success criteria, the Independent Data Monitoring Committee (IDMC) recommended expansion of the B random assignment to 160 patients to include progression-free survival (PFS) as a coprimary outcome measure.
Secondary outcome measures were toxicity of the regimens (per Common Terminology Criteria Adverse Events v4.0), PFS, and overall survival (OS).
Response was evaluated at baseline and every two cycles using cross-sectional imaging with computed tomography/ magnetic resonance imaging and I 123 MIBG scans (or fluorodeoxyglucose-positron emission tomography scans if non-MIBG avid).Bilateral bone marrow aspirates and trephine biopsies were mandatory at baseline and repeated every two cycles in patients with bone marrow involvement at study entry or clinical suspicion of progression.All patients underwent mandatory imaging of the brain before study entry.

Trial Design and Statistical Analysis
Efficacy data were analyzed on an intention-to-treat basis, and safety data were reported for all patients who received at least one dose of a trial drug.Trial sample size was calculated to detect a 15% increase in response rate with B. Assuming a baseline overall response rate (ORR) of 25% without B, using a two-stage Minimax Jung design with a and b errors set to .2, 42 patients per arm were required for the first stage and 53 patients per arm were required in total.Predefined success criteria for the trial were ≥4 more responses in the B arm compared with the non-B arm.Allowing for patients with response not assessed, the planned sample size was n 5 120.Following a recommendation of the study IDMC, the B random assignment was then expanded to include PFS as Bevacizumab and Chemotherapy in Relapsed/Refractory Neuroblastoma coprimary end point: Assuming 40% PFS at 1 year in the control arm, with 160 patients and 80 events in total, there will be 80% power to detect a difference of 15% with a Cox proportional hazards model using a one-sided test with a 5 .15.All statistical analyses performed were predefined in the BEACON Statistical Analysis Plan.
For the primary end point of PFS for the expanded B random assignment, an adjusted hazard ratio (HR) with 95% CI was produced using a Cox proportional hazards model stratified for the other treatment allocation (I or topotecan or neither).The posterior probability distribution was then plotted for PFS, using a noninformative prior, on the basis of the HR calculated from the Cox model and the number of events observed.
For all random assignments, response was a primary endpoint.For response, risk ratios (RRs) with 95% CIs were calculated using logistic regression to compare responses (complete response/partial response) between arms for each random assignment.For the chemotherapy random assignments, Bayesian Beta-Binomial conjugate analyses were used to compare treatments by derived posterior distributions for the RR.Data in each arm are assumed to follow binomial distribution.A uniform prior b (1, 1) was used for the analysis.These plots are available to view in the Data Supplement (Figs S1 -S3).
Toxicity, PFS, and OS were secondary end points.Toxicity was analyzed descriptively by presenting grade ≥3 adverse events (AEs) in each random assignment and those occurring in ≥1% of patients.Adjusted HRs with CI for the treatment effect for both PFS and OS outcomes were produced by implementing a Cox proportional hazards regression model for each treatment group (ie, for B and non-B, IT v T, and TTo v T), stratified for other treatment allocation.Kaplan-Meier survival curves were then produced for PFS and OS and survival estimates at 6, 12, and 24 months, and median survival time with 95% CIs was calculated.Heterogeneity tests in the form of forest plots were used for the PFS and OS outcomes to assess any potential interaction between arms (Data Supplement, Fig S1

Screening and Random Assignment
Between July 8, 2013, and February 7, 2019, a total of 160 eligible patients were randomly assigned from 43 sites in 10 European countries (sites are detailed in Appendix Table A1, online only): T (36), IT (30), TTo (14), BT (34), BIT (30), and BTTo (16).One hundred sixty patients were included in the B random assignment, 121 in the I random assignment, and 61 in the topotecan random assignment.Figure 1

Potential Interactions
The trial design assumed no interaction between B and the different chemotherapy regimens.At trial completion, interactions were explored by heterogeneity tests.The unadjusted heterogeneity test for PFS showed some evidence of a potential interaction between irinotecan and B (P 5 .11;Data Supplement, Fig S1A1 ), but not for OS nor for topotecan (Data Supplement, Figs S1A2 and S1B-S1K).PFS and OS for all treatment arms are shown in Figures 2G and 2H.One-year and 2-year PFS estimates for patients in the BIT arm were 0.67 (95% CI, 0.47 to 0.80) and 0.50 (95% CI, 0.31 to 0.66), respectively.One-year and 2-year OS estimates for patients in the BIT arm were 0.77 (95% CI, 0.57 to 0.88) and 0.73 (95% CI, 0.54 to 0.86), respectively.No formal statistical comparisons were planned between individual trial arms.

Safety: B Random Assignment
Grade 3 or worse AEs are shown in Table 3 and the Data Supplement (Table S5).Patients receiving B experienced more AEs than those not receiving it.More patients in the B-receiving arms experienced neutropenia, anemia, and thrombocytopenia.Ten patients receiving B experienced proteinuria.Grade ≥3 proteinuria occurred in four patients (5%) receiving B and no patients receiving chemotherapy alone.No patients had episodes of grade ≥3 bleeding, wound healing complications, fistulae, posterior reversible encephalopathy syndrome, congestive heart failure, thromboembolic events, or GI perforation.

Safety: Chemotherapy Random Assignments
Patients receiving irinotecan or topotecan experienced more AEs than those receiving T alone (Data Supplement, Table S6).Patients receiving irinotecan had an increased incidence of ≥grade 3 GI AEs, mainly diarrhea in eight patients (13%).Patients receiving topotecan had an increased incidence of grade ≥3 neutropenia, thrombocytopenia, and anemia.

DISCUSSION
To our knowledge, we report the largest randomized trial in relapsed/refractory neuroblastoma conducted to date.Using a multiarm multistage factorial design, three chemotherapy regimens and a novel agent, B, added to conventional chemotherapy, were evaluated within one international cooperative group academic clinical trial.
][11] Identifying the most appropriate backbone chemotherapy regimen is critical to evaluate novel combinations for future therapy, and only a randomized controlled comparison can provide the necessary data.
The addition of irinotecan or topotecan to T did not achieve higher response rates but appeared to result in improved PFS.One-year PFS seemed higher in irinotecan-or topotecan-containing arms (23% and 24% higher, respectively) compared with T-only arms.Toxicity profiles confirmed that the addition of irinotecan is associated with more diarrhea and GI symptoms, whereas the addition of topotecan is associated with increased myelotoxicity.Whether one or the other toxicity profile is preferable remains a subject of debate and may vary from patient to patient.The trial did not include quality-of-life questionnaires or patient-reported outcome measures, which would have been extremely useful to base future choices incorporating the patients' and parents' voices.In the B random assignment of the BEACON trial, success criteria were met for response.PFS seemed to improve with the addition of B within the limitations of evidence generated within a randomized phase II trial.The addition of B also increased toxicity, mainly related to increased myelotoxicity.Consistent with other large trials in the pediatric population, B-specific toxicities seen in adults were rare, although grade 3-4 proteinuria occurred in 5% of patients. 25,261][32] The benefits of randomized comparisons are highlighted as the results of the BEACON trial differ from a single-institution single-arm trial which showed no benefit of adding B to IT in patients with highly pretreated neuroblastoma. 17e unexpected potential interaction identified after completing the trial highlights one of the challenges of using factorial designs.The strength of this design was that it allowed four different questions to be addressed within the same trial population and generated randomized phase II data.
The current results need to be analyzed in the context of recently reported trials in the relapsed/refractory neuroblastoma setting.][35] In the Children's Oncology Group ANBL1221 trial, 53 children were treated with IT-dinutuximab-granulocytemacrophage colony-stimulating factor, with an ORR of Infections and infestations 20 Catheter-related infection 7 Abdominal infection 3 (2.5)0 (0) 3 Lung infection 1 Upper respiratory infection 0 (0) 2 (2.5) 2 General disorders and administration site conditions 7 (5) 14 ( 5 41.5% and 1-year PFS of 67.9%.After completing the B random assignment, the BEACON trial was amended to evaluate the addition of anti-GD2 therapy with dinutuximab beta to chemotherapy. Despite promising results, a significant proportion of patients still relapsed.Better understanding of the biology of refractory/relapsed and identifying biomarkers of response or resistance to therapy remain crucial.An ambitious ongoing translational research program (BEACON BIO) is currently analyzing more than 700 plasma, blood, and tumor samples collected during trial treatment. 36 conclusion, within the limits of a phase II trial, the BEACON trial provided some evidence that addition of B to T-based chemotherapy improved ORR.It also appeared to show some improvement PFS when adding irinotecan or topotecan to T. A potential interaction between irinotecan and B might explain better results for the BIT regimen.A future study in relapsed/refractory neuroblastoma will compare IT-B with a chemoimmunotherapy regimen.

15 NHS 22 Centre
Greater Glasgow and Clyde, Glasgow, United Kingdom 16 Birmingham Women's and Children's NHS Foundation Trust, Birmingham, United Kingdom 17 The Royal Marsden NHS Foundation Trust & Institute for Cancer Research, London, United Kingdom 18 Hôpital M ère-Enfant, Nantes, France 19 The Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle, United Kingdom 20 Great Ormond Street Hospital, London, United Kingdom 21 Hôpital des Enfants, H émato-Oncologie, Bordeaux, France ). Clinical trial protocol and statistical analysis plan are available as Data Supplement.the data and the fidelity of the trial to the protocol.The trial was registered before the first patient recruited on April 24, 2013 (ISRTCN40708286; Eudract2012-000072-42; ClinicalTrials.govidentifier: NCT02308527).
competent authority and ethics committee approval in all the countries where it opened.The trial was designed by the Trial Management Group and the sponsor, the University of Birmingham, in collaboration with investigators from the ITCC and SIOPEN cooperative groups.Roche provided B supply and distribution.The authors attest to the accuracy and completeness of depicts the trial CONSORT diagram.
(69%) had measurable and 49 (31%) evaluable disease at study entry; 36 (23%) had tumors with MYCN amplification, and 64 (40%) had prior anti-GD2 immunotherapy.The median follow-up for survivors was 1.5 years (IQR, 0.7-3.2years;DataSupplement,FigS2).Sixteen (10%) patients had response not evaluable and were considered nonresponders.Of them, six patients did not receive study treatment: one was ineligible post-random assignment, one experienced an AE, three withdrew from the study, and one patient progressed before treatment started.For the 10 patients who received study treatment, one patient had an unexpected change in diagnosis at cycle 2, six patients progressed before cycle 2 and one patient at cycle 2, and two patients died before cycle 2.For the present analysis, all 16 patients with response not assessed are considered nonresponders.B Random Assignment (T, IT, TTo v BT, BIT, BTTo) (95% CI, 0.27 to 0.49) in the non-B arms.PFS and OS for all random assignments and arms are shown in Figure2.The HR for OS was 1.01 (95% CI, 0.70 to 1.45).The 1-year OS was 0.69 (95% CI, 0.57 to 0.77) for B arms compared with 0.58 (95% CI, 0.47 to 0.68) for non-B arms.

TABLE 1 .
Patient Characteristics for the B RandomizationThe posterior plot for PFS for the B random assignment derived from the HR, number of events, and a noninformative prior b (0.1, 0.1) is shown in the Data Supplement (FigS3).The probability, given the data, of the true HR being <1, that is, in favor of B treatment, is 73%.
Abbreviations: B, bevacizumab; COJEC, cisplatin, vincristine, carboplatin, etoposide, and cyclophosphamide; ECOG, Eastern Cooperative Oncology Group; INSS, International Neuroblastoma Staging System; I 123 -mIBG, metaiodobenzylguanidine; NA, not applicable; SIOPEN, European Neuroblastoma Research Network; TVD, topotecan, vincristine, doxorubicin.aRefractorydiseasewas defined using the SIOPEN definition at the time of trial conduct: persistence of three or more spots on I 123 .MIBG scan (with or without persistence of bone marrow disease).Early relapse was defined as those occurring <18 months from the time of diagnosis and late relapse those occurring ≥18 months from the time of diagnosis.A posterior distribution for the RR was derived by performing simulations on the number of responses and noninformative b (0.1, 0.1) prior for the I random assignment as shown in the Data Supplement (FigS3).An RR of 0.939 (95% credible interval [CrI], 0.46 to 1.90) was obtained.A normal distribution curve is shown for comparison.The posterior probability of true RR >1.0, 1.2, 1.4, and 1.6 is presented.

TABLE 3 .
Grade ≥3 AEs in the B Random Assignment Occurring in ≥1% of Patients