Classification with a disordered dopant-atom network in silicon.

Classification is an important task at which both biological and artificial neural networks excel1,2. In machine learning, nonlinear projection into a high-dimensional feature space can make data linearly separable3,4, simplifying the classification of complex features. Such nonlinear projections are computationally expensive in conventional computers. A promising approach is to exploit physical materials systems that perform this nonlinear projection intrinsically, because of their high computational density5, inherent parallelism and energy efficiency6,7. However, existing approaches either rely on the systems' time dynamics, which requires sequential data processing and therefore hinders parallel computation5,6,8, or employ large materials systems that are difficult to scale up7. Here we use a parallel, nanoscale approach inspired by filters in the brain1 and artificial neural networks2 to perform nonlinear classification and feature extraction. We exploit the nonlinearity of hopping conduction9-11 through an electrically tunable network of boron dopant atoms in silicon, reconfiguring the network through artificial evolution to realize different computational functions. We first solve the canonical two-input binary classification problem, realizing all Boolean logic gates12 up to room temperature, demonstrating nonlinear classification with the nanomaterial system. We then evolve our dopant network to realize feature filters2 that can perform four-input binary classification on the Modified National Institute of Standards and Technology handwritten digit database. Implementation of our material-based filters substantially improves the classification accuracy over that of a linear classifier directly applied to the original data13. Our results establish a paradigm of silicon-based electronics for small-footprint and energy-efficient computation14.

A Fast, Affordable, and Minimally Invasive Diagnostic Test for Cancer of Unknown Primary Using DNA Methylation Profiling.

Currently, we cannot provide a conclusive diagnosis for 3% to 5% of people who are confronted with cancer. These patients have cancer of unknown primary (CUP), ie, a metastasized cancer for which the tissue of origin cannot be determined. Studies have shown that the DNA methylation profile is a unique "fingerprint" that can be used to classify tumors. Here we used cell-free reduced representation bisulfite sequencing (cfRRBS), a technique that allows us to identify the methylation profile starting from minimal amounts of highly fragmented DNA, for CUP diagnosis on formalin-fixed paraffin-embedded (FFPE) tissue and liquid biopsies. We collected 80 primary tumor FFPE samples covering 16 tumor entities together with 15 healthy plasma samples to use as a custom cfRRBS reference data set. Entity-specific methylation regions are defined for each entity to build a classifier based on nonnegative least squares deconvolution. This classification framework was tested on 30 FFPE, 19 plasma, and 40 pleural and peritoneal effusion samples of both known metastatic tumors and clinical CUPs for which pathological investigation finally resulted in a cancer diagnosis. Using this framework, 27 of 30 FFPE (all CUPs) and 16 of 19 plasma samples (10/13 CUPs) obtained an accurate diagnosis, with a minimal DNA input of 400 pg. Diagnosis of the 40 pleural and peritoneal effusion samples is possible in 9 of 27 samples with negative/inconclusive cytology (6/13 CUPs), showing that cell-free DNA (cfDNA) methylation profiling could complement routine cytologic analysis. However, a low "cfDNA - high-molecular weight DNA ratio" has a considerable impact on the prediction accuracy. Moreover, the accuracy improves significantly if the predicted tumor percentage is >7%. This proof-of-concept study shows the feasibility of using DNA methylation profiling on FFPE and liquid biopsy samples such as blood, ascites, and pleural effusions in a fast and affordable way. Our novel RRBS-based technique requires minimal DNA input, can be performed in <1 week, and is highly adaptable to specific diagnostic problems as we only use 5 FFPE references per tumor entity. We believe that cfRRBS methylation profiling could be a valuable addition to the pathologist's toolbox in the diagnosis of CUPs.

Digital PCR-based evaluation of nucleic acid extraction kit performance for the co-purification of cell-free DNA and RNA.

BACKGROUND: Blood plasma, one of the most studied liquid biopsies, contains various molecules that have biomarker potential for cancer detection, including cell-free DNA (cfDNA) and cell-free RNA (cfRNA). As the vast majority of cell-free nucleic acids in circulation are non-cancerous, a laboratory workflow with a high detection sensitivity of tumor-derived nucleic acids is a prerequisite for precision oncology. One way to meet this requirement is by the combined analysis of cfDNA and cfRNA from the same liquid biopsy sample. So far, no study has systematically compared the performance of cfDNA and cfRNA co-purification to increase sensitivity. RESULTS: First, we set up a framework using digital PCR (dPCR) technology to quantify cfDNA and cfRNA from human blood plasma in order to compare cfDNA/cfRNA co-purification kit performance. To that end, we optimized two dPCR duplex assays, designed to quantify both cfDNA and cfRNA with the same assays, by ensuring that primers and probes are located within a highly abundant exon. Next, we applied our optimized workflow to evaluate the co-purification performance of two manual and two semi-automated methods over a range of plasma input volumes (0.06-4 mL). Some kits result in higher nucleic acid concentrations in the eluate, while consuming only half of the plasma volume. The combined nucleic acid quantification systematically results in higher nucleic acid concentrations as compared to a parallel quantification of cfDNA and cfRNA in the eluate. CONCLUSIONS: We provide a framework to evaluate the performance of cfDNA/cfRNA co-purification kits and have tested two manual and two semi-automated co-purification kits in function of the available plasma input amount and the intended use of the nucleic acid eluate. We demonstrate that the combined quantification of cfDNA and cfRNA has a benefit compared to separate quantification. We foresee that the results of this study are instrumental for clinical applications to help increase mutation detection sensitivity, allowing improved disease detection and monitoring.

Whole genome sequencing and inheritance-based variant filtering as a tool for unraveling missing heritability in pediatric cancer.

Survival rates for pediatric cancer have significantly increased the past decades, now exceeding 70-80% for most cancer types. The cause of cancer in children and adolescents remains largely unknown and a genetic susceptibility is considered in up to 10% of the cases, but most likely this is an underestimation. Families with multiple pediatric cancer patients are rare and strongly suggestive for an underlying predisposition to cancer. The absence of identifiable mutations in known cancer predisposing genes in such families could indicate undiscovered heritability. To discover candidate susceptibility variants, whole genome sequencing was performed on germline DNA of a family with two children affected by Burkitt lymphoma. Using an inheritance-based filtering approach, 18 correctly segregating coding variants were prioritized without a biased focus on specific genes or variants. Two variants in FAT4 and DCHS2 were highlighted, both involved in the Hippo signaling pathway, which controls tissue growth and stem cell activity. Similarly, a set of nine non-coding variants was prioritized, which might contribute, in differing degrees, to the increased cancer risk within this family. In conclusion, inheritance-based whole genome sequencing in selected families or cases is a valuable approach to prioritize variants and, thus, to further unravel genetic predisposition in childhood cancer.

Cellular senescence in neuroblastoma.

Neuroblastoma is a tumour that arises from the sympathoadrenal lineage occurring predominantly in children younger than five years. About half of the patients are diagnosed with high-risk tumours and undergo intensive multi-modal therapy. The success rate of current treatments for high-risk neuroblastoma is disappointingly low and survivors suffer from multiple therapy-related long-term side effects. Most chemotherapeutics drive cancer cells towards cell death or senescence. Senescence has long been considered to represent a terminal non-proliferative state and therefore an effective barrier against tumorigenesis. This dogma, however, has been challenged by recent observations that infer a much more dynamic and reversible nature for this process, which may have implications for the efficacy of therapy-induced senescence-oriented treatment strategies. Neuroblastoma cells in a dormant, senescent-like state may escape therapy, whilst their senescence-associated secretome may promote inflammation and invasiveness, potentially fostering relapse. Conversely, due to its distinct molecular identity, senescence may also represent an opportunity for the development of novel (combination) therapies. However, the limited knowledge on the molecular dynamics and diversity of senescence signatures demands appropriate models to study this process in detail. This review summarises the molecular knowledge about cellular senescence in neuroblastoma and investigates current and future options towards therapeutic exploration.

Targeted therapy in juvenile myelomonocytic leukemia: Where are we now?

Juvenile myelomonocytic leukemia (JMML) is a rare and aggressive clonal neoplasm of early childhood, classified as an overlap myeloproliferative/myelodysplastic neoplasm by the World Health Organization. In 90% of the patients with JMML, typical initiating mutations in the canonical Ras pathway genes NF1, PTPN11, NRAS, KRAS, and CBL can be identified. Hematopoietic stem cell transplantation (HSCT) currently is the established standard of care in most patients, although long-term survival is still only 50-60%. Given the limited therapeutic options and the important morbidity and mortality associated with HSCT, new therapeutic approaches are urgently needed. Hyperactivation of the Ras pathway as disease mechanism in JMML lends itself to the use of targeted therapy. Targeted therapy could play an important role in the future treatment of patients with JMML. This review presents a comprehensive overview of targeted therapies already developed and evaluated in vitro and in vivo in patients with JMML.

mRNA Capture Sequencing and RT-qPCR for the Detection of Pathognomonic, Novel, and Secondary Fusion Transcripts in FFPE Tissue: A Sarcoma Showcase.

We assess the performance of mRNA capture sequencing to identify fusion transcripts in FFPE tissue of different sarcoma types, followed by RT-qPCR confirmation. To validate our workflow, six positive control tumors with a specific chromosomal rearrangement were analyzed using the TruSight RNA Pan-Cancer Panel. Fusion transcript calling by FusionCatcher confirmed these aberrations and enabled the identification of both fusion gene partners and breakpoints. Next, whole-transcriptome TruSeq RNA Exome sequencing was applied to 17 fusion gene-negative alveolar rhabdomyosarcoma (ARMS) or undifferentiated round cell sarcoma (URCS) tumors, for whom fluorescence in situ hybridization (FISH) did not identify the classical pathognomonic rearrangements. For six patients, a pathognomonic fusion transcript was readily detected, i.e., PAX3-FOXO1 in two ARMS patients, and EWSR1-FLI1, EWSR1-ERG, or EWSR1-NFATC2 in four URCS patients. For the 11 remaining patients, 11 newly identified fusion transcripts were confirmed by RT-qPCR, including COPS3-TOM1L2, NCOA1-DTNB, WWTR1-LINC01986, PLAA-MOB3B, AP1B1-CHEK2, and BRD4-LEUTX fusion transcripts in ARMS patients. Additionally, recurrently detected secondary fusion transcripts in patients diagnosed with EWSR1-NFATC2-positive sarcoma were confirmed (COPS4-TBC1D9, PICALM-SYTL2, SMG6-VPS53, and UBE2F-ALS2). In conclusion, this study shows that mRNA capture sequencing enhances the detection rate of pathognomonic fusions and enables the identification of novel and secondary fusion transcripts in sarcomas.

Automated Assessment of COVID-19 Reporting and Data System and Chest CT Severity Scores in Patients Suspected of Having COVID-19 Using Artificial Intelligence.

Background The coronavirus disease 2019 (COVID-19) pandemic has spread across the globe with alarming speed, morbidity, and mortality. Immediate triage of patients with chest infections suspected to be caused by COVID-19 using chest CT may be of assistance when results from definitive viral testing are delayed. Purpose To develop and validate an artificial intelligence (AI) system to score the likelihood and extent of pulmonary COVID-19 on chest CT scans using the COVID-19 Reporting and Data System (CO-RADS) and CT severity scoring systems. Materials and Methods The CO-RADS AI system consists of three deep-learning algorithms that automatically segment the five pulmonary lobes, assign a CO-RADS score for the suspicion of COVID-19, and assign a CT severity score for the degree of parenchymal involvement per lobe. This study retrospectively included patients who underwent a nonenhanced chest CT examination because of clinical suspicion of COVID-19 at two medical centers. The system was trained, validated, and tested with data from one of the centers. Data from the second center served as an external test set. Diagnostic performance and agreement with scores assigned by eight independent observers were measured using receiver operating characteristic analysis, linearly weighted κ values, and classification accuracy. Results A total of 105 patients (mean age, 62 years ± 16 [standard deviation]; 61 men) and 262 patients (mean age, 64 years ± 16; 154 men) were evaluated in the internal and external test sets, respectively. The system discriminated between patients with COVID-19 and those without COVID-19, with areas under the receiver operating characteristic curve of 0.95 (95% CI: 0.91, 0.98) and 0.88 (95% CI: 0.84, 0.93), for the internal and external test sets, respectively. Agreement with the eight human observers was moderate to substantial, with mean linearly weighted κ values of 0.60 ± 0.01 for CO-RADS scores and 0.54 ± 0.01 for CT severity scores. Conclusion With high diagnostic performance, the CO-RADS AI system correctly identified patients with COVID-19 using chest CT scans and assigned standardized CO-RADS and CT severity scores that demonstrated good agreement with findings from eight independent observers and generalized well to external data. © RSNA, 2020 Supplemental material is available for this article.

Telomerase activation by genomic rearrangements in high-risk neuroblastoma.

Neuroblastoma is a malignant paediatric tumour of the sympathetic nervous system. Roughly half of these tumours regress spontaneously or are cured by limited therapy. By contrast, high-risk neuroblastomas have an unfavourable clinical course despite intensive multimodal treatment, and their molecular basis has remained largely elusive. Here we have performed whole-genome sequencing of 56 neuroblastomas (high-risk, n = 39; low-risk, n = 17) and discovered recurrent genomic rearrangements affecting a chromosomal region at 5p15.33 proximal of the telomerase reverse transcriptase gene (TERT). These rearrangements occurred only in high-risk neuroblastomas (12/39, 31%) in a mutually exclusive fashion with MYCN amplifications and ATRX mutations, which are known genetic events in this tumour type. In an extended case series (n = 217), TERT rearrangements defined a subgroup of high-risk tumours with particularly poor outcome. Despite a large structural diversity of these rearrangements, they all induced massive transcriptional upregulation of TERT. In the remaining high-risk tumours, TERT expression was also elevated in MYCN-amplified tumours, whereas alternative lengthening of telomeres was present in neuroblastomas without TERT or MYCN alterations, suggesting that telomere lengthening represents a central mechanism defining this subtype. The 5p15.33 rearrangements juxtapose the TERT coding sequence to strong enhancer elements, resulting in massive chromatin remodelling and DNA methylation of the affected region. Supporting a functional role of TERT, neuroblastoma cell lines bearing rearrangements or amplified MYCN exhibited both upregulated TERT expression and enzymatic telomerase activity. In summary, our findings show that remodelling of the genomic context abrogates transcriptional silencing of TERT in high-risk neuroblastoma and places telomerase activation in the centre of transformation in a large fraction of these tumours.

Correction to: The pitfalls and promise of liquid biopsies for diagnosing and treating solid tumors in children: as review.

In the original version of this article, a reader pointed out that there was a mistake in the phrasing in a paragraph. This could potentially be harmful to children. The authors agree to change the wording. "vitreous fluid" will be changed to "aqueous humor".

The pitfalls and promise of liquid biopsies for diagnosing and treating solid tumors in children: a review.

Cell-free DNA profiling using patient blood is emerging as a non-invasive complementary technique for cancer genomic characterization. Since these liquid biopsies will soon be integrated into clinical trial protocols for pediatric cancer treatment, clinicians should be informed about potential applications and advantages but also weaknesses and potential pitfalls. Small retrospective studies comparing genetic alterations detected in liquid biopsies with tumor biopsies for pediatric solid tumor types are encouraging. Molecular detection of tumor markers in cell-free DNA could be used for earlier therapy response monitoring and residual disease detection as well as enabling detection of pathognomonic and therapeutically relevant genomic alterations.Conclusion: Existing analyses of liquid biopsies from children with solid tumors increasingly suggest a potential relevance for molecular diagnostics, prognostic assessment, and therapeutic decision-making. Gaps remain in the types of tumors studied and value of detection methods applied. Here we review the current stand of liquid biopsy studies for pediatric solid tumors with a dedicated focus on cell-free DNA analysis. There is legitimate hope that integrating fully validated liquid biopsy-based innovations into the standard of care will advance patient monitoring and personalized treatment of children battling solid cancers.What is Known:• Liquid biopsies are finding their way into routine oncological screening, diagnosis, and disease monitoring in adult cancer types fast.• The most widely adopted source for liquid biopsies is blood although other easily accessible body fluids, such as saliva, pleural effusions, urine, or cerebrospinal fluid (CSF) can also serve as sources for liquid biopsiesWhat is New:• Retrospective proof-of-concept studies in small cohorts illustrate that liquid biopsies in pediatric solid tumors yield tremendous potential to be used in diagnostics, for therapy response monitoring and in residual disease detection.• Liquid biopsy diagnostics could tackle some long-standing issues in the pediatric oncology field; they can enable accurate genetic diagnostics in previously unbiopsied tumor types like renal tumors or brain stem tumors leading to better treatment strategies.

Pinpointing a potential role for CLEC12B in cancer predisposition through familial exome sequencing.

Predisposition to cancer is only partly understood, and thus, the contribution of still undiscovered cancer predisposing variants necessitates further research. In search of such variants, we performed exome sequencing on the germline DNA of a family with two children affected by ganglioneuroma and neuroblastoma. Applying stringent selection criteria, we identified a potential deleterious, missense mutation in CLEC12B, coding for a lectin C-type receptor that is predicted to regulate immune function. Although further screening in a larger population and functional characterization is needed, we propose CLEC12B as a candidate cancer predisposition gene.

Severe anaphylactic reaction following anti-thymocyte globulin administration in a pediatric stem cell transplantation patient.

We present a case of a severe reaction after anti-thymocyte administration despite premedication and strict adherence to administration guidelines during the conditioning regimen. Due to the severity of the adverse drug reaction, we believe that this case should be reported.

High-throughput PCR assay design for targeted resequencing using primerXL.

BACKGROUND: Although the sequencing landscape is rapidly evolving and sequencing costs are continuously decreasing, whole genome sequencing is still too expensive for use on a routine basis. Targeted resequencing of only the regions of interest decreases both costs and the complexity of the downstream data-analysis. Various target enrichment strategies are available, but none of them obtain the degree of coverage uniformity, flexibility and specificity of PCR-based enrichment. On the other hand, the biggest limitation of target enrichment by PCR is the need to design large numbers of partially overlapping assays to cover the target. RESULTS: To overcome the aforementioned hurdles, we have developed primerXL, a state-of-the-art PCR primer design pipeline for targeted resequencing. It uses an optimized design criteria relaxation cascade and a thorough downstream in silico evaluation process to generate high quality singleplex PCR assays, reducing the need for amplicon normalization, and outperforming other target enrichment strategies and similar primer design tools when considering assay quality, coverage uniformity and target coverage. Results of four different sequencing projects with 2348 amplicons in total covering 470 kb are presented. PrimerXL can be accessed at www.primerxl.org . CONCLUSION: PrimerXL is an state-of-the-art, easy to use primer design webtool capable of generating high-quality targeted resequencing assays. The workflow is fully customizable to suit every researchers' needs, while an innovative relaxation cascade ensures maximal target coverage.

Target enrichment using parallel nanoliter quantitative PCR amplification.

BACKGROUND: Next generation targeted resequencing is replacing Sanger sequencing at high pace in routine genetic diagnosis. The need for well validated, high quality enrichment platforms to complement the bench-top next generation sequencing devices is high. RESULTS: We used the WaferGen Smartchip platform to perform highly parallelized PCR based target enrichment for a set of known cancer genes in a well characterized set of cancer cell lines from the NCI60 panel. Optimization of PCR assay design and cycling conditions resulted in a high enrichment efficiency. We provide proof of a high mutation rediscovery rate and have included technical replicates to enable SNP calling validation demonstrating the high reproducibility of our enrichment platform. CONCLUSIONS: Here we present our custom developed quantitative PCR based target enrichment platform. Using highly parallel nanoliter singleplex PCR reactions makes this a flexible and efficient platform. The high mutation validation rate shows this platform's promise as a targeted resequencing method for multi-gene routine sequencing diagnostics.

Lack of association between MDM2 promoter SNP309 and clinical outcome in patients with neuroblastoma.

While a polymorphism located within the promoter region of the MDM2 proto-oncogene, SNP309 (T > G), has previously been associated with increased risk and aggressiveness of neuroblastoma and other tumor entities, a protective effect has also been reported in certain other cancers. In this study, we evaluated the association of MDM2 SNP309 with outcome in 496 patients with neuroblastoma and its effect on MDM2 expression. No significant difference in overall or event-free survival was observed among patients with neuroblastoma with or without MDM2 SNP309. The presence of SNP309 does not affect MDM2 expression in neuroblastoma.

Diagnosis of pediatric central nervous system tumors using methylation profiling of cfDNA from cerebrospinal fluid.

Pediatric central nervous system tumors remain challenging to diagnose. Imaging approaches do not provide sufficient detail to discriminate between different tumor types, while the histopathological examination of tumor tissue shows high inter-observer variability. Recent studies have demonstrated the accurate classification of central nervous system tumors based on the DNA methylation profile of a tumor biopsy. However, a brain biopsy holds significant risk of bleeding and damaging the surrounding tissues. Liquid biopsy approaches analyzing circulating tumor DNA show high potential as an alternative and less invasive tool to study the DNA methylation pattern of tumors. Here, we explore the potential of classifying pediatric brain tumors based on methylation profiling of the circulating cell-free DNA (cfDNA) in cerebrospinal fluid (CSF). For this proof-of-concept study, we collected cerebrospinal fluid samples from 19 pediatric brain cancer patients via a ventricular drain placed for reasons of increased intracranial pressure. Analyses on the cfDNA showed high variability of cfDNA quantities across patients ranging from levels below the limit of quantification to 40 ng cfDNA per milliliter of CSF. Classification based on methylation profiling of cfDNA from CSF was correct for 7 out of 20 samples in our cohort. Accurate results were mostly observed in samples of high quality, more specifically those with limited high molecular weight DNA contamination. Interestingly, we show that centrifugation of the CSF prior to processing increases the fraction of fragmented cfDNA to high molecular weight DNA. In addition, classification was mostly correct for samples with high tumoral cfDNA fraction as estimated by computational deconvolution (> 40%). In summary, analysis of cfDNA in the CSF shows potential as a tool for diagnosing pediatric nervous system tumors especially in patients with high levels of tumoral cfDNA in the CSF. Further optimization of the collection procedure, experimental workflow and bioinformatic approach is required to also allow classification for patients with low tumoral fractions in the CSF.

Unleashing the strengths of unlabelled data in deep learning-assisted pan-cancer abdominal organ quantification: the FLARE22 challenge.

Deep learning has shown great potential to automate abdominal organ segmentation and quantification. However, most existing algorithms rely on expert annotations and do not have comprehensive evaluations in real-world multinational settings. To address these limitations, we organised the FLARE 2022 challenge to benchmark fast, low-resource, and accurate abdominal organ segmentation algorithms. We first constructed an intercontinental abdomen CT dataset from more than 50 clinical research groups. We then independently validated that deep learning algorithms achieved a median dice similarity coefficient (DSC) of 90·0% (IQR 87·4-91·3%) by use of 50 labelled images and 2000 unlabelled images, which can substantially reduce manual annotation costs. The best-performing algorithms successfully generalised to holdout external validation sets, achieving a median DSC of 89·4% (85·2-91·3%), 90·0% (84·3-93·0%), and 88·5% (80·9-91·9%) on North American, European, and Asian cohorts, respectively. These algorithms show the potential to use unlabelled data to boost performance and alleviate annotation shortages for modern artificial intelligence models.

NBAtlas: A harmonized single-cell transcriptomic reference atlas of human neuroblastoma tumors.

Neuroblastoma, a rare embryonic tumor arising from neural crest development, is responsible for 15% of pediatric cancer-related deaths. Recently, several single-cell transcriptome studies were performed on neuroblastoma patient samples to investigate the cell of origin and tumor heterogeneity. However, these individual studies involved a small number of tumors and cells, limiting the conclusions that could be drawn. To overcome this limitation, we integrated seven single-cell or single-nucleus datasets into a harmonized cell atlas covering 362,991 cells across 61 patients. We use this atlas to decipher the transcriptional landscape of neuroblastoma at single-cell resolution, revealing associations between transcriptomic profiles and clinical outcomes within the tumor compartment. In addition, we characterize the complex immune-cell landscape and uncover considerable heterogeneity among tumor-associated macrophages. Finally, we showcase the utility of our atlas as a resource by expanding it with additional data and using it as a reference for data-driven cell-type annotation.

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.