Feasibility of real-time molecular profiling for patients with newly diagnosed glioblastoma without MGMT promoter hypermethylation—the NCT Neuro Master Match (N2M2) pilot study

Feasibility of real-time molecular profiling for patients with newly diagnosed glioblastoma... Abstract Background O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status is a predictive biomarker in glioblastoma patients. Glioblastoma without hypermethylated MGMT promoter is largely resistant to treatment with temozolomide. These patients are in particular need of new treatment approaches, which are offered by biomarker-driven clinical trials with targeted drugs based on molecular characterization of individual tumors. Methods In preparation for an upcoming clinical study, a comprehensive molecular profiling approach was undertaken on tissues from 43 glioblastoma patients harboring an unmethylated MGMT promoter at diagnosis. The diagnostic pipeline covered various levels of molecular characteristics, including whole-exome sequencing, low-coverage whole-genome sequencing, RNA sequencing, as well as microarray-based gene expression profiling and DNA methylation arrays. Results Complex multilayer molecular diagnostics were feasible in this setting with a median turnaround time of 4–5 weeks from surgery to the molecular tumor board. In 35% of cases, potentially relevant therapeutic decisions were derived from the data. Alterations were most frequently found in receptor tyrosine kinases, members of the phosphoinositide 3-kinase/Akt/mechanistic target of rapamycin and mitogen-activated protein kinase pathway as well as cell cycle control and p53 regulation cascades. Individual tumors harbored clonal alterations such as oncogenic fusions of tyrosine kinases which constitute promising targets for targeted therapies. A prioritization algorithm is proposed to allocate patients with multiple targets to the potentially best treatment option. Conclusion With this feasibility study, a comprehensive molecular profiling approach for patients with newly diagnosed glioblastoma harboring an unmethylated MGMT promoter is presented. Analyses in this pilot cohort serve as a basis for trials based on targetable alterations and on the question of allocation of patients to the best treatment arm. glioblastoma, MGMT, N2M2, precision oncology Importance of the study The diagnosis of glioblastoma is associated with a poor prognosis—this holds especially true for patients with glioblastoma without MGMT hypermethylation showing resistance toward the standard-of-care treatment with temozolomide. With this study, we present a comprehensive approach for molecular diagnostics applicable at diagnosis in a clinically meaningful timeframe. Implementing methods to investigate various molecular levels reveals well-known alterations in glioblastoma, but also identifies rare promising targets for targeted therapies. The proposed prioritization algorithm serves as a basis for determining the most suitable targeted treatment for the individual patient in cases with multiple druggable targets. It also allows for planning and allocating patients to biomarker-driven clinical trials. Glioblastoma is the most frequent malignant brain tumor in adults, constituting 60%–75% of astrocytic tumors and ~15% of all intracranial tumors.1 The incidence is about 3–4 new cases per 100000 people/year.1 The prognosis for patients with glioblastoma is particularly unfavorable, with a median survival of ~15 months and 5-year survival rate of less than 10% in trial cohorts.1,2 The current standard-of-care (SOC) treatment comprises induction concomitant with radiotherapy combined with oral temozolomide, followed by a maintenance phase with six to twelve 28-day cycles of temozolomide.3 Temozolomide as an alkylating agent exerts its cytotoxic effect through the formation of DNA crosslinks which lead to DNA strand breaks. Alkylation at guanine O6 can be reversed by the DNA repair protein O6-methylguanine-DNA methyltransferase (MGMT)4,5—thereby preventing the formation of lethal crosslinks. In line with this, patients harboring tumors with a hypermethylated MGMT promoter and thus silencing of MGMT gene expression experience a significant survival benefit from temozolomide chemotherapy. In contrast, patients with an unmethylated MGMT promoter show only minimal response.6,7 Therefore, particularly the latter group is in specific need of new treatments. Frequent genetic alterations in glioblastoma affect receptor tyrosine kinases (RTKs), such as epidermal growth factor receptor (EGFR), platelet derived growth factor receptor alpha (PDGFRA), fibroblast growth factor receptor (FGFR), and MET proto oncogene RTK (MET).8EGFR is one of the most commonly altered genes, mainly affected by amplifications but also activated by somatic single nucleotide variants (SNVs)8 or rearrangements including constitutively active EGFR variant III (EGFRvIII).9 Although several therapeutic agents targeting EGFR and EGFRvIII have been investigated in clinical trials,10,11 no clinical benefit has been proven to result from addressing this pathway in glioblastoma. Downstream of RTKs, alterations in different members and regulators of the phosphoinositide 3-kinase (PI3K)/Akt/mechanistic target of rapamycin (mTOR) as well as the mitogen activated protein kinase (MAPK) signaling pathways are regularly found. Cell cycle regulation is altered in approximately 80% of glioblastoma,8,12 mainly due to cyclin-dependent kinase inhibitor (CDKN) 2A/B deletion or cyclin-dependent kinase (CDK) 4/6 amplification. Finally, TP53 mutations or mouse double minute (MDM) 2/4 alterations impact p53 signaling, senescence, and apoptosis in glioblastoma. Thus, the spectrum of molecular characteristics in glioblastoma comprises, at a relatively high frequency, genetic alterations which can be targeted by specific drugs. Therefore, clinical trials focusing on targeted therapies based on molecular characteristics of the individual tumor represent an attractive next step in the treatment of glioblastoma. However, glioblastomas display extensive intratumoral heterogeneity at the level of the genome, transcriptome, and epigenome.13 This heterogeneity, with subclones showing different characteristics, is one critical aspect to consider when planning biomarker-driven clinical trials. Several clinical studies including targeted agents in newly diagnosed glioblastoma have failed to demonstrate efficacy in unselected patient cohorts.12,14–16 Therefore, well-considered allocation of patients to clinical trials based on molecular characteristics of the tumor as well as necessary retrospective validation of potential biomarkers are essential in a clinical setting. The aim of this study was to evaluate the feasibility of timely, comprehensive molecular diagnostics and interpretation in a clinical setting. Our analysis of a prospective cohort of 43 matched tumor and germline samples using next-generation sequencing, gene-expression profiling, and DNA methylation analysis also serves as a basis for development of stratification algorithms to allocate patients to biomarker-driven clinical trials with targeted drugs. The applicability of a similar pipeline in a clinically meaningful manner has previously been shown in the INFORM pilot phase for high-risk relapsed pediatric malignancies,17 and we here aimed to demonstrate that such a model was transferable to a routine adult neuro-oncology setting at initial diagnosis. Materials and Methods Cohort The NCT Neuro Master Match (N2M2) pilot cohort includes 43 patients with newly diagnosed glioblastoma without MGMT promoter hypermethylation determined by pyrosequencing.18 Thirty-nine of them were diagnosed at the Heidelberg University Hospital between July 15, 2014 and November 11, 2016. Glioblastomas in 4 patients were first diagnosed in other hospitals in Germany (Koblenz, Mannheim, Karlsruhe, Munich) and tissue was sent to our institution for analysis. Patients provided informed consent concerning the use of their tissue samples for research purposes. The concept of the investigation was approved by the local ethics committee (no. 206/2005, University of Heidelberg, Medical Faculty Ethics Committee). Pyrosequencing for Determining MGMT Promoter Methylation Analysis of MGMT promoter methylation status through pyrosequencing was performed with the Therascreen MGMT Pyro Kit (Qiagen) according to the manufacturer’s instructions. Quantitative measurement of methylation in 4 cytosine-phosphate-guanine (CpG) sites in exon 1 of the MGMT gene was performed. The cutoff was set at 8%. Infinium HumanMethylation450 Bead Chip and MethylationEPIC Kits The Illumina Infinium HumanMethylation450 (450k) bead chip and MethylationEPIC kits were used to obtain the DNA methylation status at >450000 and >850000 CpG sites, respectively, according to the manufacturer’s instructions at the Genomics and Proteomics Core Facility of the German Cancer Research Center in Heidelberg, Germany. Samples were analyzed using the R (www.r-project.org) based methylation pipeline “ChAMP.”19 In brief, filtering was done for multihit sites, SNPs, and XY chromosome–related CpGs, then data were normalized with a Beta-Mixture Quantile–based method and analyzed for batch effects with a singular value decomposition algorithm. Batch effects related to the tissue used (formalin-fixed paraffin-embedded [FFPE] vs fresh frozen) were corrected using ComBat. MGMT promoter methylation status was determined by the algorithm of Bady et al.20 Next-Generation Sequencing and Gene Expression Profiling Comprehensive molecular profiling (low-coverage whole-genome sequencing [lc-WGS] and whole-exome sequencing [WES], RNA-sequencing, and gene expression microarray) was conducted as described before.17 Confirmation of specific copy number variation (CNV) was performed using fluorescence in situ hybridization (FISH). Furthermore, gene panel sequencing using an in-house gene panel comprising 130 neuro-oncology relevant genes was added as previously specified.21 A detailed description of the molecular profiling can be found in the Supplementary material. Comparison of Different Diagnostic Methods In order to address the extensive intratumoral heterogeneity known in glioblastoma, molecular diagnostic approaches should comprise different techniques and should be applied at sufficient depth to investigate events occurring at various levels of clonality. We additionally aimed to maximize the chance to identify important hits and potential heterogeneity of their abundance by analyzing different areas of the same tumor specimen. Samples collected for routine pathology were paraffin embedded and used for gene panel sequencing and DNA methylation profiling, whereas fresh frozen material was taken for the dedicated molecular workup of the study including WES. mTOR Immunohistochemistry Immunohistochemistry to detect phospho-mTOR was performed as described previously12 using a heat antigen retrieval procedure (citrate buffer) and the phospho-mTOR antibody (Ser-2448, #2976, Cell Signaling Technology) in the dilution 1:100 according to manufacturer’s recommendations. Algorithm for Potential Assignment to Targeted Therapies Assignment to potential therapies was performed based on the molecular profile of the glioblastoma samples. Patients with glioblastoma harboring one of the molecular characteristics given in Supplementary Table S1 were allocated to the respective hypothetical group. For patients with tumors harboring more than one of the described molecular characteristics, a prioritization assignment algorithm (Table 1) was developed to rank the targeted therapies with respective biomarkers to allocate the patient to that targeted therapy with the highest estimated chance of response. This algorithm is partly inspired by the INFORM algorithm17 and considers clinical and preclinical evidence for benefit of a specific drug-biomarker combination in glioblastoma or other types of cancer as well as the nature of the genetic alteration. Of note, for each target, one test was considered as the formal reference—ie, WES for BRAF V600E mutation (vemurafenib) and anaplastic lymphoma kinase (ALK) point mutation (alectinib); RNA-Seq for fibroblast growth factor receptor (FGFR)–transforming acidic coiled-coil protein (TACC) fusion (ADZ4547) and MET fusion (crizotinib); methylation array for CDK4/6 amplification (palbociclib), MDM2 amplification (idasanutlin), and sonic hedgehog (SHH) amplification (vismodegib); and immunohistochemistry for mTOR Ser-2448 phosphorylation (temsirolimus). We allocated the patients of the above-mentioned alterations to the group that matches the molecular profile of the tumor and has the highest rank as given in Supplementary Table S1. This list gives a theoretical rank and may vary in a real trial according to drug and trial arm availability. Patients without any of these characteristics cannot be assigned to a specific group and in case of the N2M222 clinical trial will undergo block randomization to treatments with either a programmed cell death ligand 1 (PD-L1) antibody, cluster of differentiation 95 ligand (CD95L) inhibitory recombinant protein, or temozolomide. Table 1 Prioritization algorithm for biomarker-based targeted treatment Group Criterion 1 Biomarker with approved biomarker specific treatment in glioblastoma + with strong survival benefit – with moderate survival benefit or inconsistent 2A Biomarker with approved biomarker specific treatment in another cancer indication with compelling clinical evidence in glioblastoma 2B Biomarker with approved biomarker specific treatment in another cancer indication not tested in glioblastoma in a clinical setting 3A Clinical evidence in glioblastoma, but not approved in glioblastoma or any other cancer indication + mutation – amp/expression 3B Clinical evidence in another cancer indication, makes biological sense in glioblastoma, but no clinical evidence in glioblastoma + mutation – amp/expression 4A No compelling clinical evidence in any cancer indication but retrospective biomarker assessment in + glioblastoma – another cancer indication 4B No compelling clinical evidence in any cancer indication, but stable genetic change (mutation/fusion) and makes biological sense in glioblastoma ± preclinical evidence 4C No compelling clinical evidence in any cancer indication, expression change, phosphorylation, etc. is a direct drug target and makes biological sense in glioblastoma 4D Alteration of pathways or genetic alteration/expression change regulates drug target Group Criterion 1 Biomarker with approved biomarker specific treatment in glioblastoma + with strong survival benefit – with moderate survival benefit or inconsistent 2A Biomarker with approved biomarker specific treatment in another cancer indication with compelling clinical evidence in glioblastoma 2B Biomarker with approved biomarker specific treatment in another cancer indication not tested in glioblastoma in a clinical setting 3A Clinical evidence in glioblastoma, but not approved in glioblastoma or any other cancer indication + mutation – amp/expression 3B Clinical evidence in another cancer indication, makes biological sense in glioblastoma, but no clinical evidence in glioblastoma + mutation – amp/expression 4A No compelling clinical evidence in any cancer indication but retrospective biomarker assessment in + glioblastoma – another cancer indication 4B No compelling clinical evidence in any cancer indication, but stable genetic change (mutation/fusion) and makes biological sense in glioblastoma ± preclinical evidence 4C No compelling clinical evidence in any cancer indication, expression change, phosphorylation, etc. is a direct drug target and makes biological sense in glioblastoma 4D Alteration of pathways or genetic alteration/expression change regulates drug target View Large Table 1 Prioritization algorithm for biomarker-based targeted treatment Group Criterion 1 Biomarker with approved biomarker specific treatment in glioblastoma + with strong survival benefit – with moderate survival benefit or inconsistent 2A Biomarker with approved biomarker specific treatment in another cancer indication with compelling clinical evidence in glioblastoma 2B Biomarker with approved biomarker specific treatment in another cancer indication not tested in glioblastoma in a clinical setting 3A Clinical evidence in glioblastoma, but not approved in glioblastoma or any other cancer indication + mutation – amp/expression 3B Clinical evidence in another cancer indication, makes biological sense in glioblastoma, but no clinical evidence in glioblastoma + mutation – amp/expression 4A No compelling clinical evidence in any cancer indication but retrospective biomarker assessment in + glioblastoma – another cancer indication 4B No compelling clinical evidence in any cancer indication, but stable genetic change (mutation/fusion) and makes biological sense in glioblastoma ± preclinical evidence 4C No compelling clinical evidence in any cancer indication, expression change, phosphorylation, etc. is a direct drug target and makes biological sense in glioblastoma 4D Alteration of pathways or genetic alteration/expression change regulates drug target Group Criterion 1 Biomarker with approved biomarker specific treatment in glioblastoma + with strong survival benefit – with moderate survival benefit or inconsistent 2A Biomarker with approved biomarker specific treatment in another cancer indication with compelling clinical evidence in glioblastoma 2B Biomarker with approved biomarker specific treatment in another cancer indication not tested in glioblastoma in a clinical setting 3A Clinical evidence in glioblastoma, but not approved in glioblastoma or any other cancer indication + mutation – amp/expression 3B Clinical evidence in another cancer indication, makes biological sense in glioblastoma, but no clinical evidence in glioblastoma + mutation – amp/expression 4A No compelling clinical evidence in any cancer indication but retrospective biomarker assessment in + glioblastoma – another cancer indication 4B No compelling clinical evidence in any cancer indication, but stable genetic change (mutation/fusion) and makes biological sense in glioblastoma ± preclinical evidence 4C No compelling clinical evidence in any cancer indication, expression change, phosphorylation, etc. is a direct drug target and makes biological sense in glioblastoma 4D Alteration of pathways or genetic alteration/expression change regulates drug target View Large Results Clinical Characteristics of the Patient Cohort The N2M2 pilot cohort consisted of 43 patients with newly diagnosed glioblastoma. Inclusion was based on MGMT promoter methylation status determined by pyrosequencing and IDH1/2 wild-type status in all 43 cases. Twenty-one (49%) women and 22 (51%) men were included with an average age at diagnosis of 62.8 years (Table 2). Most patients had an Eastern Cooperative Oncology Group performance score of 0 (63%), and 19 (44%) patients were on steroid treatment at the time of diagnosis. Follow-up data on the primary therapy was available for 40 (93%) patients. Twenty-eight (65%) received primary radiotherapy with temozolomide, 11 patients (26%) were treated with radiotherapy only. One patient had not started further therapy at the time of analysis (Table 2 and Supplementary Table S2). Table 2. Clinical characteristics of the patient cohort Clinical Characteristics Pilot Phase Cohort (n = 43) Sex, no. (%)  Female 21 (48.8)  Male 22 (51.2) Age, no. (%)  <60 y 16 (37.2)  60–69 y 12 (27.9)  ≥70 y 15 (34.9) Age, mean (SD) 62.8 (11.5) ECOG, no. (%)  0 27 (62.8)  1 8 (18.7)  ≥2 8 (18.7) Steroids at baseline, no. (%)  Yes 19 (44.2)  No 23 (53.5)  Missing data 1 (2.3) Localization, no. (%)  Left hemisphere 26 (60.4)  Right hemisphere 14 (32.6)  Corpus callosum 1 (2.3)  Cerebellar 1 (2.3)  Intramedular 1 (2.3) Primary therapy, no. (%)  RT + TMZ 28 (65.1)  RT alone 11 (25.6)  TMZ alone 0 (0.0)  Other 0 (0.0)  None 1 (2.3)  Missing data 3 (7.0) Clinical Characteristics Pilot Phase Cohort (n = 43) Sex, no. (%)  Female 21 (48.8)  Male 22 (51.2) Age, no. (%)  <60 y 16 (37.2)  60–69 y 12 (27.9)  ≥70 y 15 (34.9) Age, mean (SD) 62.8 (11.5) ECOG, no. (%)  0 27 (62.8)  1 8 (18.7)  ≥2 8 (18.7) Steroids at baseline, no. (%)  Yes 19 (44.2)  No 23 (53.5)  Missing data 1 (2.3) Localization, no. (%)  Left hemisphere 26 (60.4)  Right hemisphere 14 (32.6)  Corpus callosum 1 (2.3)  Cerebellar 1 (2.3)  Intramedular 1 (2.3) Primary therapy, no. (%)  RT + TMZ 28 (65.1)  RT alone 11 (25.6)  TMZ alone 0 (0.0)  Other 0 (0.0)  None 1 (2.3)  Missing data 3 (7.0) Abbreviations: ECOG, Eastern Cooperative Oncology Group; RT, radiotherapy; TMZ, temozolomide. View Large Table 2. Clinical characteristics of the patient cohort Clinical Characteristics Pilot Phase Cohort (n = 43) Sex, no. (%)  Female 21 (48.8)  Male 22 (51.2) Age, no. (%)  <60 y 16 (37.2)  60–69 y 12 (27.9)  ≥70 y 15 (34.9) Age, mean (SD) 62.8 (11.5) ECOG, no. (%)  0 27 (62.8)  1 8 (18.7)  ≥2 8 (18.7) Steroids at baseline, no. (%)  Yes 19 (44.2)  No 23 (53.5)  Missing data 1 (2.3) Localization, no. (%)  Left hemisphere 26 (60.4)  Right hemisphere 14 (32.6)  Corpus callosum 1 (2.3)  Cerebellar 1 (2.3)  Intramedular 1 (2.3) Primary therapy, no. (%)  RT + TMZ 28 (65.1)  RT alone 11 (25.6)  TMZ alone 0 (0.0)  Other 0 (0.0)  None 1 (2.3)  Missing data 3 (7.0) Clinical Characteristics Pilot Phase Cohort (n = 43) Sex, no. (%)  Female 21 (48.8)  Male 22 (51.2) Age, no. (%)  <60 y 16 (37.2)  60–69 y 12 (27.9)  ≥70 y 15 (34.9) Age, mean (SD) 62.8 (11.5) ECOG, no. (%)  0 27 (62.8)  1 8 (18.7)  ≥2 8 (18.7) Steroids at baseline, no. (%)  Yes 19 (44.2)  No 23 (53.5)  Missing data 1 (2.3) Localization, no. (%)  Left hemisphere 26 (60.4)  Right hemisphere 14 (32.6)  Corpus callosum 1 (2.3)  Cerebellar 1 (2.3)  Intramedular 1 (2.3) Primary therapy, no. (%)  RT + TMZ 28 (65.1)  RT alone 11 (25.6)  TMZ alone 0 (0.0)  Other 0 (0.0)  None 1 (2.3)  Missing data 3 (7.0) Abbreviations: ECOG, Eastern Cooperative Oncology Group; RT, radiotherapy; TMZ, temozolomide. View Large Timelines One of the aims of this feasibility study was to determine whether a comprehensive molecular profiling could be performed in the clinically demanded timeframe of 6 weeks from surgery. In 39 of 43 (90.7%) cases, surgery was performed locally at the Heidelberg University Hospital, and material arrived in the neuropathology laboratory at the day of operation. Four patients were operated on in external centers and tissue was subsequently analyzed at the Department of Neuropathology at the Heidelberg University Hospital. For timeline calculations, only prospectively included cases (n = 35) were considered. The median time interval for sample processing before submission to pyrosequencing was 4 days (range 0–21). Pyrosequencing and DNA methylation array analysis were completed within a median of 3.5 days (range 0–25) and 26 days (range 8–84), respectively. WES, WGS, and microarray gene expression profile were performed within a median of 12 days (range 8–33) and 9 days (range 7–16), respectively. Panel sequencing was completed within a median of 23 days (range 7–42). Bioinformatic processing of sequencing data required 2 days as median (range 0–12) and data interpretation could be completed within 2–3 days (Fig. 1A). Thus, when running as a routine part of a clinical study, the whole pipeline, including clinical decision for study arm allocation, is feasible within 4–5 weeks (Fig. 1B). This also corresponds to our experience in now more than 300 patients prospectively enrolled in the INFORM23 study. Fig. 1 View largeDownload slide Schematic illustration of the workflow and timeline of the N2M2 molecular diagnostic pipeline. (A) The real median time for each step of the process is calculated for the prospectively included cases, range of time intervals is given in brackets. (B) Ideal time intervals for prospectively analyzed cases with the whole molecular diagnostics pipeline being performed within 4 weeks from day of tissue arrival to discussion in the molecular tumorboard. Fig. 1 View largeDownload slide Schematic illustration of the workflow and timeline of the N2M2 molecular diagnostic pipeline. (A) The real median time for each step of the process is calculated for the prospectively included cases, range of time intervals is given in brackets. (B) Ideal time intervals for prospectively analyzed cases with the whole molecular diagnostics pipeline being performed within 4 weeks from day of tissue arrival to discussion in the molecular tumorboard. MGMT Methylation Status: Pyrosequencing versus 450k/850k Methylation Array Determination of MGMT status for inclusion in the present study was based on pyrosequencing, which is considered the gold standard.18,24 Prior studies of our group have suggested discordance in different methods for MGMT promoter methylation testing.25 In 86.0% (37/43) of cases, the pyrosequencing and DNA methylation analysis results were consistent. Two cases showed an unmethylated MGMT promoter based on pyrosequencing (4% and 6% methylation, respectively), but the MGMT promoter was methylated based on DNA methylation array. Four cases were not classifiable with sufficient confidence to either methylated or unmethylated based on DNA methylation array. Notably, 2 distinct CpG sites are analyzed for MGMT promoter methylation analysis using the Infinium 450k bead chip and MethylationEPIC kits, respectively,20 which differ from the CpGs in the pyrosequencing approach. Single Nucleotide Variants: WES versus Gene Panel The mean count of SNVs was 47.7 per exome (range 19–80) based on WES, in line with previous reports of adult glioblastoma26 and higher than in pediatric high-grade glioma.27 To explore intratumoral heterogeneity, targeted sequencing using an in-house panel of 130 neuro-oncology–relevant genes (Panel Seq)21 was performed on FFPE material and compared with WES conducted on fresh frozen material from a different region of the same tumor. For 36 cases (84%) gene panel data were available from FFPE material, whereas for the remaining 7 cases, Panel Seq was performed on the same fresh frozen material as WES. Alterations detected in these 7 cases were completely consistent in both methods. In the group of cases with sequencing data from different material types (FFPE and frozen), we altogether detected 69 alterations (SNVs and insertions/deletions), which we would consider as either potentially druggable targets or tumor biologically interesting findings (Fig. 2). In total, 74% of alterations (48/69) were detected by both methods. One case harbored 2 different nucleotide substitutions within the phosphatase and tensin homolog (PTEN) gene, each of the SNVs revealed by one of the 2 methods, respectively. Three point mutations were detected by WES in genes which were not covered in the gene panel. Fig. 2 View largeDownload slide Overview of selected genetic alterations for each case, subgroup affiliation based on DNA methylation as well as expression analysis and mTOR staining results. Fig. 2 View largeDownload slide Overview of selected genetic alterations for each case, subgroup affiliation based on DNA methylation as well as expression analysis and mTOR staining results. Of the 25% of alterations that were found by only one method, 5 of 69 were high allele frequency mutations, including one alteration found only by WES (MDM2) and not in the Panel Seq, and 4 alterations found only by Panel Seq and not WES (phosphoinositide -4,5-bisphosphate 3-kinase catalytic subunit alpha [PIK3CA], 2x EGFR, FGFR4), respectively. For these alterations, a technical reason such as poor coverage or very low allele frequency was ruled out, therefore the discrepancy points to intratumoral heterogeneity. This is underlined by the fact that no discrepant alterations were found in the cases with sequencing data from the same material. Another 9% of alterations could be detected by only one of the methods due to a low variant allele frequency (of lower than 10%). The remaining 9% of discrepancies were due to differences in the bioinformatic pipelines used. Therefore, the pipelines are now being further optimized based on these results. Copy Number Variations: Low-Coverage WGS versus 450k/850k Copy number plots obtained from lc-WGS were compared with those obtained from methylation array analysis for 36 cases. The overlap between both methods accounted for 88% (77/88) of alterations considered to be potential drug targets and/or tumor-biologically relevant (Fig. 2). Four of 88 druggable alterations could be found only in lc-WGS data, whereas 7 of 88 were detected in copy number plots from methylation array but were not observed in lc-WGS. The likely explanation for these discrepancies is again assumed to be intratumoral heterogeneity, since the relevant aberrations were of a type and size that should be detectable by both methods. One case (Pilot_GBM_26)—besides harboring CDK4 and MDM2 amplification (detected with both methods)—showed EGFR amplification in the lc-WGS data only and PDGFRA amplification in the methylation analysis only. Both alterations were high-level focal amplifications, but each was observed by only one of the 2 methods. EGFR amplification was validated by FISH in the piece used for lc-WGS but was not detectable in the piece for methylation analysis (Supplementary Figure S1). The tissue pieces taken for methylation analysis and lc-WGS therefore likely showed a strong enrichment for one of these 2 subclones, to the extent that the minor population was below the level of detection. Mosaic amplification of RTKs (ie, EGFR and PDGFRA) was previously described in glioblastoma with distinct tumor cell populations harboring either EGFR or PDGFRA amplification.28 The phenomenon of chromothripsis as a single catastrophic event leading to multiple chromosomal rearrangements was also described in glioblastoma,29 especially linked to MDM2/4, CDK4, and EGFR amplification.30 Three cases in the cohort investigated showed chromothripsis of 1–2 chromosomes/chromosome arms. Notably, 2 of the tumors exhibited concomitant CDK4 and MDM2 amplification, whereas the third tumor was EGFR amplified (Supplementary Figure S2). Assignment to Previously Described Subgroups Based on DNA Methylation Analysis and Gene Expression Profiling Using an in-house classifier scoring algorithm31 based on genome-wide DNA methylation patterns, samples could be allocated to previously established methylation subgroups.32 Of all cases, 88.4% (38/43) fall into the glioblastoma methylation groups of the mesenchymal, RTK I, or RTK II subtypes (Fig. 2). For 2 samples (pilot_GBM_22 and pilot_GBM_32) the methylation profile was most similar to pediatric glioblastoma groups (one to the pediatric RTK subtype and one to a methylation group enriched for MYCN amplification,33 respectively). One tumor (pilot_GBM_41) had a high leukocyte infiltration reflected by the highest score for the glioblastoma methylation group with high leukocyte infiltration. This indicates a low tumor cell content (in line with low variant allele frequency and rather flat copy number plot in this case), which also needs to be considered in such personalized medicine efforts. Two cases (pilot_GBM_2, pilot_GBM_13) could not be allocated clearly to one specific subgroup, but rather to a general group of high-grade gliomas. Samples were furthermore assigned to the proneural, classical, and mesenchymal expression groups of Verhaak et al (Fig. 2, Supplementary Figure S3).34 The concordance between the methylation and expression based groups was relatively low. Frequently Affected Oncogenic Pathways The most frequently altered pathways in the cohort investigated were (receptor) tyrosine kinase signaling (RTK), PI3K/Akt/mTOR pathway, MAPK pathway, cell cycle control, and TP53 regulation. Frequencies of alterations in these pathways as well as in individual pathway members are depicted in Fig. 3. Fig. 3 View largeDownload slide Alterations of well-established oncogenic pathways. Frequencies of alterations are specifically illustrated for selected members of the single pathways. Only genetic alterations (SNV, indel, CNV, fusion) are considered. Fig. 3 View largeDownload slide Alterations of well-established oncogenic pathways. Frequencies of alterations are specifically illustrated for selected members of the single pathways. Only genetic alterations (SNV, indel, CNV, fusion) are considered. Amplifications and mutations of the EGFR gene were the most commonly occurring alterations of RTKs (18/43; 42%), followed by PDGFRA alterations (5/43; 12%). Less frequently affected tyrosine kinases included FGFR3/4, VEGFR, MET, and NTRK3. Taken together, 61% (26/43) of cases showed a genetic alteration in at least one RTK, with 6 cases harboring alterations in 2 or 3 different RTKs. Regarding alterations of the PI3K/Akt/mTOR pathway, PTEN loss-of-function mutations and deletions were most frequent (24/43; 56%). To a lesser extent, PIK3CA/B and PIK3R1 mutations were also observed (10/43; 23%). Overall, genetic alterations of the PI3K/Akt/mTOR pathway were found in 70% (30/43) of cases. Alterations affecting members of the MAPK pathway (downstream of RTKs) were detected in 21% (9/43), partly overlapping with RTK alterations. Dysregulation of cell cycle control was mostly due to CDKN2A/B deletion (occurring in 29/43 cases; 67%) and at a lower frequency due to CDK4 amplification (8/43 cases; 19%). No CDK6 amplifications were observed. Altogether, only 14% of cases showed no obvious alterations of cell cycle control mechanisms (6/43). TP53 gene mutations were detected in 12 of 43 (28%) cases, in line with previous reports.8,35 Amplifications of the p53 regulators MDM2 and MDM4 were each found in 9% (4/43) of cases. None of these cases harbored a concomitant TP53 mutation. MDM2 overexpression, defined as reads per kilobase of exon model per million mapped reads >50 based on RNA-Seq data, correlated with gene amplification in 3 of 4 cases. One further case with high MDM2 expression showed no gene amplification. Less frequent genetic alterations occurring in single cases included genes involved in transcription, chromatin remodeling, DNA repair, Notch and Wnt signaling pathway, as well as telomerase maintenance. mTOR Staining A subset of patients with MGMT unmethylated glioblastoma showing phosphorylation of mTOR at Ser-2448 may have some benefit from mTOR inhibitor treatment.12 Therefore, 25 samples with material available were stained for mTOR Ser-2448. Four of 25 cases (16%) were evaluated as “mTOR positive” with 100% of cells showing a score of ≥2 (see Fig. 2 and Supplementary Figure S4). Alterations as Promising Targets for Biomarker-Driven Therapies The present study does not constitute an interventional trial, nor have patients been assigned to any particular therapy based on these data. However, hypothetically we could identify alterations that are promising druggable targets in glioblastoma in 15 of 43 (35%) cases (Fig. 4). Given the parallel design of an interventional trial, N2M2,22 we have tested available arms and found matches for the MDM2 inhibitor idasanutlin, the CDK4/6 inhibitor palbociclib, and the mTOR inhibitor temsirolimus in 11 of 43 (26%) cases. Eight patients would fit into the palbociclib group, 4 patients into the idasanutlin group, and 4 patients into the temsirolimus group, including 5 patients fitting into 2 of these groups. Thirty-two of 43 (74%) patients had no biomarker match for this trial, but 4 cases had a further promising biomarker not currently planned to be included in the trial design. Fig. 4 View largeDownload slide Flowchart of trial arm allocation according to matching biomarkers. Flowchart shows possible allocation of patients to different trials or randomization according to druggable targets detected. *According to evidence level, **best matching target or where recruitment possible; ITT: intention to treat. Fig. 4 View largeDownload slide Flowchart of trial arm allocation according to matching biomarkers. Flowchart shows possible allocation of patients to different trials or randomization according to druggable targets detected. *According to evidence level, **best matching target or where recruitment possible; ITT: intention to treat. Of these, a hotspot BRAF V600E mutation was found. The overall incidence of this mutation in adult glioblastoma is considered relatively low.36–38 Single case reports describe a response to BRAF inhibitor treatment in pediatric patients with high-grade glioma.36,37 Clinical trials applying BRAF inhibitors in glioma patients are planned or ongoing (eg, ClinicalTrials.gov Identifier: NCT02684058). Two cases harbored FGFR3-TACC3 fusions. Oncogenic translocations fusing the kinase domain of FGFR1 or FGFR3 to TACC1 or TACC3, respectively, are described in ~3% of glioblastoma and were shown to be sensitive to FGFR inhibitor treatment in in vivo models.39 Fusion of MET and protein tyrosine phosphatase receptor type Z1 could be detected in one further case. Rearrangements affecting MET were described as being oncogenic and a promising drug target in glioblastoma for treatment with MET inhibitors.23,39 For alteration patterns of the tumors, see Fig. 2. The N2M2 clinical trial furthermore aims to include patients also based on the rare genetic alterations ALK fusion/point mutation for treatment with the ALK inhibitor alectinib and on SHH amplification to the SHH inhibitor vismodegib. No such alterations have been found in the 43 patients within the pilot cohort. We furthermore compared targetable alterations in our cohort with the previously described cohort from The Cancer Genome Atlas (TCGA). Frequencies of the alterations which were also assessed in TCGA were similar in both cohorts (Supplementary Table S3). Prioritization Algorithm for Patients with Multiple Potential Biomarkers Six of 43 (14%) tumors harbored more than one promising target specifically consisting of the overlap between CDK4 and MDM2 amplifications as well as mTOR phosphorylation. This constitutes a challenging group, as a decision must be made on the best target. The refined prioritization algorithm ranked potentially targetable alterations regardless of N2M2 trial availability together with the corresponding available treatments according to (i) the evidence level for benefit of patients harboring the respective biomarker, (ii) whether this evidence comes from glioblastoma or other cancer entities, and (iii) type and stability of the genetic alteration (Table 1, Supplementary Table S1). In the starting clinical trial, patients will be allocated to the best matching N2M2 arm. If high confidence targets could be detected, but the compound arm for targeting this alteration is currently not available in N2M2, inclusion into other appropriate studies based on the molecular aberration will be considered. Patients without a matching biomarker will undergo block randomization between a PD-L1 antibody, a CD95L inhibitory recombinant protein, and SOC temozolomide (Fig. 4). Discussion In contrast to the growing knowledge about molecular characteristics and tumorigenesis of glioblastoma during the last few years, progress in clinical management is still limited. Several clinical studies evaluating targeted agents in unselected patient populations in newly diagnosed glioblastoma have failed to demonstrate efficacy.12,14–16 Therefore, well-considered allocation of patients to clinical trials based on molecular characteristics of the individual tumor is required and critical evaluation of potential biomarkers in a clinical set is essential for further progress. This especially applies for the subgroup of MGMT unmethylated glioblastoma poorly responding to treatment with temozolomide. Inclusion criteria of current and planned studies are increasingly based on MGMT promoter methylation. As these studies may withhold temozolomide in at least one study arm for the poorly responding MGMT unmethylated patients, accurate determination of MGMT promoter methylation status is crucial to avoid withholding temozolomide in patients who might benefit from this drug. In our study, 6 of 43 (14%) tumors with an unmethylated MGMT promoter determined by pyrosequencing and data available from both methods were classified as methylated or were unclassifiable with array-based methods. The discrepancy between the 2 methods can likely be explained by the differing CpG regions assessed in the 2 approaches, as noted above. It is therefore possible that these discrepancies represent true differences in methylation status depending on the precise CpGs analyzed, and it is important to keep this in mind for future stratification. We recently suggested that the response to temozolomide is more complex, and defining the right subgroups that do not benefit may also involve global methylation profiles and status of telomerase reverse transcriptase.40 For the N2M2 clinical trial, we will use pyrosequencing as the currently accepted gold standard to determine MGMT promoter status that defines the inclusion criteria. Patient samples with an unmethylated MGMT promoter defined by methylation array will be validated with pyrosequencing, and patients with MGMT methylation in either method excluded. Nevertheless, further research is needed, especially regarding those patients with discrepant MGMT promoter methylation. With this feasibility study, we present a comprehensive approach including various techniques to allow for an in-depth molecular characterization of each individual tumor. For applicability in a clinical setting, a reasonable timeframe is crucial. In the N2M2 pilot study presented here, molecular characterization was performed within 4–5 weeks after tumor operation, allowing timely decision making for further therapy. In line with previous large-scale studies,8 we could demonstrate that several biologically relevant pathways are affected in the investigated glioblastoma cohort, and in many patients at least one potential treatment biomarker was identified. Since targeted approaches are under consideration for clinical trials in glioblastoma, molecular profiling at diagnosis can serve as a basis for allocation of patients to biomarker-driven clinical trials. One example is the upcoming N2M2 clinical trial that includes patients with an unmethylated MGMT promoter and allocates these patients according to their molecular profile to different treatment arms. We virtually assigned the patients from this pilot cohort into the best fitting treatment arm. In the upcoming trial, patients without matching biomarkers will be randomized between a PD-L1 antibody, a CD95L inhibitory recombinant protein, and SOC temozolomide. For CD95L inhibition, a promising biomarker, low CD95L promoter methylation,41 was present in 18 of 43 (42%) patients. Together, this shows that reasonable biomarkers for the treatment groups are present in a large subset of patients. N2M2 provides a useful structure to test the translation of molecular diagnostics into clinical decisions as well as safety and efficacy evaluations. One problem is the allocation of patients with tumors harboring 2 or more potentially druggable targets to the best specific treatment. Since combined treatments often result in severe side effects and cannot be analyzed in clinical trials due to their complexity, we developed a prioritization algorithm based on clinical and preclinical evidence in glioblastoma and other cancer indications. The algorithm yields the most promising biomarker and targeted treatment for patients with more than one potential target. It can be easily adapted in competing situations with new potential biomarkers evolving in the future or shifting evidence on existing biomarkers according to clinical trial results. Chromothripsis has been observed in 2 patients in the present study. It is questionable whether patients with such large numbers of genetic alterations related to a single event should be excluded from study entry. However, genomic instability may also drive senescence and may enhance dependence on growth-promoting pathways, therefore we plan not to exclude such patients, but to specifically follow up on the disease course. Certainly, inter- as well as intratumoral heterogeneity remain an extreme challenge for the design of biomarker-driven clinical trials, and the full spectrum is not addressed in this study. Several examples are presented which suggest that parallel, complementary analyses from different tumor pieces may be a rational strategy. Future studies will also have to take into account that biomarkers might be unevenly distributed both spatially and temporally within one and between different tumors, which is an important consideration for maximizing the predictive value of biomarkers. A further practical problem is the low frequency of some of the alterations. N2M2 will assess each subtrial individually and only use the temozolomide subtrial as a contemporary reassurance for the validity of the assumptions of progression-free survival. As a next step, randomized, controlled trials will be performed for successful arms with high-frequency molecular lesions like the CDK4/6 amplification (CDKN2A loss) or the mTOR Ser-2448 phosphorylation or for yet-to-be-discovered parameters for the APG101 or atezolizumab subtrials. The lesions with lower frequency will need more time for accrual, a preselection, or different efficacy assessments like objective response rates. In summary, with this feasibility study, we present a comprehensive molecular profiling approach for glioblastoma without MGMT promoter hypermethylation, which can be applied at diagnosis in a clinically relevant timeframe. Knowing the molecular characteristics of the individual tumor allows for allocation to respective biomarker-driven clinical trials as well as retrospective identification of promising biomarkers to hopefully advance the pace of therapeutic progress for glioblastoma patients. Supplementary Material Supplementary material is available at Neuro-Oncology online. Funding We thank the DKFZ-Heidelberg Center for Personalized Oncology (DKFZ-HIPO) for technical support and funding through HIPO project numbers H57 and K25. Conflict of interest statement. No conflicts disclosed. References 1. Ostrom QT , Gittleman H , Fulop J , et al. CBTRUS Statistical Report: primary brain and central nervous system tumors diagnosed in the United States in 2008–2012 . Neuro Oncol . 2015 ; 17 ( Suppl 4 ): iv1 – iv62 . Google Scholar CrossRef Search ADS PubMed 2. Stupp R , Hegi ME , Mason WP , et al. ; European Organisation for Research and Treatment of Cancer Brain Tumour and Radiation Oncology Groups; National Cancer Institute of Canada Clinical Trials Group . Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial . Lancet Oncol . 2009 ; 10 ( 5 ): 459 – 466 . Google Scholar CrossRef Search ADS PubMed 3. Stupp R , Mason WP , van den Bent MJ , et al. ; European Organisation for Research and Treatment of Cancer Brain Tumor and Radiotherapy Groups; National Cancer Institute of Canada Clinical Trials Group . Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma . N Engl J Med . 2005 ; 352 ( 10 ): 987 – 996 . Google Scholar CrossRef Search ADS PubMed 4. Pegg AE , Dolan ME , Moschel RC . Structure, function, and inhibition of O6-alkylguanine-DNA alkyltransferase . Prog Nucleic Acid Res Mol Biol . 1995 ; 51 : 167 – 223 . Google Scholar CrossRef Search ADS PubMed 5. Ludlum DB . DNA alkylation by the haloethylnitrosoureas: nature of modifications produced and their enzymatic repair or removal . Mutat Res . 1990 ; 233 ( 1-2 ): 117 – 126 . Google Scholar CrossRef Search ADS PubMed 6. Esteller M , Garcia-Foncillas J , Andion E , et al. Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents . N Engl J Med . 2000 ; 343 ( 19 ): 1350 – 1354 . Google Scholar CrossRef Search ADS PubMed 7. Hegi ME , Diserens AC , Gorlia T , et al. MGMT gene silencing and benefit from temozolomide in glioblastoma . N Engl J Med . 2005 ; 352 ( 10 ): 997 – 1003 . Google Scholar CrossRef Search ADS PubMed 8. Brennan CW , Verhaak RG , McKenna A , et al. ; TCGA Research Network . The somatic genomic landscape of glioblastoma . Cell . 2013 ; 155 ( 2 ): 462 – 477 . Google Scholar CrossRef Search ADS PubMed 9. Batra SK , Castelino-Prabhu S , Wikstrand CJ , et al. Epidermal growth factor ligand-independent, unregulated, cell-transforming potential of a naturally occurring human mutant EGFRvIII gene . Cell Growth Differ . 1995 ; 6 ( 10 ): 1251 – 1259 . Google Scholar PubMed 10. Gan HK , Kaye AH , Luwor RB . The EGFRvIII variant in glioblastoma multiforme . J Clin Neurosci . 2009 ; 16 ( 6 ): 748 – 754 . Google Scholar CrossRef Search ADS PubMed 11. Weller M , Butowski N , Tran DD , et al. ; ACT IV trial investigators . Rindopepimut with temozolomide for patients with newly diagnosed, EGFRvIII-expressing glioblastoma (ACT IV): a randomised, double-blind, international phase 3 trial . Lancet Oncol . 2017 ; 18 ( 10 ): 1373 – 1385 . Google Scholar CrossRef Search ADS PubMed 12. Wick W , Gorlia T , Bady P , et al. Phase II study of radiotherapy and temsirolimus versus radiochemotherapy with temozolomide in patients with newly diagnosed glioblastoma without MGMT promoter hypermethylation (EORTC 26082) . Clin Cancer Res . 2016 ; 22 ( 19 ): 4797 – 4806 . Google Scholar CrossRef Search ADS PubMed 13. Bonavia R , Inda MM , Cavenee WK , Furnari FB . Heterogeneity maintenance in glioblastoma: a social network . Cancer Res . 2011 ; 71 ( 12 ): 4055 – 4060 . Google Scholar CrossRef Search ADS PubMed 14. Herrlinger U , Schäfer N , Steinbach JP , et al. Bevacizumab plus irinotecan versus temozolomide in newly diagnosed O6-methylguanine-DNA methyltransferase nonmethylated glioblastoma: the randomized GLARIUS trial . J Clin Oncol . 2016 ; 34 ( 14 ): 1611 – 1619 . Google Scholar CrossRef Search ADS PubMed 15. Nabors LB , Fink KL , Mikkelsen T , et al. Two cilengitide regimens in combination with standard treatment for patients with newly diagnosed glioblastoma and unmethylated MGMT gene promoter: results of the open-label, controlled, randomized phase II CORE study . Neuro Oncol . 2015 ; 17 ( 5 ): 708 – 717 . Google Scholar CrossRef Search ADS PubMed 16. Wick W , Steinbach JP , Platten M , et al. Enzastaurin before and concomitant with radiation therapy, followed by enzastaurin maintenance therapy, in patients with newly diagnosed glioblastoma without MGMT promoter hypermethylation . Neuro Oncol . 2013 ; 15 ( 10 ): 1405 – 1412 . Google Scholar CrossRef Search ADS PubMed 17. Worst BC , van Tilburg CM , Balasubramanian GP , et al. Next-generation personalised medicine for high-risk paediatric cancer patients—the INFORM pilot study . Eur J Cancer . 2016 ; 65 : 91 – 101 . Google Scholar CrossRef Search ADS PubMed 18. Mikeska T , Bock C , El-Maarri O , et al. Optimization of quantitative MGMT promoter methylation analysis using pyrosequencing and combined bisulfite restriction analysis . J Mol Diagn . 2007 ; 9 ( 3 ): 368 – 381 . Google Scholar CrossRef Search ADS PubMed 19. Morris TJ , Butcher LM , Feber A , et al. ChAMP: 450k chip analysis methylation pipeline . Bioinformatics . 2014 ; 30 ( 3 ): 428 – 430 . Google Scholar CrossRef Search ADS PubMed 20. Bady P , Sciuscio D , Diserens AC , et al. MGMT methylation analysis of glioblastoma on the Infinium methylation BeadChip identifies two distinct CpG regions associated with gene silencing and outcome, yielding a prediction model for comparisons across datasets, tumor grades, and CIMP-status . Acta Neuropathol . 2012 ; 124 ( 4 ): 547 – 560 . Google Scholar CrossRef Search ADS PubMed 21. Sahm F , Schrimpf D , Jones DT , et al. Next-generation sequencing in routine brain tumor diagnostics enables an integrated diagnosis and identifies actionable targets . Acta Neuropathol . 2016 ; 131 ( 6 ): 903 – 910 . Google Scholar CrossRef Search ADS PubMed 22. Hertenstein A , Jones D , Sahm F , et al. Umbrella protocol for phase I/IIa trials of molecularly matched targeted therapies plus radiotherapy in patients with newly diagnosed glioblastoma without MGMT promoter methylation Neuro Master Match (N2M2) . J Clin Oncol . 2016 ; 34 ( 15 suppl ):doi: 10.1200/JCO.2016.34.15_suppl.TPS2084 . 23. INFORM . INFORM registry webpage . http://www.dkfz.de/en/inform/index.html. 24. Wick W , Weller M , van den Bent M , et al. MGMT testing—the challenges for biomarker-based glioma treatment . Nat Rev Neurol . 2014 ; 10 ( 7 ): 372 – 385 . Google Scholar CrossRef Search ADS PubMed 25. Wiestler B , Capper D , Hovestadt V , et al. Assessing CpG island methylator phenotype, 1p/19q codeletion, and MGMT promoter methylation from epigenome-wide data in the biomarker cohort of the NOA-04 trial . Neuro Oncol . 2014 ; 16 ( 12 ): 1630 – 1638 . Google Scholar CrossRef Search ADS PubMed 26. Alexandrov LB , Nik-Zainal S , Wedge DC , et al. ; Australian Pancreatic Cancer Genome Initiative; ICGC Breast Cancer Consortium; ICGC MMML-Seq Consortium; ICGC PedBrain . Signatures of mutational processes in human cancer . Nature . 2013 ; 500 ( 7463 ): 415 – 421 . Google Scholar CrossRef Search ADS PubMed 27. International Cancer Genome Consortium PedBrain Tumor P . Recurrent MET fusion genes represent a drug target in pediatric glioblastoma . Nat Med . 2016 ; 22 ( 11 ): 1314 – 1320 . CrossRef Search ADS PubMed 28. Snuderl M , Fazlollahi L , Le LP , et al. Mosaic amplification of multiple receptor tyrosine kinase genes in glioblastoma . Cancer Cell . 2011 ; 20 ( 6 ): 810 – 817 . Google Scholar CrossRef Search ADS PubMed 29. Malhotra A , Lindberg M , Faust GG , et al. Breakpoint profiling of 64 cancer genomes reveals numerous complex rearrangements spawned by homology-independent mechanisms . Genome Res . 2013 ; 23 ( 5 ): 762 – 776 . Google Scholar CrossRef Search ADS PubMed 30. Furgason JM , Koncar RF , Michelhaugh SK , et al. Whole genome sequence analysis links chromothripsis to EGFR, MDM2, MDM4, and CDK4 amplification in glioblastoma . Oncoscience . 2015 ; 2 ( 7 ): 618 – 628 . Google Scholar CrossRef Search ADS PubMed 31. MolecularNeuropathology . Platform for next generation neuropathology . https://www.molecularneuropathology.org/mnp. 32. Sturm D , Witt H , Hovestadt V , et al. Hotspot mutations in H3F3A and IDH1 define distinct epigenetic and biological subgroups of glioblastoma . Cancer Cell . 2012 ; 22 ( 4 ): 425 – 437 . Google Scholar CrossRef Search ADS PubMed 33. Sturm D , Orr BA , Toprak UH , et al. New brain tumor entities emerge from molecular classification of CNS-PNETs . Cell . 2016 ; 164 ( 5 ): 1060 – 1072 . Google Scholar CrossRef Search ADS PubMed 34. Verhaak RG , Hoadley KA , Purdom E , et al. ; Cancer Genome Atlas Research Network . Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1 . Cancer Cell . 2010 ; 17 ( 1 ): 98 – 110 . Google Scholar CrossRef Search ADS PubMed 35. Felsberg J , Rapp M , Loeser S , et al. Prognostic significance of molecular markers and extent of resection in primary glioblastoma patients . Clin Cancer Res . 2009 ; 15 ( 21 ): 6683 – 6693 . Google Scholar CrossRef Search ADS PubMed 36. Knobbe CB , Reifenberger J , Reifenberger G . Mutation analysis of the Ras pathway genes NRAS, HRAS, KRAS and BRAF in glioblastomas . Acta Neuropathol . 2004 ; 108 ( 6 ): 467 – 470 . Google Scholar CrossRef Search ADS PubMed 37. Behling F , Barrantes-Freer A , Skardelly M , et al. Frequency of BRAF V600E mutations in 969 central nervous system neoplasms . Diagn Pathol . 2016 ; 11 ( 1 ): 55 . Google Scholar CrossRef Search ADS PubMed 38. Hatae R , Hata N , Suzuki SO , et al. A comprehensive analysis identifies BRAF hotspot mutations associated with gliomas with peculiar epithelial morphology . Neuropathology . 2017 ; 37 ( 3 ): 191 – 199 . Google Scholar CrossRef Search ADS PubMed 39. Singh D , Chan JM , Zoppoli P , et al. Transforming fusions of FGFR and TACC genes in human glioblastoma . Science . 2012 ; 337 ( 6099 ): 1231 – 1235 . Google Scholar CrossRef Search ADS PubMed 40. Kessler T , Sahm F , Sadik A , et al. Molecular differences in IDH wildtype glioblastoma according to MGMT promoter methylation . Neuro Oncol . 2018 ; 20 ( 3 ): 367 – 379 . Google Scholar CrossRef Search ADS PubMed 41. Wick W , Fricke H , Junge K , et al. A phase II, randomized, study of weekly APG101+reirradiation versus reirradiation in progressive glioblastoma . Clin Cancer Res . 2014 ; 20 ( 24 ): 6304 – 6313 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Neuro-Oncology Oxford University Press

Feasibility of real-time molecular profiling for patients with newly diagnosed glioblastoma without MGMT promoter hypermethylation—the NCT Neuro Master Match (N2M2) pilot study

Loading next page...
 
/lp/ou_press/feasibility-of-real-time-molecular-profiling-for-patients-with-newly-O2SVtWTC2f
Publisher
Oxford University Press
Copyright
© The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1522-8517
eISSN
1523-5866
D.O.I.
10.1093/neuonc/nox216
Publisher site
See Article on Publisher Site

Abstract

Abstract Background O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status is a predictive biomarker in glioblastoma patients. Glioblastoma without hypermethylated MGMT promoter is largely resistant to treatment with temozolomide. These patients are in particular need of new treatment approaches, which are offered by biomarker-driven clinical trials with targeted drugs based on molecular characterization of individual tumors. Methods In preparation for an upcoming clinical study, a comprehensive molecular profiling approach was undertaken on tissues from 43 glioblastoma patients harboring an unmethylated MGMT promoter at diagnosis. The diagnostic pipeline covered various levels of molecular characteristics, including whole-exome sequencing, low-coverage whole-genome sequencing, RNA sequencing, as well as microarray-based gene expression profiling and DNA methylation arrays. Results Complex multilayer molecular diagnostics were feasible in this setting with a median turnaround time of 4–5 weeks from surgery to the molecular tumor board. In 35% of cases, potentially relevant therapeutic decisions were derived from the data. Alterations were most frequently found in receptor tyrosine kinases, members of the phosphoinositide 3-kinase/Akt/mechanistic target of rapamycin and mitogen-activated protein kinase pathway as well as cell cycle control and p53 regulation cascades. Individual tumors harbored clonal alterations such as oncogenic fusions of tyrosine kinases which constitute promising targets for targeted therapies. A prioritization algorithm is proposed to allocate patients with multiple targets to the potentially best treatment option. Conclusion With this feasibility study, a comprehensive molecular profiling approach for patients with newly diagnosed glioblastoma harboring an unmethylated MGMT promoter is presented. Analyses in this pilot cohort serve as a basis for trials based on targetable alterations and on the question of allocation of patients to the best treatment arm. glioblastoma, MGMT, N2M2, precision oncology Importance of the study The diagnosis of glioblastoma is associated with a poor prognosis—this holds especially true for patients with glioblastoma without MGMT hypermethylation showing resistance toward the standard-of-care treatment with temozolomide. With this study, we present a comprehensive approach for molecular diagnostics applicable at diagnosis in a clinically meaningful timeframe. Implementing methods to investigate various molecular levels reveals well-known alterations in glioblastoma, but also identifies rare promising targets for targeted therapies. The proposed prioritization algorithm serves as a basis for determining the most suitable targeted treatment for the individual patient in cases with multiple druggable targets. It also allows for planning and allocating patients to biomarker-driven clinical trials. Glioblastoma is the most frequent malignant brain tumor in adults, constituting 60%–75% of astrocytic tumors and ~15% of all intracranial tumors.1 The incidence is about 3–4 new cases per 100000 people/year.1 The prognosis for patients with glioblastoma is particularly unfavorable, with a median survival of ~15 months and 5-year survival rate of less than 10% in trial cohorts.1,2 The current standard-of-care (SOC) treatment comprises induction concomitant with radiotherapy combined with oral temozolomide, followed by a maintenance phase with six to twelve 28-day cycles of temozolomide.3 Temozolomide as an alkylating agent exerts its cytotoxic effect through the formation of DNA crosslinks which lead to DNA strand breaks. Alkylation at guanine O6 can be reversed by the DNA repair protein O6-methylguanine-DNA methyltransferase (MGMT)4,5—thereby preventing the formation of lethal crosslinks. In line with this, patients harboring tumors with a hypermethylated MGMT promoter and thus silencing of MGMT gene expression experience a significant survival benefit from temozolomide chemotherapy. In contrast, patients with an unmethylated MGMT promoter show only minimal response.6,7 Therefore, particularly the latter group is in specific need of new treatments. Frequent genetic alterations in glioblastoma affect receptor tyrosine kinases (RTKs), such as epidermal growth factor receptor (EGFR), platelet derived growth factor receptor alpha (PDGFRA), fibroblast growth factor receptor (FGFR), and MET proto oncogene RTK (MET).8EGFR is one of the most commonly altered genes, mainly affected by amplifications but also activated by somatic single nucleotide variants (SNVs)8 or rearrangements including constitutively active EGFR variant III (EGFRvIII).9 Although several therapeutic agents targeting EGFR and EGFRvIII have been investigated in clinical trials,10,11 no clinical benefit has been proven to result from addressing this pathway in glioblastoma. Downstream of RTKs, alterations in different members and regulators of the phosphoinositide 3-kinase (PI3K)/Akt/mechanistic target of rapamycin (mTOR) as well as the mitogen activated protein kinase (MAPK) signaling pathways are regularly found. Cell cycle regulation is altered in approximately 80% of glioblastoma,8,12 mainly due to cyclin-dependent kinase inhibitor (CDKN) 2A/B deletion or cyclin-dependent kinase (CDK) 4/6 amplification. Finally, TP53 mutations or mouse double minute (MDM) 2/4 alterations impact p53 signaling, senescence, and apoptosis in glioblastoma. Thus, the spectrum of molecular characteristics in glioblastoma comprises, at a relatively high frequency, genetic alterations which can be targeted by specific drugs. Therefore, clinical trials focusing on targeted therapies based on molecular characteristics of the individual tumor represent an attractive next step in the treatment of glioblastoma. However, glioblastomas display extensive intratumoral heterogeneity at the level of the genome, transcriptome, and epigenome.13 This heterogeneity, with subclones showing different characteristics, is one critical aspect to consider when planning biomarker-driven clinical trials. Several clinical studies including targeted agents in newly diagnosed glioblastoma have failed to demonstrate efficacy in unselected patient cohorts.12,14–16 Therefore, well-considered allocation of patients to clinical trials based on molecular characteristics of the tumor as well as necessary retrospective validation of potential biomarkers are essential in a clinical setting. The aim of this study was to evaluate the feasibility of timely, comprehensive molecular diagnostics and interpretation in a clinical setting. Our analysis of a prospective cohort of 43 matched tumor and germline samples using next-generation sequencing, gene-expression profiling, and DNA methylation analysis also serves as a basis for development of stratification algorithms to allocate patients to biomarker-driven clinical trials with targeted drugs. The applicability of a similar pipeline in a clinically meaningful manner has previously been shown in the INFORM pilot phase for high-risk relapsed pediatric malignancies,17 and we here aimed to demonstrate that such a model was transferable to a routine adult neuro-oncology setting at initial diagnosis. Materials and Methods Cohort The NCT Neuro Master Match (N2M2) pilot cohort includes 43 patients with newly diagnosed glioblastoma without MGMT promoter hypermethylation determined by pyrosequencing.18 Thirty-nine of them were diagnosed at the Heidelberg University Hospital between July 15, 2014 and November 11, 2016. Glioblastomas in 4 patients were first diagnosed in other hospitals in Germany (Koblenz, Mannheim, Karlsruhe, Munich) and tissue was sent to our institution for analysis. Patients provided informed consent concerning the use of their tissue samples for research purposes. The concept of the investigation was approved by the local ethics committee (no. 206/2005, University of Heidelberg, Medical Faculty Ethics Committee). Pyrosequencing for Determining MGMT Promoter Methylation Analysis of MGMT promoter methylation status through pyrosequencing was performed with the Therascreen MGMT Pyro Kit (Qiagen) according to the manufacturer’s instructions. Quantitative measurement of methylation in 4 cytosine-phosphate-guanine (CpG) sites in exon 1 of the MGMT gene was performed. The cutoff was set at 8%. Infinium HumanMethylation450 Bead Chip and MethylationEPIC Kits The Illumina Infinium HumanMethylation450 (450k) bead chip and MethylationEPIC kits were used to obtain the DNA methylation status at >450000 and >850000 CpG sites, respectively, according to the manufacturer’s instructions at the Genomics and Proteomics Core Facility of the German Cancer Research Center in Heidelberg, Germany. Samples were analyzed using the R (www.r-project.org) based methylation pipeline “ChAMP.”19 In brief, filtering was done for multihit sites, SNPs, and XY chromosome–related CpGs, then data were normalized with a Beta-Mixture Quantile–based method and analyzed for batch effects with a singular value decomposition algorithm. Batch effects related to the tissue used (formalin-fixed paraffin-embedded [FFPE] vs fresh frozen) were corrected using ComBat. MGMT promoter methylation status was determined by the algorithm of Bady et al.20 Next-Generation Sequencing and Gene Expression Profiling Comprehensive molecular profiling (low-coverage whole-genome sequencing [lc-WGS] and whole-exome sequencing [WES], RNA-sequencing, and gene expression microarray) was conducted as described before.17 Confirmation of specific copy number variation (CNV) was performed using fluorescence in situ hybridization (FISH). Furthermore, gene panel sequencing using an in-house gene panel comprising 130 neuro-oncology relevant genes was added as previously specified.21 A detailed description of the molecular profiling can be found in the Supplementary material. Comparison of Different Diagnostic Methods In order to address the extensive intratumoral heterogeneity known in glioblastoma, molecular diagnostic approaches should comprise different techniques and should be applied at sufficient depth to investigate events occurring at various levels of clonality. We additionally aimed to maximize the chance to identify important hits and potential heterogeneity of their abundance by analyzing different areas of the same tumor specimen. Samples collected for routine pathology were paraffin embedded and used for gene panel sequencing and DNA methylation profiling, whereas fresh frozen material was taken for the dedicated molecular workup of the study including WES. mTOR Immunohistochemistry Immunohistochemistry to detect phospho-mTOR was performed as described previously12 using a heat antigen retrieval procedure (citrate buffer) and the phospho-mTOR antibody (Ser-2448, #2976, Cell Signaling Technology) in the dilution 1:100 according to manufacturer’s recommendations. Algorithm for Potential Assignment to Targeted Therapies Assignment to potential therapies was performed based on the molecular profile of the glioblastoma samples. Patients with glioblastoma harboring one of the molecular characteristics given in Supplementary Table S1 were allocated to the respective hypothetical group. For patients with tumors harboring more than one of the described molecular characteristics, a prioritization assignment algorithm (Table 1) was developed to rank the targeted therapies with respective biomarkers to allocate the patient to that targeted therapy with the highest estimated chance of response. This algorithm is partly inspired by the INFORM algorithm17 and considers clinical and preclinical evidence for benefit of a specific drug-biomarker combination in glioblastoma or other types of cancer as well as the nature of the genetic alteration. Of note, for each target, one test was considered as the formal reference—ie, WES for BRAF V600E mutation (vemurafenib) and anaplastic lymphoma kinase (ALK) point mutation (alectinib); RNA-Seq for fibroblast growth factor receptor (FGFR)–transforming acidic coiled-coil protein (TACC) fusion (ADZ4547) and MET fusion (crizotinib); methylation array for CDK4/6 amplification (palbociclib), MDM2 amplification (idasanutlin), and sonic hedgehog (SHH) amplification (vismodegib); and immunohistochemistry for mTOR Ser-2448 phosphorylation (temsirolimus). We allocated the patients of the above-mentioned alterations to the group that matches the molecular profile of the tumor and has the highest rank as given in Supplementary Table S1. This list gives a theoretical rank and may vary in a real trial according to drug and trial arm availability. Patients without any of these characteristics cannot be assigned to a specific group and in case of the N2M222 clinical trial will undergo block randomization to treatments with either a programmed cell death ligand 1 (PD-L1) antibody, cluster of differentiation 95 ligand (CD95L) inhibitory recombinant protein, or temozolomide. Table 1 Prioritization algorithm for biomarker-based targeted treatment Group Criterion 1 Biomarker with approved biomarker specific treatment in glioblastoma + with strong survival benefit – with moderate survival benefit or inconsistent 2A Biomarker with approved biomarker specific treatment in another cancer indication with compelling clinical evidence in glioblastoma 2B Biomarker with approved biomarker specific treatment in another cancer indication not tested in glioblastoma in a clinical setting 3A Clinical evidence in glioblastoma, but not approved in glioblastoma or any other cancer indication + mutation – amp/expression 3B Clinical evidence in another cancer indication, makes biological sense in glioblastoma, but no clinical evidence in glioblastoma + mutation – amp/expression 4A No compelling clinical evidence in any cancer indication but retrospective biomarker assessment in + glioblastoma – another cancer indication 4B No compelling clinical evidence in any cancer indication, but stable genetic change (mutation/fusion) and makes biological sense in glioblastoma ± preclinical evidence 4C No compelling clinical evidence in any cancer indication, expression change, phosphorylation, etc. is a direct drug target and makes biological sense in glioblastoma 4D Alteration of pathways or genetic alteration/expression change regulates drug target Group Criterion 1 Biomarker with approved biomarker specific treatment in glioblastoma + with strong survival benefit – with moderate survival benefit or inconsistent 2A Biomarker with approved biomarker specific treatment in another cancer indication with compelling clinical evidence in glioblastoma 2B Biomarker with approved biomarker specific treatment in another cancer indication not tested in glioblastoma in a clinical setting 3A Clinical evidence in glioblastoma, but not approved in glioblastoma or any other cancer indication + mutation – amp/expression 3B Clinical evidence in another cancer indication, makes biological sense in glioblastoma, but no clinical evidence in glioblastoma + mutation – amp/expression 4A No compelling clinical evidence in any cancer indication but retrospective biomarker assessment in + glioblastoma – another cancer indication 4B No compelling clinical evidence in any cancer indication, but stable genetic change (mutation/fusion) and makes biological sense in glioblastoma ± preclinical evidence 4C No compelling clinical evidence in any cancer indication, expression change, phosphorylation, etc. is a direct drug target and makes biological sense in glioblastoma 4D Alteration of pathways or genetic alteration/expression change regulates drug target View Large Table 1 Prioritization algorithm for biomarker-based targeted treatment Group Criterion 1 Biomarker with approved biomarker specific treatment in glioblastoma + with strong survival benefit – with moderate survival benefit or inconsistent 2A Biomarker with approved biomarker specific treatment in another cancer indication with compelling clinical evidence in glioblastoma 2B Biomarker with approved biomarker specific treatment in another cancer indication not tested in glioblastoma in a clinical setting 3A Clinical evidence in glioblastoma, but not approved in glioblastoma or any other cancer indication + mutation – amp/expression 3B Clinical evidence in another cancer indication, makes biological sense in glioblastoma, but no clinical evidence in glioblastoma + mutation – amp/expression 4A No compelling clinical evidence in any cancer indication but retrospective biomarker assessment in + glioblastoma – another cancer indication 4B No compelling clinical evidence in any cancer indication, but stable genetic change (mutation/fusion) and makes biological sense in glioblastoma ± preclinical evidence 4C No compelling clinical evidence in any cancer indication, expression change, phosphorylation, etc. is a direct drug target and makes biological sense in glioblastoma 4D Alteration of pathways or genetic alteration/expression change regulates drug target Group Criterion 1 Biomarker with approved biomarker specific treatment in glioblastoma + with strong survival benefit – with moderate survival benefit or inconsistent 2A Biomarker with approved biomarker specific treatment in another cancer indication with compelling clinical evidence in glioblastoma 2B Biomarker with approved biomarker specific treatment in another cancer indication not tested in glioblastoma in a clinical setting 3A Clinical evidence in glioblastoma, but not approved in glioblastoma or any other cancer indication + mutation – amp/expression 3B Clinical evidence in another cancer indication, makes biological sense in glioblastoma, but no clinical evidence in glioblastoma + mutation – amp/expression 4A No compelling clinical evidence in any cancer indication but retrospective biomarker assessment in + glioblastoma – another cancer indication 4B No compelling clinical evidence in any cancer indication, but stable genetic change (mutation/fusion) and makes biological sense in glioblastoma ± preclinical evidence 4C No compelling clinical evidence in any cancer indication, expression change, phosphorylation, etc. is a direct drug target and makes biological sense in glioblastoma 4D Alteration of pathways or genetic alteration/expression change regulates drug target View Large Results Clinical Characteristics of the Patient Cohort The N2M2 pilot cohort consisted of 43 patients with newly diagnosed glioblastoma. Inclusion was based on MGMT promoter methylation status determined by pyrosequencing and IDH1/2 wild-type status in all 43 cases. Twenty-one (49%) women and 22 (51%) men were included with an average age at diagnosis of 62.8 years (Table 2). Most patients had an Eastern Cooperative Oncology Group performance score of 0 (63%), and 19 (44%) patients were on steroid treatment at the time of diagnosis. Follow-up data on the primary therapy was available for 40 (93%) patients. Twenty-eight (65%) received primary radiotherapy with temozolomide, 11 patients (26%) were treated with radiotherapy only. One patient had not started further therapy at the time of analysis (Table 2 and Supplementary Table S2). Table 2. Clinical characteristics of the patient cohort Clinical Characteristics Pilot Phase Cohort (n = 43) Sex, no. (%)  Female 21 (48.8)  Male 22 (51.2) Age, no. (%)  <60 y 16 (37.2)  60–69 y 12 (27.9)  ≥70 y 15 (34.9) Age, mean (SD) 62.8 (11.5) ECOG, no. (%)  0 27 (62.8)  1 8 (18.7)  ≥2 8 (18.7) Steroids at baseline, no. (%)  Yes 19 (44.2)  No 23 (53.5)  Missing data 1 (2.3) Localization, no. (%)  Left hemisphere 26 (60.4)  Right hemisphere 14 (32.6)  Corpus callosum 1 (2.3)  Cerebellar 1 (2.3)  Intramedular 1 (2.3) Primary therapy, no. (%)  RT + TMZ 28 (65.1)  RT alone 11 (25.6)  TMZ alone 0 (0.0)  Other 0 (0.0)  None 1 (2.3)  Missing data 3 (7.0) Clinical Characteristics Pilot Phase Cohort (n = 43) Sex, no. (%)  Female 21 (48.8)  Male 22 (51.2) Age, no. (%)  <60 y 16 (37.2)  60–69 y 12 (27.9)  ≥70 y 15 (34.9) Age, mean (SD) 62.8 (11.5) ECOG, no. (%)  0 27 (62.8)  1 8 (18.7)  ≥2 8 (18.7) Steroids at baseline, no. (%)  Yes 19 (44.2)  No 23 (53.5)  Missing data 1 (2.3) Localization, no. (%)  Left hemisphere 26 (60.4)  Right hemisphere 14 (32.6)  Corpus callosum 1 (2.3)  Cerebellar 1 (2.3)  Intramedular 1 (2.3) Primary therapy, no. (%)  RT + TMZ 28 (65.1)  RT alone 11 (25.6)  TMZ alone 0 (0.0)  Other 0 (0.0)  None 1 (2.3)  Missing data 3 (7.0) Abbreviations: ECOG, Eastern Cooperative Oncology Group; RT, radiotherapy; TMZ, temozolomide. View Large Table 2. Clinical characteristics of the patient cohort Clinical Characteristics Pilot Phase Cohort (n = 43) Sex, no. (%)  Female 21 (48.8)  Male 22 (51.2) Age, no. (%)  <60 y 16 (37.2)  60–69 y 12 (27.9)  ≥70 y 15 (34.9) Age, mean (SD) 62.8 (11.5) ECOG, no. (%)  0 27 (62.8)  1 8 (18.7)  ≥2 8 (18.7) Steroids at baseline, no. (%)  Yes 19 (44.2)  No 23 (53.5)  Missing data 1 (2.3) Localization, no. (%)  Left hemisphere 26 (60.4)  Right hemisphere 14 (32.6)  Corpus callosum 1 (2.3)  Cerebellar 1 (2.3)  Intramedular 1 (2.3) Primary therapy, no. (%)  RT + TMZ 28 (65.1)  RT alone 11 (25.6)  TMZ alone 0 (0.0)  Other 0 (0.0)  None 1 (2.3)  Missing data 3 (7.0) Clinical Characteristics Pilot Phase Cohort (n = 43) Sex, no. (%)  Female 21 (48.8)  Male 22 (51.2) Age, no. (%)  <60 y 16 (37.2)  60–69 y 12 (27.9)  ≥70 y 15 (34.9) Age, mean (SD) 62.8 (11.5) ECOG, no. (%)  0 27 (62.8)  1 8 (18.7)  ≥2 8 (18.7) Steroids at baseline, no. (%)  Yes 19 (44.2)  No 23 (53.5)  Missing data 1 (2.3) Localization, no. (%)  Left hemisphere 26 (60.4)  Right hemisphere 14 (32.6)  Corpus callosum 1 (2.3)  Cerebellar 1 (2.3)  Intramedular 1 (2.3) Primary therapy, no. (%)  RT + TMZ 28 (65.1)  RT alone 11 (25.6)  TMZ alone 0 (0.0)  Other 0 (0.0)  None 1 (2.3)  Missing data 3 (7.0) Abbreviations: ECOG, Eastern Cooperative Oncology Group; RT, radiotherapy; TMZ, temozolomide. View Large Timelines One of the aims of this feasibility study was to determine whether a comprehensive molecular profiling could be performed in the clinically demanded timeframe of 6 weeks from surgery. In 39 of 43 (90.7%) cases, surgery was performed locally at the Heidelberg University Hospital, and material arrived in the neuropathology laboratory at the day of operation. Four patients were operated on in external centers and tissue was subsequently analyzed at the Department of Neuropathology at the Heidelberg University Hospital. For timeline calculations, only prospectively included cases (n = 35) were considered. The median time interval for sample processing before submission to pyrosequencing was 4 days (range 0–21). Pyrosequencing and DNA methylation array analysis were completed within a median of 3.5 days (range 0–25) and 26 days (range 8–84), respectively. WES, WGS, and microarray gene expression profile were performed within a median of 12 days (range 8–33) and 9 days (range 7–16), respectively. Panel sequencing was completed within a median of 23 days (range 7–42). Bioinformatic processing of sequencing data required 2 days as median (range 0–12) and data interpretation could be completed within 2–3 days (Fig. 1A). Thus, when running as a routine part of a clinical study, the whole pipeline, including clinical decision for study arm allocation, is feasible within 4–5 weeks (Fig. 1B). This also corresponds to our experience in now more than 300 patients prospectively enrolled in the INFORM23 study. Fig. 1 View largeDownload slide Schematic illustration of the workflow and timeline of the N2M2 molecular diagnostic pipeline. (A) The real median time for each step of the process is calculated for the prospectively included cases, range of time intervals is given in brackets. (B) Ideal time intervals for prospectively analyzed cases with the whole molecular diagnostics pipeline being performed within 4 weeks from day of tissue arrival to discussion in the molecular tumorboard. Fig. 1 View largeDownload slide Schematic illustration of the workflow and timeline of the N2M2 molecular diagnostic pipeline. (A) The real median time for each step of the process is calculated for the prospectively included cases, range of time intervals is given in brackets. (B) Ideal time intervals for prospectively analyzed cases with the whole molecular diagnostics pipeline being performed within 4 weeks from day of tissue arrival to discussion in the molecular tumorboard. MGMT Methylation Status: Pyrosequencing versus 450k/850k Methylation Array Determination of MGMT status for inclusion in the present study was based on pyrosequencing, which is considered the gold standard.18,24 Prior studies of our group have suggested discordance in different methods for MGMT promoter methylation testing.25 In 86.0% (37/43) of cases, the pyrosequencing and DNA methylation analysis results were consistent. Two cases showed an unmethylated MGMT promoter based on pyrosequencing (4% and 6% methylation, respectively), but the MGMT promoter was methylated based on DNA methylation array. Four cases were not classifiable with sufficient confidence to either methylated or unmethylated based on DNA methylation array. Notably, 2 distinct CpG sites are analyzed for MGMT promoter methylation analysis using the Infinium 450k bead chip and MethylationEPIC kits, respectively,20 which differ from the CpGs in the pyrosequencing approach. Single Nucleotide Variants: WES versus Gene Panel The mean count of SNVs was 47.7 per exome (range 19–80) based on WES, in line with previous reports of adult glioblastoma26 and higher than in pediatric high-grade glioma.27 To explore intratumoral heterogeneity, targeted sequencing using an in-house panel of 130 neuro-oncology–relevant genes (Panel Seq)21 was performed on FFPE material and compared with WES conducted on fresh frozen material from a different region of the same tumor. For 36 cases (84%) gene panel data were available from FFPE material, whereas for the remaining 7 cases, Panel Seq was performed on the same fresh frozen material as WES. Alterations detected in these 7 cases were completely consistent in both methods. In the group of cases with sequencing data from different material types (FFPE and frozen), we altogether detected 69 alterations (SNVs and insertions/deletions), which we would consider as either potentially druggable targets or tumor biologically interesting findings (Fig. 2). In total, 74% of alterations (48/69) were detected by both methods. One case harbored 2 different nucleotide substitutions within the phosphatase and tensin homolog (PTEN) gene, each of the SNVs revealed by one of the 2 methods, respectively. Three point mutations were detected by WES in genes which were not covered in the gene panel. Fig. 2 View largeDownload slide Overview of selected genetic alterations for each case, subgroup affiliation based on DNA methylation as well as expression analysis and mTOR staining results. Fig. 2 View largeDownload slide Overview of selected genetic alterations for each case, subgroup affiliation based on DNA methylation as well as expression analysis and mTOR staining results. Of the 25% of alterations that were found by only one method, 5 of 69 were high allele frequency mutations, including one alteration found only by WES (MDM2) and not in the Panel Seq, and 4 alterations found only by Panel Seq and not WES (phosphoinositide -4,5-bisphosphate 3-kinase catalytic subunit alpha [PIK3CA], 2x EGFR, FGFR4), respectively. For these alterations, a technical reason such as poor coverage or very low allele frequency was ruled out, therefore the discrepancy points to intratumoral heterogeneity. This is underlined by the fact that no discrepant alterations were found in the cases with sequencing data from the same material. Another 9% of alterations could be detected by only one of the methods due to a low variant allele frequency (of lower than 10%). The remaining 9% of discrepancies were due to differences in the bioinformatic pipelines used. Therefore, the pipelines are now being further optimized based on these results. Copy Number Variations: Low-Coverage WGS versus 450k/850k Copy number plots obtained from lc-WGS were compared with those obtained from methylation array analysis for 36 cases. The overlap between both methods accounted for 88% (77/88) of alterations considered to be potential drug targets and/or tumor-biologically relevant (Fig. 2). Four of 88 druggable alterations could be found only in lc-WGS data, whereas 7 of 88 were detected in copy number plots from methylation array but were not observed in lc-WGS. The likely explanation for these discrepancies is again assumed to be intratumoral heterogeneity, since the relevant aberrations were of a type and size that should be detectable by both methods. One case (Pilot_GBM_26)—besides harboring CDK4 and MDM2 amplification (detected with both methods)—showed EGFR amplification in the lc-WGS data only and PDGFRA amplification in the methylation analysis only. Both alterations were high-level focal amplifications, but each was observed by only one of the 2 methods. EGFR amplification was validated by FISH in the piece used for lc-WGS but was not detectable in the piece for methylation analysis (Supplementary Figure S1). The tissue pieces taken for methylation analysis and lc-WGS therefore likely showed a strong enrichment for one of these 2 subclones, to the extent that the minor population was below the level of detection. Mosaic amplification of RTKs (ie, EGFR and PDGFRA) was previously described in glioblastoma with distinct tumor cell populations harboring either EGFR or PDGFRA amplification.28 The phenomenon of chromothripsis as a single catastrophic event leading to multiple chromosomal rearrangements was also described in glioblastoma,29 especially linked to MDM2/4, CDK4, and EGFR amplification.30 Three cases in the cohort investigated showed chromothripsis of 1–2 chromosomes/chromosome arms. Notably, 2 of the tumors exhibited concomitant CDK4 and MDM2 amplification, whereas the third tumor was EGFR amplified (Supplementary Figure S2). Assignment to Previously Described Subgroups Based on DNA Methylation Analysis and Gene Expression Profiling Using an in-house classifier scoring algorithm31 based on genome-wide DNA methylation patterns, samples could be allocated to previously established methylation subgroups.32 Of all cases, 88.4% (38/43) fall into the glioblastoma methylation groups of the mesenchymal, RTK I, or RTK II subtypes (Fig. 2). For 2 samples (pilot_GBM_22 and pilot_GBM_32) the methylation profile was most similar to pediatric glioblastoma groups (one to the pediatric RTK subtype and one to a methylation group enriched for MYCN amplification,33 respectively). One tumor (pilot_GBM_41) had a high leukocyte infiltration reflected by the highest score for the glioblastoma methylation group with high leukocyte infiltration. This indicates a low tumor cell content (in line with low variant allele frequency and rather flat copy number plot in this case), which also needs to be considered in such personalized medicine efforts. Two cases (pilot_GBM_2, pilot_GBM_13) could not be allocated clearly to one specific subgroup, but rather to a general group of high-grade gliomas. Samples were furthermore assigned to the proneural, classical, and mesenchymal expression groups of Verhaak et al (Fig. 2, Supplementary Figure S3).34 The concordance between the methylation and expression based groups was relatively low. Frequently Affected Oncogenic Pathways The most frequently altered pathways in the cohort investigated were (receptor) tyrosine kinase signaling (RTK), PI3K/Akt/mTOR pathway, MAPK pathway, cell cycle control, and TP53 regulation. Frequencies of alterations in these pathways as well as in individual pathway members are depicted in Fig. 3. Fig. 3 View largeDownload slide Alterations of well-established oncogenic pathways. Frequencies of alterations are specifically illustrated for selected members of the single pathways. Only genetic alterations (SNV, indel, CNV, fusion) are considered. Fig. 3 View largeDownload slide Alterations of well-established oncogenic pathways. Frequencies of alterations are specifically illustrated for selected members of the single pathways. Only genetic alterations (SNV, indel, CNV, fusion) are considered. Amplifications and mutations of the EGFR gene were the most commonly occurring alterations of RTKs (18/43; 42%), followed by PDGFRA alterations (5/43; 12%). Less frequently affected tyrosine kinases included FGFR3/4, VEGFR, MET, and NTRK3. Taken together, 61% (26/43) of cases showed a genetic alteration in at least one RTK, with 6 cases harboring alterations in 2 or 3 different RTKs. Regarding alterations of the PI3K/Akt/mTOR pathway, PTEN loss-of-function mutations and deletions were most frequent (24/43; 56%). To a lesser extent, PIK3CA/B and PIK3R1 mutations were also observed (10/43; 23%). Overall, genetic alterations of the PI3K/Akt/mTOR pathway were found in 70% (30/43) of cases. Alterations affecting members of the MAPK pathway (downstream of RTKs) were detected in 21% (9/43), partly overlapping with RTK alterations. Dysregulation of cell cycle control was mostly due to CDKN2A/B deletion (occurring in 29/43 cases; 67%) and at a lower frequency due to CDK4 amplification (8/43 cases; 19%). No CDK6 amplifications were observed. Altogether, only 14% of cases showed no obvious alterations of cell cycle control mechanisms (6/43). TP53 gene mutations were detected in 12 of 43 (28%) cases, in line with previous reports.8,35 Amplifications of the p53 regulators MDM2 and MDM4 were each found in 9% (4/43) of cases. None of these cases harbored a concomitant TP53 mutation. MDM2 overexpression, defined as reads per kilobase of exon model per million mapped reads >50 based on RNA-Seq data, correlated with gene amplification in 3 of 4 cases. One further case with high MDM2 expression showed no gene amplification. Less frequent genetic alterations occurring in single cases included genes involved in transcription, chromatin remodeling, DNA repair, Notch and Wnt signaling pathway, as well as telomerase maintenance. mTOR Staining A subset of patients with MGMT unmethylated glioblastoma showing phosphorylation of mTOR at Ser-2448 may have some benefit from mTOR inhibitor treatment.12 Therefore, 25 samples with material available were stained for mTOR Ser-2448. Four of 25 cases (16%) were evaluated as “mTOR positive” with 100% of cells showing a score of ≥2 (see Fig. 2 and Supplementary Figure S4). Alterations as Promising Targets for Biomarker-Driven Therapies The present study does not constitute an interventional trial, nor have patients been assigned to any particular therapy based on these data. However, hypothetically we could identify alterations that are promising druggable targets in glioblastoma in 15 of 43 (35%) cases (Fig. 4). Given the parallel design of an interventional trial, N2M2,22 we have tested available arms and found matches for the MDM2 inhibitor idasanutlin, the CDK4/6 inhibitor palbociclib, and the mTOR inhibitor temsirolimus in 11 of 43 (26%) cases. Eight patients would fit into the palbociclib group, 4 patients into the idasanutlin group, and 4 patients into the temsirolimus group, including 5 patients fitting into 2 of these groups. Thirty-two of 43 (74%) patients had no biomarker match for this trial, but 4 cases had a further promising biomarker not currently planned to be included in the trial design. Fig. 4 View largeDownload slide Flowchart of trial arm allocation according to matching biomarkers. Flowchart shows possible allocation of patients to different trials or randomization according to druggable targets detected. *According to evidence level, **best matching target or where recruitment possible; ITT: intention to treat. Fig. 4 View largeDownload slide Flowchart of trial arm allocation according to matching biomarkers. Flowchart shows possible allocation of patients to different trials or randomization according to druggable targets detected. *According to evidence level, **best matching target or where recruitment possible; ITT: intention to treat. Of these, a hotspot BRAF V600E mutation was found. The overall incidence of this mutation in adult glioblastoma is considered relatively low.36–38 Single case reports describe a response to BRAF inhibitor treatment in pediatric patients with high-grade glioma.36,37 Clinical trials applying BRAF inhibitors in glioma patients are planned or ongoing (eg, ClinicalTrials.gov Identifier: NCT02684058). Two cases harbored FGFR3-TACC3 fusions. Oncogenic translocations fusing the kinase domain of FGFR1 or FGFR3 to TACC1 or TACC3, respectively, are described in ~3% of glioblastoma and were shown to be sensitive to FGFR inhibitor treatment in in vivo models.39 Fusion of MET and protein tyrosine phosphatase receptor type Z1 could be detected in one further case. Rearrangements affecting MET were described as being oncogenic and a promising drug target in glioblastoma for treatment with MET inhibitors.23,39 For alteration patterns of the tumors, see Fig. 2. The N2M2 clinical trial furthermore aims to include patients also based on the rare genetic alterations ALK fusion/point mutation for treatment with the ALK inhibitor alectinib and on SHH amplification to the SHH inhibitor vismodegib. No such alterations have been found in the 43 patients within the pilot cohort. We furthermore compared targetable alterations in our cohort with the previously described cohort from The Cancer Genome Atlas (TCGA). Frequencies of the alterations which were also assessed in TCGA were similar in both cohorts (Supplementary Table S3). Prioritization Algorithm for Patients with Multiple Potential Biomarkers Six of 43 (14%) tumors harbored more than one promising target specifically consisting of the overlap between CDK4 and MDM2 amplifications as well as mTOR phosphorylation. This constitutes a challenging group, as a decision must be made on the best target. The refined prioritization algorithm ranked potentially targetable alterations regardless of N2M2 trial availability together with the corresponding available treatments according to (i) the evidence level for benefit of patients harboring the respective biomarker, (ii) whether this evidence comes from glioblastoma or other cancer entities, and (iii) type and stability of the genetic alteration (Table 1, Supplementary Table S1). In the starting clinical trial, patients will be allocated to the best matching N2M2 arm. If high confidence targets could be detected, but the compound arm for targeting this alteration is currently not available in N2M2, inclusion into other appropriate studies based on the molecular aberration will be considered. Patients without a matching biomarker will undergo block randomization between a PD-L1 antibody, a CD95L inhibitory recombinant protein, and SOC temozolomide (Fig. 4). Discussion In contrast to the growing knowledge about molecular characteristics and tumorigenesis of glioblastoma during the last few years, progress in clinical management is still limited. Several clinical studies evaluating targeted agents in unselected patient populations in newly diagnosed glioblastoma have failed to demonstrate efficacy.12,14–16 Therefore, well-considered allocation of patients to clinical trials based on molecular characteristics of the individual tumor is required and critical evaluation of potential biomarkers in a clinical set is essential for further progress. This especially applies for the subgroup of MGMT unmethylated glioblastoma poorly responding to treatment with temozolomide. Inclusion criteria of current and planned studies are increasingly based on MGMT promoter methylation. As these studies may withhold temozolomide in at least one study arm for the poorly responding MGMT unmethylated patients, accurate determination of MGMT promoter methylation status is crucial to avoid withholding temozolomide in patients who might benefit from this drug. In our study, 6 of 43 (14%) tumors with an unmethylated MGMT promoter determined by pyrosequencing and data available from both methods were classified as methylated or were unclassifiable with array-based methods. The discrepancy between the 2 methods can likely be explained by the differing CpG regions assessed in the 2 approaches, as noted above. It is therefore possible that these discrepancies represent true differences in methylation status depending on the precise CpGs analyzed, and it is important to keep this in mind for future stratification. We recently suggested that the response to temozolomide is more complex, and defining the right subgroups that do not benefit may also involve global methylation profiles and status of telomerase reverse transcriptase.40 For the N2M2 clinical trial, we will use pyrosequencing as the currently accepted gold standard to determine MGMT promoter status that defines the inclusion criteria. Patient samples with an unmethylated MGMT promoter defined by methylation array will be validated with pyrosequencing, and patients with MGMT methylation in either method excluded. Nevertheless, further research is needed, especially regarding those patients with discrepant MGMT promoter methylation. With this feasibility study, we present a comprehensive approach including various techniques to allow for an in-depth molecular characterization of each individual tumor. For applicability in a clinical setting, a reasonable timeframe is crucial. In the N2M2 pilot study presented here, molecular characterization was performed within 4–5 weeks after tumor operation, allowing timely decision making for further therapy. In line with previous large-scale studies,8 we could demonstrate that several biologically relevant pathways are affected in the investigated glioblastoma cohort, and in many patients at least one potential treatment biomarker was identified. Since targeted approaches are under consideration for clinical trials in glioblastoma, molecular profiling at diagnosis can serve as a basis for allocation of patients to biomarker-driven clinical trials. One example is the upcoming N2M2 clinical trial that includes patients with an unmethylated MGMT promoter and allocates these patients according to their molecular profile to different treatment arms. We virtually assigned the patients from this pilot cohort into the best fitting treatment arm. In the upcoming trial, patients without matching biomarkers will be randomized between a PD-L1 antibody, a CD95L inhibitory recombinant protein, and SOC temozolomide. For CD95L inhibition, a promising biomarker, low CD95L promoter methylation,41 was present in 18 of 43 (42%) patients. Together, this shows that reasonable biomarkers for the treatment groups are present in a large subset of patients. N2M2 provides a useful structure to test the translation of molecular diagnostics into clinical decisions as well as safety and efficacy evaluations. One problem is the allocation of patients with tumors harboring 2 or more potentially druggable targets to the best specific treatment. Since combined treatments often result in severe side effects and cannot be analyzed in clinical trials due to their complexity, we developed a prioritization algorithm based on clinical and preclinical evidence in glioblastoma and other cancer indications. The algorithm yields the most promising biomarker and targeted treatment for patients with more than one potential target. It can be easily adapted in competing situations with new potential biomarkers evolving in the future or shifting evidence on existing biomarkers according to clinical trial results. Chromothripsis has been observed in 2 patients in the present study. It is questionable whether patients with such large numbers of genetic alterations related to a single event should be excluded from study entry. However, genomic instability may also drive senescence and may enhance dependence on growth-promoting pathways, therefore we plan not to exclude such patients, but to specifically follow up on the disease course. Certainly, inter- as well as intratumoral heterogeneity remain an extreme challenge for the design of biomarker-driven clinical trials, and the full spectrum is not addressed in this study. Several examples are presented which suggest that parallel, complementary analyses from different tumor pieces may be a rational strategy. Future studies will also have to take into account that biomarkers might be unevenly distributed both spatially and temporally within one and between different tumors, which is an important consideration for maximizing the predictive value of biomarkers. A further practical problem is the low frequency of some of the alterations. N2M2 will assess each subtrial individually and only use the temozolomide subtrial as a contemporary reassurance for the validity of the assumptions of progression-free survival. As a next step, randomized, controlled trials will be performed for successful arms with high-frequency molecular lesions like the CDK4/6 amplification (CDKN2A loss) or the mTOR Ser-2448 phosphorylation or for yet-to-be-discovered parameters for the APG101 or atezolizumab subtrials. The lesions with lower frequency will need more time for accrual, a preselection, or different efficacy assessments like objective response rates. In summary, with this feasibility study, we present a comprehensive molecular profiling approach for glioblastoma without MGMT promoter hypermethylation, which can be applied at diagnosis in a clinically relevant timeframe. Knowing the molecular characteristics of the individual tumor allows for allocation to respective biomarker-driven clinical trials as well as retrospective identification of promising biomarkers to hopefully advance the pace of therapeutic progress for glioblastoma patients. Supplementary Material Supplementary material is available at Neuro-Oncology online. Funding We thank the DKFZ-Heidelberg Center for Personalized Oncology (DKFZ-HIPO) for technical support and funding through HIPO project numbers H57 and K25. Conflict of interest statement. No conflicts disclosed. References 1. Ostrom QT , Gittleman H , Fulop J , et al. CBTRUS Statistical Report: primary brain and central nervous system tumors diagnosed in the United States in 2008–2012 . Neuro Oncol . 2015 ; 17 ( Suppl 4 ): iv1 – iv62 . Google Scholar CrossRef Search ADS PubMed 2. Stupp R , Hegi ME , Mason WP , et al. ; European Organisation for Research and Treatment of Cancer Brain Tumour and Radiation Oncology Groups; National Cancer Institute of Canada Clinical Trials Group . Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial . Lancet Oncol . 2009 ; 10 ( 5 ): 459 – 466 . Google Scholar CrossRef Search ADS PubMed 3. Stupp R , Mason WP , van den Bent MJ , et al. ; European Organisation for Research and Treatment of Cancer Brain Tumor and Radiotherapy Groups; National Cancer Institute of Canada Clinical Trials Group . Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma . N Engl J Med . 2005 ; 352 ( 10 ): 987 – 996 . Google Scholar CrossRef Search ADS PubMed 4. Pegg AE , Dolan ME , Moschel RC . Structure, function, and inhibition of O6-alkylguanine-DNA alkyltransferase . Prog Nucleic Acid Res Mol Biol . 1995 ; 51 : 167 – 223 . Google Scholar CrossRef Search ADS PubMed 5. Ludlum DB . DNA alkylation by the haloethylnitrosoureas: nature of modifications produced and their enzymatic repair or removal . Mutat Res . 1990 ; 233 ( 1-2 ): 117 – 126 . Google Scholar CrossRef Search ADS PubMed 6. Esteller M , Garcia-Foncillas J , Andion E , et al. Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents . N Engl J Med . 2000 ; 343 ( 19 ): 1350 – 1354 . Google Scholar CrossRef Search ADS PubMed 7. Hegi ME , Diserens AC , Gorlia T , et al. MGMT gene silencing and benefit from temozolomide in glioblastoma . N Engl J Med . 2005 ; 352 ( 10 ): 997 – 1003 . Google Scholar CrossRef Search ADS PubMed 8. Brennan CW , Verhaak RG , McKenna A , et al. ; TCGA Research Network . The somatic genomic landscape of glioblastoma . Cell . 2013 ; 155 ( 2 ): 462 – 477 . Google Scholar CrossRef Search ADS PubMed 9. Batra SK , Castelino-Prabhu S , Wikstrand CJ , et al. Epidermal growth factor ligand-independent, unregulated, cell-transforming potential of a naturally occurring human mutant EGFRvIII gene . Cell Growth Differ . 1995 ; 6 ( 10 ): 1251 – 1259 . Google Scholar PubMed 10. Gan HK , Kaye AH , Luwor RB . The EGFRvIII variant in glioblastoma multiforme . J Clin Neurosci . 2009 ; 16 ( 6 ): 748 – 754 . Google Scholar CrossRef Search ADS PubMed 11. Weller M , Butowski N , Tran DD , et al. ; ACT IV trial investigators . Rindopepimut with temozolomide for patients with newly diagnosed, EGFRvIII-expressing glioblastoma (ACT IV): a randomised, double-blind, international phase 3 trial . Lancet Oncol . 2017 ; 18 ( 10 ): 1373 – 1385 . Google Scholar CrossRef Search ADS PubMed 12. Wick W , Gorlia T , Bady P , et al. Phase II study of radiotherapy and temsirolimus versus radiochemotherapy with temozolomide in patients with newly diagnosed glioblastoma without MGMT promoter hypermethylation (EORTC 26082) . Clin Cancer Res . 2016 ; 22 ( 19 ): 4797 – 4806 . Google Scholar CrossRef Search ADS PubMed 13. Bonavia R , Inda MM , Cavenee WK , Furnari FB . Heterogeneity maintenance in glioblastoma: a social network . Cancer Res . 2011 ; 71 ( 12 ): 4055 – 4060 . Google Scholar CrossRef Search ADS PubMed 14. Herrlinger U , Schäfer N , Steinbach JP , et al. Bevacizumab plus irinotecan versus temozolomide in newly diagnosed O6-methylguanine-DNA methyltransferase nonmethylated glioblastoma: the randomized GLARIUS trial . J Clin Oncol . 2016 ; 34 ( 14 ): 1611 – 1619 . Google Scholar CrossRef Search ADS PubMed 15. Nabors LB , Fink KL , Mikkelsen T , et al. Two cilengitide regimens in combination with standard treatment for patients with newly diagnosed glioblastoma and unmethylated MGMT gene promoter: results of the open-label, controlled, randomized phase II CORE study . Neuro Oncol . 2015 ; 17 ( 5 ): 708 – 717 . Google Scholar CrossRef Search ADS PubMed 16. Wick W , Steinbach JP , Platten M , et al. Enzastaurin before and concomitant with radiation therapy, followed by enzastaurin maintenance therapy, in patients with newly diagnosed glioblastoma without MGMT promoter hypermethylation . Neuro Oncol . 2013 ; 15 ( 10 ): 1405 – 1412 . Google Scholar CrossRef Search ADS PubMed 17. Worst BC , van Tilburg CM , Balasubramanian GP , et al. Next-generation personalised medicine for high-risk paediatric cancer patients—the INFORM pilot study . Eur J Cancer . 2016 ; 65 : 91 – 101 . Google Scholar CrossRef Search ADS PubMed 18. Mikeska T , Bock C , El-Maarri O , et al. Optimization of quantitative MGMT promoter methylation analysis using pyrosequencing and combined bisulfite restriction analysis . J Mol Diagn . 2007 ; 9 ( 3 ): 368 – 381 . Google Scholar CrossRef Search ADS PubMed 19. Morris TJ , Butcher LM , Feber A , et al. ChAMP: 450k chip analysis methylation pipeline . Bioinformatics . 2014 ; 30 ( 3 ): 428 – 430 . Google Scholar CrossRef Search ADS PubMed 20. Bady P , Sciuscio D , Diserens AC , et al. MGMT methylation analysis of glioblastoma on the Infinium methylation BeadChip identifies two distinct CpG regions associated with gene silencing and outcome, yielding a prediction model for comparisons across datasets, tumor grades, and CIMP-status . Acta Neuropathol . 2012 ; 124 ( 4 ): 547 – 560 . Google Scholar CrossRef Search ADS PubMed 21. Sahm F , Schrimpf D , Jones DT , et al. Next-generation sequencing in routine brain tumor diagnostics enables an integrated diagnosis and identifies actionable targets . Acta Neuropathol . 2016 ; 131 ( 6 ): 903 – 910 . Google Scholar CrossRef Search ADS PubMed 22. Hertenstein A , Jones D , Sahm F , et al. Umbrella protocol for phase I/IIa trials of molecularly matched targeted therapies plus radiotherapy in patients with newly diagnosed glioblastoma without MGMT promoter methylation Neuro Master Match (N2M2) . J Clin Oncol . 2016 ; 34 ( 15 suppl ):doi: 10.1200/JCO.2016.34.15_suppl.TPS2084 . 23. INFORM . INFORM registry webpage . http://www.dkfz.de/en/inform/index.html. 24. Wick W , Weller M , van den Bent M , et al. MGMT testing—the challenges for biomarker-based glioma treatment . Nat Rev Neurol . 2014 ; 10 ( 7 ): 372 – 385 . Google Scholar CrossRef Search ADS PubMed 25. Wiestler B , Capper D , Hovestadt V , et al. Assessing CpG island methylator phenotype, 1p/19q codeletion, and MGMT promoter methylation from epigenome-wide data in the biomarker cohort of the NOA-04 trial . Neuro Oncol . 2014 ; 16 ( 12 ): 1630 – 1638 . Google Scholar CrossRef Search ADS PubMed 26. Alexandrov LB , Nik-Zainal S , Wedge DC , et al. ; Australian Pancreatic Cancer Genome Initiative; ICGC Breast Cancer Consortium; ICGC MMML-Seq Consortium; ICGC PedBrain . Signatures of mutational processes in human cancer . Nature . 2013 ; 500 ( 7463 ): 415 – 421 . Google Scholar CrossRef Search ADS PubMed 27. International Cancer Genome Consortium PedBrain Tumor P . Recurrent MET fusion genes represent a drug target in pediatric glioblastoma . Nat Med . 2016 ; 22 ( 11 ): 1314 – 1320 . CrossRef Search ADS PubMed 28. Snuderl M , Fazlollahi L , Le LP , et al. Mosaic amplification of multiple receptor tyrosine kinase genes in glioblastoma . Cancer Cell . 2011 ; 20 ( 6 ): 810 – 817 . Google Scholar CrossRef Search ADS PubMed 29. Malhotra A , Lindberg M , Faust GG , et al. Breakpoint profiling of 64 cancer genomes reveals numerous complex rearrangements spawned by homology-independent mechanisms . Genome Res . 2013 ; 23 ( 5 ): 762 – 776 . Google Scholar CrossRef Search ADS PubMed 30. Furgason JM , Koncar RF , Michelhaugh SK , et al. Whole genome sequence analysis links chromothripsis to EGFR, MDM2, MDM4, and CDK4 amplification in glioblastoma . Oncoscience . 2015 ; 2 ( 7 ): 618 – 628 . Google Scholar CrossRef Search ADS PubMed 31. MolecularNeuropathology . Platform for next generation neuropathology . https://www.molecularneuropathology.org/mnp. 32. Sturm D , Witt H , Hovestadt V , et al. Hotspot mutations in H3F3A and IDH1 define distinct epigenetic and biological subgroups of glioblastoma . Cancer Cell . 2012 ; 22 ( 4 ): 425 – 437 . Google Scholar CrossRef Search ADS PubMed 33. Sturm D , Orr BA , Toprak UH , et al. New brain tumor entities emerge from molecular classification of CNS-PNETs . Cell . 2016 ; 164 ( 5 ): 1060 – 1072 . Google Scholar CrossRef Search ADS PubMed 34. Verhaak RG , Hoadley KA , Purdom E , et al. ; Cancer Genome Atlas Research Network . Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1 . Cancer Cell . 2010 ; 17 ( 1 ): 98 – 110 . Google Scholar CrossRef Search ADS PubMed 35. Felsberg J , Rapp M , Loeser S , et al. Prognostic significance of molecular markers and extent of resection in primary glioblastoma patients . Clin Cancer Res . 2009 ; 15 ( 21 ): 6683 – 6693 . Google Scholar CrossRef Search ADS PubMed 36. Knobbe CB , Reifenberger J , Reifenberger G . Mutation analysis of the Ras pathway genes NRAS, HRAS, KRAS and BRAF in glioblastomas . Acta Neuropathol . 2004 ; 108 ( 6 ): 467 – 470 . Google Scholar CrossRef Search ADS PubMed 37. Behling F , Barrantes-Freer A , Skardelly M , et al. Frequency of BRAF V600E mutations in 969 central nervous system neoplasms . Diagn Pathol . 2016 ; 11 ( 1 ): 55 . Google Scholar CrossRef Search ADS PubMed 38. Hatae R , Hata N , Suzuki SO , et al. A comprehensive analysis identifies BRAF hotspot mutations associated with gliomas with peculiar epithelial morphology . Neuropathology . 2017 ; 37 ( 3 ): 191 – 199 . Google Scholar CrossRef Search ADS PubMed 39. Singh D , Chan JM , Zoppoli P , et al. Transforming fusions of FGFR and TACC genes in human glioblastoma . Science . 2012 ; 337 ( 6099 ): 1231 – 1235 . Google Scholar CrossRef Search ADS PubMed 40. Kessler T , Sahm F , Sadik A , et al. Molecular differences in IDH wildtype glioblastoma according to MGMT promoter methylation . Neuro Oncol . 2018 ; 20 ( 3 ): 367 – 379 . Google Scholar CrossRef Search ADS PubMed 41. Wick W , Fricke H , Junge K , et al. A phase II, randomized, study of weekly APG101+reirradiation versus reirradiation in progressive glioblastoma . Clin Cancer Res . 2014 ; 20 ( 24 ): 6304 – 6313 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Neuro-OncologyOxford University Press

Published: Nov 18, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off