GENOMIC CLASSIFIER THAT PREDICTS RESPONSE TO MULTI-KINASE INHIBITOR TREATMENT INTRODUCTION
20170342499 · 2017-11-30
Inventors
- Jean-Yves BLAY (CALUIRE, FR)
- Xiaojun JIANG (LYON, FR)
- Olivier TREDAN (ECULLY, FR)
- Isabelle RAY COQUARD (CALUIRE, FR)
Cpc classification
C12Q2600/106
CHEMISTRY; METALLURGY
C12Q1/6809
CHEMISTRY; METALLURGY
C12Q1/6874
CHEMISTRY; METALLURGY
International classification
Abstract
The method for predicting the anti-tumor response in a human or animal having a tumor to multiple kinase inhibitors, using any multiple kinase inhibitor, comprises selection of genes encoding for protein kinases targeted by the said tyrosine kinase inhibitor, for each one of these genes, providing at least one nucleic acid probe which hybridizes to said gene under stringent conditions, thus providing an array of nucleic acid probes, having a biological sample containing cancer cells from said human or animal, extracting DNA from the sample, fragmenting into DNA fragments, optionally labeling the DNA fragments, submitting the optionally labeled DNA fragments to hybridization with the array of nucleic acid probes, recovering and quantifying for all the genes the gains or losses in gene copy numbers, wherein gains and losses of gene copy numbers of each selected gene are used to determine whether the tumor is sensitive or not to said kinase inhibitor.
Claims
1. A method for predicting the anti-tumor response in a human or animal having a tumor to multiple kinase inhibitors, using any multiple kinase inhibitor, comprising: selection of genes encoding for protein kinases targeted by the said tyrosine kinase inhibitor, for each one of these genes, providing at least one nucleic acid probe which hybridizes to said gene under stringent conditions, thus providing an array of nucleic acid probes, having a biological sample containing cancer cells from said human or animal, extracting DNA from the sample, fragmenting into DNA fragments, optionally labeling the DNA fragments, submitting the optionally labeled DNA fragments to hybridization with the array of nucleic acid probes, and recovering and quantifying for all the genes the gains and losses in gene copy numbers, wherein gains and losses of gene copy numbers of each selected gene are used to determine whether the tumor is sensitive or not to said kinase inhibitor.
2. The method of claim 1, wherein the array of nucleic acid probes is an array of oligonucleotides.
3. The method of claim 1, comprising mixing the optionally labeled DNA fragments with optionally labelled non-tumoral reference DNA and subjecting the mixture to Comparative Genomic Hybridization (CGH).
4. The method of claim 3, wherein CGH is performed using a human whole-genome oligonucleotide array.
5. The method of claim 1, wherein: the sum TTC of the copy-number gains of all selected genes encoding for protein kinases targeted by said kinase inhibitor is calculated, and the sum TTL of the copy-number losses of all selected genes encoding for protein kinases targeted by said kinase inhibitor is calculated, wherein a sum of the copy-number gains greater than the sum of copy-number losses is indicative that the tumor is sensitive to the kinase inhibitor.
6. The method of claim 5, wherein: a TTC≧4 and TTC≧TTL is indicative that the tumor is sensitive to the kinase inhibitor; a TTC≧4 and TTC<TTL is indicative that the tumor is resistant to the kinase inhibitor; a TTC such as 1<TTC<4 combined to TTC>TTL is indicative that the tumor is sensitive to the kinase inhibitor; a TTC such as 1<TTC<4 combined to TTC≦TTL is indicative that the tumor is resistant to the kinase inhibitor; and/or a TTC≦1 is indicative that the tumor is resistant to the kinase inhibitor.
7. The method of claim 5, wherein the following algorithm is used:
8. The method of claim 1, wherein CGH is an array comparative genomic hybridization.
9. The method of claim 1, wherein the kinase inhibitor is regorafenib, imatinib, sorafenib, pazopanib, axitinib, cabozantinib or vandetanib.
10. The method of claim 9, wherein the kinase inhibitor is regorafenib and the genes are selected from the group consisting of ieRET, (VEGFR1(FLT1), VEGFR2(KDR), VEGFR3(FLT4), KIT, PDGFRα, PDGFRβ, FGFR1, FGFR2, angiopoietin-1 receptor(TEK), DDR2, High affinity nerve growth factor receptor(NTRK1), EPHA2, RAF1, BRAF, MAPK11, FRK and ABL1.
11. A multi-kinase inhibitor for use in treating a cancer in a human or animal that has been predicted as sensitive to the multi-kinase inhibitor.
12. The inhibitor for the use of claim 11, wherein the human or animal that has been predicted as sensitive to the multi-kinase inhibitor by Comparative Genomic Hybridization (CGH).
13. The inhibitor for the use of claim 12, wherein the human or animal that has been predicted as sensitive to the multi-kinase inhibitor by the method according to claim 1.
14. The inhibitor for the use of claim 12, wherein the human or animal that has been predicted as sensitive to the multi-kinase inhibitor by Comparative Genomic Hybridization (CGH) and use of the following algorithm:
Description
[0069] The present invention will now be described in more detail using non-limiting embodiments referring to the appended figures.
[0070]
[0071]
[0072]
[0073]
[0074]
[0075]
EXAMPLE 1:
[0076] 1. Methods
[0077] Study Design and Patients
[0078] Patients included in the profiLER program (Program to Establish the Genetic and Immunologic Profile of Patient's Tumor for All Types of Advanced Cancer, NCT01774409) treated with MTKI in advanced stage were included. The profiLER study enrolls patients with advanced solid tumors and aims to establish a genetic profile by CGH and targeted mutation sequencing. As of Nov 2014, 1163 patients have been included.
[0079] Patients
[0080] 58 patients were analyzed in this work
[0081] Patients Treated with Regorafenib, n=25
[0082] Among the first 700 patients enrolled in the program from March 2013 to March 2014, 23 patients with metastatic colorectal cancer (mCRC) and 5 patients with advanced soft tissue sarcomas (STS) pretreated with chemotherapy received regorafenib from February 2011 to February 2014 (Tables 2 and 3) under the ATU, a compassionate use procedure of the French National Agency of Medicine and Health Products Safety. Three tumor samples did not fulfill DNA quality requirements for analysis. Twenty five tumor samples were therefore analyzed. This group was split into a discovery cohort of 13 CRC patients and the first validation cohort of 12 patients, with 7 mCRC and 5 STS patients. The patients received regorafenib at standard dose of 160 mg or 120 mg daily as first-line MTKI according the performance status. The regorafenib dosage was adjusted by the treating physician on the basis of presence/absence of adverse events.
[0083] Patients Treated with Other MTKI, n=33
[0084] To generalize our hypothesis in other tumors types treated with other MTKIs, a second validation cohort of 33 profiLER patients treated with one of the 6 MTKIs (Sorafenib, Sunitinib, Pazopanib, Axitinib, Vandetanib and Cabozantinib) as 1.sup.st line MTKI therapy. The predictive model was finally tested in a third set of 22 of these 33 patients who have received 2.sup.nd or more MTKI.
TABLE-US-00002 TABLE 2 Discovery cohort 1.sup.st Validation 2.sup.nd Validation Total 13 12 33 Age (median, 63.1 (40.7-75.8) 56.0 (41.6-70.5) 55.9 (24.6-76.1) range) Main tumor type n (%) CRC.sup.1) 13 (100%) 7 (58.3%) 3 (9.1%) STS.sup.2) 0 (0%) 5 (41.7%) 5 (15.2%) RCC.sup.3) 0 (0%) 0 (0%) 12 (36.3%) Thyroid.sup.4) 0 (0%) 0 (0%) 7 (21.2%) HCC.sup.5) 0 (0%) 0 (0%) 4 (12.1%) MTKIs concerned n (%) Regorafenib 13 (100%) 12 (100%) 0 (0%) Sorafenib 0 (0%) 0 (0%) 16 (48.4%) Sunitinib 0 (0%) 0 (0%) 12 (36.4%) Pazopanib 0 (0%) 0 (0%) 1 (3.0%) Axitinib 0 (0%) 0 (0%) 1 (3.0%) Vandetanib 0 (0%) 0 (0%) 2 (6.1%) Cabozantinib 0 (0%) 0 (0%) 1 (3.0%) Baseline ECOG score n (%) 0 2 (15.4%) 5 (41.7%) 13 (39.4%) 1 4 (30.8%) 3 (25%) 14 (42.4%) 2 5 (38.5%) 2 (16.7%) 2 (6.1%) NA.sup.6) 2 (15.4%) 2 (16.7%) 4 (12.1%) .sup.1)CRC: Colorectal cancer .sup.2)STS: Soft Tissue Sarcoma .sup.3)RCC: Renal Cell Carcinoma .sup.4)Thyroid: thyroid carcinoma .sup.5)HCC: Hepatocellular Carcinoma .sup.6)NA: not available
TABLE-US-00003 TABLE 3 Time to Number of metastasis MTKI previous lines of No Gender Age Tumor Type (month) received chemotherapy Discovery Cohort 1 BR M 64.5 Colorectal 14.1 Regorafenib 2 2 LJ F 61.5 Colorectal 7.6 Regorafenib 5 3 CB F 65.4 Colorectal 25.1 Regorafenib 10 4 DH M 72.2 Colorectal 12.0 Regorafenib 5 5 CP M 69.0 Colorectal 0 Regorafenib 4 6 BZ F 53.2 Colorectal 34.3 Regorafenib 4 7 JJ F 60.2 Colorectal 0 Regorafenib 2 8 BM M 69.7 Colorectal 0 Regorafenib 5 9 MB M 72.7 Colorectal 0 Regorafenib 5 10 RS F 40.7 Colorectal 0 Regorafenib 5 11 AM M 75.8 Colorectal 11.0 Regorafenib 3 12 MC M 65.4 Colorectal 0 Regorafenib 2 13 LH* M 51.2 Colorectal 1.9 Regorafenib 4 1.sup.st Validation Cohort 1 PN F 55.1 Colorectal 0 Regorafenib >2 2 AS F 41.6 Sarcoma 23.1 Regorafenib 2 3 PA M 68.9 Sarcoma 4.4 Regorafenib 1 4 GN F 57.3 Sarcoma 0 Regorafenib 2 5 GE F 67.3 Sarcoma 5.3 Regorafenib 3 6 CJC M 53.4 Colorectal 0 Regorafenib 6 7 PE F 42.8 Colorectal NA Regorafenib 3 8 PN F 58.6 Colorectal 0 Regorafenib 2 9 NA F 70.5 Colorectal 28 Regorafenib 4 10 CF F 67.8 Colorectal 0 Regorafenib 3 11 BE M 44.1 Sarcoma 1.3 Regorafenib 2 12 RD M 44.4 Colorectal 0 Regorafenib 2 2.sup.nd Validation Cohort 1 VOM F 53.9 ACC.sup.4) 0 Sunitinib 2 2 NL* F 76.1 Sarcoma NA Sorafenib 2 3 DE M 45.3 Colorectal 0 Sorafenib 6 4 F 70.7 Sarcoma 0 Sorafenib 3 5 MC* M 47.1 Thyroid 4.1 Sorafenib 0 6 CC F 49.8 RCC.sup.3) 86.4 Sunitinib 0 7 DP* M 53.4 RCC 20.4 Sunitinib 0 8 GAM M 69.1 RCC 23.8 Sunitinib 1 9 PJ* M 54.4 Thyroid 44.2 Vandetanib 0 10 TM F 65.6 HCC 5.0 Sorafenib 0 11 BR* M 59.0 Thyroid 24.7 Cabozantinib 0 12 M 59.8 RCC 2 Axitinib 0 13 GS M 64.9 CHC 19.9 Sorafenib 0 14 PF F 62.9 HCC .sup.1) 6.5 Sorafenib 0 15 CC F 54 Head&Neck 0 Pazopanib 3 16 JD* M 55.1 Sarcoma 21.2 Sorafenib 3 17 DG* M 68.2 Thyroid.sup.2) 0 Sorafenib 0 18 AJ F 50.4 HCC 4.5 Sorafenib 1 19 RJ* M 65.5 RCC 2.3 Sunitinib 0 20 GF F 43.5 Colorectal 0 Sorafenib 3 21 AL F 54.3 RCC 0 Sunitinib 0 22 AR* M 60.0 RCC 0 Sunitinib 0 23 NM M 49.9 RCC 1.6 Sunitinib 0 24 GR* M 51.7 Thyroid 15.7 Vandetanib 0 25 DC M 49.2 Colorectal 0 Sorafenib 3 26 M 56.4 Thyroid 36.9 Sorafenib 0 27 AJ* M 32.4 Sarcoma 5.7 Sorafenib 3 28 MM F 55.4 Sarcoma 135.1 Sorafenib 1 29 TG M 61.1 Thyroid 0 Sorafenib 0 30 FA* M 24.6 RCC 11.7 Sunitinib 0 31 MP* M 75.4 RCC 21.4 Sunitinib 0 32 TA* F 58.0 RCC 0 Sunitinib 0 33 MB* M 49.9 RCC 0 Sunitinib 0 .sup.1) HCC: Hepatocellular Carcinoma .sup.2)Thyroid: Differentiated thyroid carcinoma .sup.3)RCC: Renal Cell Carcinoma .sup.4)ACC: Adrenal Cortical Carcinoma Patients with * were treated by 2 or more MTKIs.
Response and Progression Free Survival
[0085] All patients had progressive disease before initiation of MTKI. Patients included in this analysis had thoraco-abdomino-pelvic CT examinations performed at the center, 4+/−2 weeks before and 8+/−2 weeks after initiation of regorafenib or other MTKI treatment. Baseline demographic and clinical data were collected, with the site and dates of metastases, previous systemic therapies, MTKI treatment, treatment duration, date and results of follow-up imaging, responses and progression free survival (PFS) determined by the radiologist and physician at each follow-up visit with Response Evaluation Criteria in Solid Tumors (RECIST, version 1.1)27.
[0086] Best response during the treatment and the PFS (defined as time from the initiation of treatment to first radiological or clinical progression or death) were collected. Patients with complete response, partial response and stable disease lasting at least 2 months were defined as “MTKI sensitive”; those with progressive disease as best response at 2 months were classified as MTKI resistant.
[0087] Array-Based Comparative Genomic Hybridization (CGH)
[0088] Sample Selection and DNA Extraction
[0089] All the tumor samples (formalin fixed paraffin embedded—FFPE) were stored in the center before treatment by MTKIs. Tumor samples were collected from the primary tumor (n=40, 69.0%) or from metastasis (n=18, 31.0%), mainly from formalin fixed archival tissues. All samples were collected before the MTKI treatments.
[0090] Array CGH
[0091] Fragmentation and labeling were done according to the manufacturer's recommendations for the CGH array (Agilent Technologies, Santa Clara, Calif.). In brief, 1.5 μg of tumor DNA and 1.5 μg of reference DNA (Promega #G1471 or #G1521, WI, USA) were heat denatured and fragmented during 10 min at 95° C. Then, tumor DNA was chemically labeled with Kreatech's Universal Linkage System (ULSTM) Cy5-dye, whereas reference DNA was labeled with Cy3-dye (Agilent #5190-0450). Labeled samples were then purified using KREApure columns (Agilent #5190-0418). Co-hybridization was performed on 4*180K Agilent SurePrint G3 Human whole-genome oligonucleotide arrays (Agilent #G4449A), containing 180 000 oligonucleotide probes. Slides were washed, dried and scanned on the Agilent Surescan scanner according to the manufacturer's recommendations. Scanned images were processed using Agilent Feature Extraction software V11.0 and the analysis was carried out using the Agilent Genomic Workbench software V7.0. The identification of aberrant copy number segments was based on ADM-2 segmentation algorithm with default settings (Threshold of 15.0). A null Log2 ratio corresponds to a balanced tumor/normal DNA ratio. Low-level and high-level copy number gains/losses were defined as a log2 (ratio)>0.25 and 1.5.
[0092] Somatic Mutation Detection with NGS (Ion Personal Genome Machine)
[0093] Ten nanograms of DNA were used for the Ion Torrent library preparation of a panel covering 59 key cancer genes (Table S2) following the manufacturer's protocol for the Ion AmpliSeq Library Kit 2.0 (Life Technologies). The size distribution of the DNA amplicons was analyzed on the 2200 TapeStation (Agilent) using the High sensitivity kit (Agilent). Template preparation, emulsion PCR, and Ion Sphere Particle (ISP) enrichment was performed using the One Touch 2 kit (Life Technologies) according to manufacturer's instructions. The ISPs were loaded onto a 318 chip (Life Technologies) and sequenced using an Ion PGM 200 v2 sequencing kit (Life Technologies) on the Ion Torrent PGM for 500 cycles.
[0094] After a successful sequencing reaction, the raw signal data were analyzed using NextGENe Software Suite v3.4.2 (Soft genetics). The pipeline includes quality score assignment, alignment to human genome 19 reference, mapping quality QC, coverage analysis and variant calling. After completion of the primary data analysis, lists of detected sequence variants (SNVs and INDELs) were compiled in a VCF (Variant Call File) format. For downstream analysis, variants with minimum coverage of 100 reads containing at least 10 of the mutant reads were selected. Variant calls were further analyzed using variant filtering and annotation using COSMIC v.64 and dbSNP build 135.
TABLE-US-00004 ABL1 CSF1 IGF1R MET PIK3CA ROS1 SRC VEGFR1 AKT1 CSF1R JAK2 MPL PIK3R1 RYK STK11 VEGFR2 AKT2 DDB2 JAK3 MST1R PTCH SDHAF2 TEK VEGFR3 ALK DDR1 KIT mTOR PTEN SDHB TIE1 VHL APC DDR2 KRAS MUSK RB1 SDHC TP53 AXL EGFR NRAS PDGFA RET SDHD TSC1 BRAF ERBB2 HRAS PDGFRA ROR1 SMARCB1 TSC2 CRAF FLT3 MERTK PDGFRB ROR2 SMO TYRO3
Statistical Analysis
[0095] The association between SCNAs and categorical variables was tested using the Mann-Whitney U-test. Association between categorical variables was assessed using Chi-square test. All p-values were two-sided. Survival curves were plotted using the Kaplan Meier method and compared using a log rank test. Statistical analysis was conducted using the SPSS 19.1 Package (SPSS, IBM France). Table 4: List of analysed mutations status by NGC (Ion torrent).
[0096] 2. Results
Patient Characteristics
[0097] In total, all evaluable 58 patients included in the ProfilER study, who had received at least one of the 7 MTKIs listed in Table 1 in the first line setting were included in this analysis. Patients' characteristics are detailed (Tables 2 and 3).
[0098] Regorafenib Treated Patients: Discovery Cohort
[0099] 25 patients treated with regorafenib as first-line MTKI. The discovery cohort consisted of 13 patients with metastatic colorectal cancers (mCRC) who had received the regorafenib as first-line MTKI treatment, after having progressed under irinotecan, oxaliplatin containing regimens. The median duration of regorafenib treatment was 3 months (range, 0.5-25). Six patients who had achieved a stable disease (SD) or objective response (partial response, PR or complete response, CR) at 8 weeks were considered as regorafenib sensitive and 7 other patients with progressive disease≦8 weeks were qualified as regorafenib resistant.
Establishment of the Target Copy Number Change Pattern
[0100] The copy number changes of the panel of 18 genes encoding for kinases whose enzymatic activity is blocked by regorafenib was investigated: RET, VEGFR1(FLT1), VEGFR2(KDR), VEGFR3(FLT4), KIT, PDGFRα, PDGFRβ, FGFR1, FGFR2, angiopoietin-1 receptor(TEK), DDR2, High affinity nerve growth factor receptor(NTRK1), EPHA2, RAF1, BRAF, MAPK11, FRK, ABL128. Analyzing the SCNA of 18 target genes in these 13 tumors using CGH array (Agilent), we explored a possible correlation between the clinical outcome and SCNAs of these target genes. The sum of gains on target genes were termed as tumor target charge (TTC), while the sum of deletions of target genes were termed as tumor target loss (TTL).
TABLE-US-00005 TABLE 5 Copy-number change pattern of the 18 target genes of regorafenib in the discovery cohort. The copy number change pattern of 18 targets genes is displayed as heatmap. Top and bottom show the grouped results of 6 regorafenib sensitive and 7 regorafenib resistant tumors, respectively. Table 5 EPHA2 NTRK1 DDR2 RAF1 KIT KDR PDGFRA PDGFRB FLT4 FRK Regoranib 1 1 1 1 1 1 1 1 1 1 Sensitive 0 1 1 0 0 0 0 0 0 1 0 1 1 0 0 0 0 1 1 0 0 1 1 −1 0 0 0 1 1 1 0 1 1 0 0 0 0 0 0 1 0 0 0 −1 −1 −1 −1 0 0 0 Regorafenib 0 0 0 0 0 0 0 0 0 0 Resistant −1 1 1 −1 0 0 0 0 0 0 −1 1 2 0 −1 −1 −1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 −1 0 0 −1 −1 −1 −1 0 0 0.5 Table 5 BRAF FGFR1 TEK ABL1 RET FGFR2 FLT1 MAPK11 Regoranib 1 0 0 1 1 0 1.5 0 Sensitive 0 0 1 1 0 0 1.5 0 1 0 0 0 0 0 1 −1 1 0 1 1 0 0 1 −1 1 0 0 0 0 0 1 0 0 0 0.5 −1 0 0 0 0 Regorafenib 0 −1 1 0.5 0 0 0.5 0.5 Resistant 0 −1 0 0 0 0 0 −1 0.5 −1 0.5 0.5 0.5 0.5 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0.5 0 0.5 0 0 0 0 0 0 0 0 0 −1 −1 0 0 0.5 0 2 = Gene Amplification; 1.5 = High Gain; 1 = Gain; 0.5 = Heterogeneous Gain; 0 = Normal; −0.5 = Heterogeneous Deletion; −1 = Deletion; −2 = Gene Loss
[0101] An enrichment in gains on genes encoding for regorafenib target was observed in sensitive patients with a total of 41 gains across 6 samples (mean: 6.8; range 1-14) versus 20 gains across 7 samples (mean: 2.1; range 0-7) in the regorafenib resistant group. The regorafenib sensitive tumors had a total of 8 deletions (mean: 1.3; range 0-5), while the resistant tumors had a total of 17 deletions (mean: 2.4; range 0-7). The differences between TTC and negative TTC were significantly higher in the sensitive group (P=0.038; Mann Whitney). In addition, five of six sensitive tumors had a TTC≧4, vs 2 of 7 resistant tumors (P=0.048). The details of SCNAs of all patients are listed in Table 6.
TABLE-US-00006 TABLE 6 Patients characteristics, target gene copy number change and sensitivity predicted Gains on Losses on Best targeted targeted Sensitivity Patients Tumor MTKIs Response kinases kinases Predicted Discovery Cohort 1 BR CRC.sup.1) Regorafenib SD 14 0 / 2 LJ CRC Regorafenib PR 6 0 / 3 CB CRC Regorafenib SD 6 1 / 4 DH CRC Regorafenib SD 9 2 / 5 CP CRC Regorafenib SD 5 0 / 6 BZ CRC Regorafenib SD 0 0 / 7 AM CRC Regorafenib PD 2 0 / 8 BM CRC Regorafenib PD 2 4 / 9 MB CRC Regorafenib PD 2 5 / 10 MC CRC Regorafenib PD 1 0 / 11 JJ CRC Regorafenib PD 0 0 / 12 RS CRC Regorafenib PD 0 0 / 13 LH CRC Regorafenib PD 0 0 / Validation Cohort I 1 GN STS.sup.2) Regorafenib SD 14 0 Sensitive 2 PA STS Regorafenib SD 12 0 Sensitive 3 GE STS Regorafenib SD 7 1 Sensitive 4 AS STS Regorafenib PR 6 5 Sensitive 5 PN CRC Regorafenib PR 2 0 Sensitive 6 CF CRC Regorafenib PD 5 6 Resistant 7 PE CRC Regorafenib PD 5 6 Resistant 8 PN CRC Regorafenib PD 2 0 Sensitive 9 BE STS Regorafenib PD 2 2 Resistant 10 CJC CRC Regorafenib PD 1 4 Resistant 11 RD CRC Regorafenib PD 1 0 Resistant 12 NA CRC Regorafenib PD 0 0 Resistant 13 RD CRC Regorafenib PD 1 0 Resistant Validation Cohort II 1 VOM ACC.sup.3) Sunitinib SD 9 0 Sensitive 2 NL STS Sorafenib SD 4 2 Sensitive 3 DE CRC Sorafenib SD 3 0 Sensitive 4 DBO STS Sorafenib SD 3 1 Sensitive 5 MC Thyroid .sup.5) Sorafenib SD 3 1 Sensitive 6 CC RCC.sup.4) Sunitinib PR 3 0 Sensitive 7 DP RCC Sunitinib SD 3 2 Sensitive 8 GAM RCC Sunitinib SD 3 0 Sensitive 9 PJ Thyroid Vandetanib SD 3 0 Sensitive 10 TM HCC Sorafenib SD 2 0 Sensitive 11 BR Thyroid Cabozantinib PR 2 0 Sensitive 12 BJM RCC Axitinib SD 2 1 Sensitive 13 GS HCC Sorafenib SD 2 0 Sensitive 14 PF HCC.sup.6) Sorafenib SD 1 1 Resistant 15 CC Head&neck Pazopanib SD 1 0 Resistant 16 JD STS Sorafenib SD 0 2 Resistant 17 DG Thyroid Sorafenib PR 0 0 Resistant 18 AJ HCC Sorafenib PR 0 4 Resistant 19 RJ RCC Sunitinib CR 0 0 Resistant 20 GF CRC Sorafenib SD 0 2 Resistant 21 AL RCC Sunitinib PD 3 0 Resistant 22 AR RCC Sunitinib PD 3 0 Resistant 23 NM RCC Sunitinib PD 3 6 Resistant 24 GR Thyroid Vandetanib PD 1 2 Resistant 25 DC CRC Sorafenib PD 0 0 Resistant 26 VJL Thyroid Sorafenib PD 0 0 Resistant 27 AJ STS Sorafenib PD 0 3 Resistant 28 MM STS Sorafenib PD 0 3 Resistant 29 TG Thyroid Sorafenib PD 0 2 Resistant 30 FA RCC Sunitinib PD 0 3 Resistant 31 MP RCC Sunitinib PD 0 0 Resistant 32 TA RCC Sunitinib PD 0 0 Resistant 33 MB RCC Sunitinib PD 0 0 Resistant Validation cohort III 1 PJ Thyroid Sunitinib PR 6 0 Sensitive 2 PJ Thyroid Sorafenib SD 6 0 Sensitive 3 DBO STS Regorafenib SD 5 1 Sensitive 4 NL STS Pazopanib SD 5 1 Sensitive 5 LP GIST Pazopanib SD 4 1 Sensitive 6 PF HCC Regorafenib SD 2 1 Sensitive 7 BR Thyroid Vandetanib PR 2 0 Sensitive 8 MC Thyroid Pazopanib SD 2 0 Sensitive 9 LP GIST Regorafenib SD 2 4 Resistant 10 DG Thyroid Pazopanib SD 0 0 Resistant 11 DP RCC Axitinib SD 0 0 Sensitive 12 LP GIST Sunitinib PD 3 2 Sensitive 13 TH CRC Regorafenib PD 1 0 Resistant 14 CJC CRC Sorafenib PD 1 1 Resistant 15 AJ STS Pazopanib PD 1 1 Resistant 16 LH CRC Sorafenib PD 0 0 Resistant 17 AR RCC Sorafenib PD 0 0 Resistant 18 JD STS Pazopanib PD 0 1 Resistant 19 VJL Thyroid Pazopanib PD 0 0 Resistant 20 GR Thyroid Sunitinib PD 0 8 Resistant 21 FA RCC Axitinib PD 0 2 Resistant 22 AR RCC Axitinib PD 0 0 Resistant 23 TA RCC Axitinib PD 0 0 Resistant 24 RJ RCC Axitinib PD 0 0 Resistant 25 MB RCC Axitinib PD 0 0 Resistant
The Regorafenib Validation Series
[0102] This difference of TTC and TTL was assessed in the 1st validation cohort composed of the 12 patients treated with regorafenib as first-line MTKI. In this analysis, as well , gains of gene encoding for targets of regorafenib was observed in the sensitive tumors, as compared to resistant tumors (Table 7). The differences between TTC and TTL was significantly higher in sensitive tumors (Mann Whitney, p=0.014). Additionally, four of five sensitive tumors had TTC≧4, versus two of seven resistant tumors (P=0.07; chi-square test) (Table 7). Even though several of these 18 individual genes gains predicted well for regorafenib sensitivity, the numerical combination of target gene gains and deletions was actually the most efficient predictor of sensitivity to regorafenib.
TABLE-US-00007 TABLE 7 Copy-number change pattern of 18 target genes of regorafenib in the validation cohort I. Top and bottom show the grouped results of 6 regorafenib sensitive and 7 regorafenib resistant tumors, respectively. EPHA2 NTRK1 DDR2 RAF1 KIT KDR PDGFRA PDGFRβ FLT4 FRK Regorafenib 0 1 1 1 0.5 0.5 0.5 0 0 0 Sensitive 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 0 0 0 0 0 1 0 Regorafenib 0 0 0 0 0 0 0 0.5 0.5 0 Resistant 0 0 0 0 0 0 0 0 0 0 −1 0 0 0 0 0 1 −1 −1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 −1 0 0 1 −1 −1 −1 −1 −1 0.5 0 0 0 0 0 0 0 0 0 0 Tumor BRAF FGFR1 TEK ABL1 RET FGFR2 FLT1 MAPK11 Type Regorafenib 0 −1 −1 0 −1 −1 0 −1 STS Sensitive 0 1 0 0 0 0 1 0 CRC 1 1 1 1 0 0 0 −1 STS 1 1 1 0.5 0 0 1 0.5 STS 1 2 −1 1 −1 −1 0 −1 STS Regorafenib 0 0 −1 0 0 0 −1 0 STS Resistant 1 −1 0 0 −1 −1 0 −1 CRC −1 1 −1 0 0.5 0.5 1.5 −1 CRC 1 0 0 0 0 0 1 0 CRC 0 0 0 0 0 0 0 0 CRC 1 0 1 0 1 0 1 0 CRC 0 0 0 0 0 0 1 0 CRC 2 = Gene Amplification; 1.5 = High Gain; 1 = Gain; 0.5 = Heterogeneous Gain; 0 = Normal; −0.5 = Heterogeneous Deletion; −1 = Deletion; −2 = Gene Loss All gain events are considered as 1 TTC and all lost events are considered as 1 TTL.
Pool of the Regorafenib Series
[0103] The pooled integral analysis of the 2 regorafenib sets (Table 8) revealed that the TTL tends to outnumber TTC in resistant tumors. As expected, the difference between TTC and
[0104] TTL was significantly higher in the sensitive tumors (P=0.003; Mann Whitney). Of note, all 8 patients with a difference no less than five achieved clinical benefit, vs 3 of the 17 remaining patients (p=0.0001). Combining the two parameters enabled to delineate an algorithm, termed as SUMSCAN (see above). Using this algorithm, ten of the eleven sensitive tumors would have been identified as responders, vs five of the fourteen remaining patients (P=0.005; chi-square test), resulting in a sensitivity of 90.9% and a specificity of 66.7%. A prediction accuracy of 76% (19 of 25) was achieved.
[0105] Furthermore, the prognostic significance of SUMSCAN was evaluated using univariate Kaplan-Meier survival analysis. The progression free survival and overall survival of patients treated with regorafenib in 2 cohorts are significantly better in patients with a favorable SUMSCAN profile (defined as TTC≧4; P.sub.PFS=0.001, P.sub.OS=0.017, respectively; Log-rank test,
TABLE-US-00008 TABLE 8 Integral analysis of expression pattern of 18 target genes in 25 patients treated with regorafenib. EPHA2 NTRK1 DDR2 RAF1 KIT KDR PDGFRA CSF1R PDGFRB FLT4 FRK Regorafenib 1 1 1 1 1 1 1 1 1 1 1 Sensitive 0 1 1 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 1 1 1 0 0 1 1 0 0 0 0 0 0 0 1 0 1 1 −1 0 0 0 1 1 1 1 0 0 0 −1 −1 −1 −1 0 0 0 0 0 1 1 1 0.5 0.5 0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 0 0 0 0 1 0 Regorafenib 0 0 0 0 0 0 0 0 0 0 0 Resistant −1 1 1 −1 0 0 0 0 0 0 0 −1 1 2 0 −1 −1 −1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 −1 0 0 −1 −1 −1 −1 0 0 0 0.5 0 0 0 0 0 0 0 0.5 0.5 0.5 0 0 0 0 0 0 0 0 0 0 0 0 −1 0 0 0 0 0 1 −1 −1 −1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 −1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Tumor BRAF FGFR1 TEK ABL1 RET FGFR2 FLT1 MAPK11 type Regorafenib 1 0 0 1 1 0 1.5 0 CRC Sensitive 0 0 1 1 0 0 1.5 0 CRC 1 0 0 0 0 0 1 −1 CRC 1 0 0 0 0 0 1 0 CRC 1 0 1 1 0 0 1 −1 CRC 0 0 0.5 −1 0 0 0 0 CRC 0 −1 −1 0 −1 −1 0 −1 STS 0 1 0 0 0 0 1 0 CRC 1 1 1 1 0 0 0 −1 STS 1 1 1 0.5 0 0 1 0.5 STS 1 2 −1 1 −1 0 0 0 STS Regorafenib 0 −1 0.5 0.5 0 0 0.5 0.5 CRC Resistant 0 −1 0 0 0 0 0 −1 CRC 0.5 −1 0.5 0.5 0.5 0.5 0 0 CRC 0 0 0 0 0 0 0 0 CRC 1 0 1 0 0 0 0.5 0 CRC 0.5 0 0 0 0 0 0 0 CRC 0 0 −1 −1 0 0 0.5 0 CRC 0 0 −1 0 0 0 −1 0 STS 1 −1 0 0 −1 −1 0 −1 CRC −1 1 −1 0 0.5 0.5 1.5 −1 CRC 1 0 0 0 0 0 1 0 CRC 0 0 0 0 0 0 0 0 CRC 0 0 0 0 0 0 0 0 CRC 0 0 0 0 0 0 1 0 CRC 2 = Gene Amplification; 1.5 = High Gain; 1 = Gain; 0.5 = Heterogeneous Gain; 0 = Normal; −0.5 = Heterogeneous Deletion; −1 = Deletion; −2 = Gene Loss All gain events are considered as 1 TTC and all lost events are considered as 1 TTL.
Gains and Deletions of Specific Target Genes
[0106] A significant difference in the gain frequencies between regorafenib sensitive tumor and regorafenib resistant tumors was observed for genes DDR2, NTRK1 (High affinity nerve growth factor receptor) and FLT4 (
Overall Gains in the Genome of Sensitive vs Resistant Patients
[0107] As a control, we compared whole genome gains and deletions in the same experiment (
Added Predictive Ability of Sequence Mutations of the Same Samples Using NGS
[0108] To gain more insight in the discrepancies between the SUMSCAN model and efficacy of the treatments, we then investigated the correlation between mutations in the genes of NGS panel (Gene list Table 4) and response to regorafenib. Target mutation sequencing was feasible for 100% (25 of 25) of the samples. PIK3CA mutation (42.9% in resistant tumors vs none of the sensitive tumors, P=0.0196; Fisher exact test) was found associated with regorafenib resistance, but not TP53 (72.7% in sensitive tumor vs 64.3% in resistant tumors) nor KRAS (27.3% in sensitive tumor vs 57.1% in resistant tumors). Two of six PIK3CA mutant tumors were predicted sensitive by SUMSCAN. Six PIK3CA hotspot mutations were located in exon 9 (E542K*2, E545K*2, Q546P) and one in exon 20 (H1047R), all with relatively low tumor target charge (TTC≦2) and 5 of 6 predicted to be resistant by SUMSCAN. This association was observed only in mCRC, not in STS.
Second Validation Series: Assessment of the Model in Tumors Treated by Other Mtkis
[0109] The prediction power of SUMSCAN was further assessed in 33 patients treated with other 6 MTKIs as 1.sup.st line MTKI. Twenty-two patients have subsequently received 2 or more MTKIs (see Table 6). The definition of target genes varied between each MTKI according to DrugBank (http://www.drugbank.ca; see Table 1; MET, KDR, RET for Cabozantinib). The SUMSCAN was applied to this group of patients treated with different MTKI for a variety of neoplastic diseases (Tables 6 and 8). Again, the difference between TTC and TTL was significantly different between sensitive tumors (n=20) and resistant tumors (n=13) (P=0.008; Mann Whitney). Fourteen of twenty responsive patients were predicted sensitive by SUMSCAN, vs two of the thirteen resistant patients (P=0.002; chi-square) resulting in a sensitivity of 70% and a predictive accuracy of 75.8%. The positive predictive value (PPV) and negative predictive value (NPV) were 87.5% and 64.7% respectively (P=0.002, chi-square). Patients with a favorable SUMSCAN profile had a significantly better PFS, with a median PFS of 9.9 months vs only 2.8 months for patients with unfavorable SUMSCAN profile and a trend for a longer survival (
[0110] We then compared whole-genome gains and losses of sensitive tumor profiles (n=20) and resistant tumor profiles (n=13), and found that a total of 622 gains in sensitive tumors (mean: 31.1; range0-87) versus 186 gains in resistant tumors (mean:14.3; range 0-43) (p=0.053, Mann-Whitney), and 247 losses in 20 sensitive samples (mean: 12.4; range0-37) versus 327 losses in resistant samples(mean: 25.2; range 1-68)(P=0.036; Mann-Whitney). The difference of TTC and TTL was marginally different between sensitive and resistant tumors (P=0.06; Mann-Whitney).
Second Line and Beyond
[0111] The relevance of SUMSCAN was analyzed in 22 of the 33 patients treated with 2.sup.nd line or more MTKIs. Three patients were treated with more than 2 MKTI. We had included also the 3.sup.rd line and 4.sup.th line MTKI in the analysis. This third validation cohort consists of 26 cases treated by a second line MTKI or beyond. SUMSCAN predicted 10 out of 12 MTKIs sensitive patients in second line and 12 out of 14 MTKIs resistant patients in second line, with an accuracy rate of 84.6% (22/26), a sensitivity of 83.3% and a specificity of 85.7% (P=0.0011, Fisher exact test). Interestingly, two patients had a TTC superior to that of the 1.sup.st line MKTI, both had experienced a longer PFS than that in the 1.sup.st line setting, one with thyroid carcinoma had a PFS of 19 months with vandetanib (TTC=3) and 35 months with sunitinib (TTC=6) in the 2.sup.nd line. Patients with a favorable SUMSCAN profile had a significantly better PFS and OS (
SUMSCAN Performance in 5 Histological Types Treated by MTKIs
[0112] See
[0113] See
SUMSCAN is Not Predictive of the Response to Conventional Chemotherapy
[0114] To evaluate the specificity of the SUMSCAN algorithm, the correlation between the SUMSCAN and response to conventional chemotherapy regimens (irinotecan and oxaliplatin containing regime) received beforehand was evaluated in 21 CRC patients. These patients were divided into 2 groups according to the SUMSCAN (predicted sensitive, predicted resistant; see Table 10). No correlation between the SUMSCAN and response to conventional chemotherapy were observed.
TABLE-US-00009 TABLE 10 Prediction model and response to chemotherapy in CRC patients (n = 21) Irinotecan Predicted Predicted Sensitivity Sensitive Resistant Total Sensitivity 9 (42.9%) 7 (33.3%) 16 (76.2%) Resistant 3 (14.3%) 2 (9.5%) 5 (23.8%) P = 0.882 (Chi-square) Oxaliplatin Predicted Predicted Sensitivity Sensitive Resistant Total Sensitivity 6 (28.6%) 7 (33.3%) 13 (61.9%) Resistant 6 (28.6%) 2 (9.5%) 8 (38.1%) P = 0.195 (Chi-square)
[0115] Discussion
[0116] The hypothesis explored in this work was that the antitumor activity of MTKI may be related to the sum of gains and losses of genes encoding for the receptors and targets of these MTKI in a given tumor. The underlying biological rationale is that a high level of target genes copy number gains in tumor may indicate tumor's oncogenes dependence and therefore their increased sensitivity to molecules targeting these genes.
[0117] Actually, standard biomarkers of response to MTKI, in particular those inhibiting VEGF receptors, have not been identified yet. This is in contrast with what is observed when a key driver event such as KIT mutations, BRAF mutations, ALK translocations is present in the tumor.
[0118] Here, it is reported that the antitumor activity of MTKIs in tumors lacking a well-defined strong oncogenic driver is strongly correlated to the gains of additional copies and/or losses of genes encoding for the protein kinases which are the targets of these MTKIs. The sum of gains of the genes encoding for targets of MKTIs, termed as tumor target charge (TTC) was found higher in responding patients, enabling to delineate a predictive score. Based on the concept of TTC, we created the predictive score SUMSCAN, identifying a favorable and an unfavorable group based on the classification of the patients into three TTC groups (low, medium and high).
[0119] Regorafenib was investigated firstly in patients with mCRC progressing after irinotecan, oxaliplatin and 5FU. The difference between TTC and TTL was found significantly higher in patients who experienced tumor control to regorafenib. This was confirmed in the first validation series including mCRC patients and advanced STS progressing after standard chemotherapy with doxorubicin. A predictive score SUMSCAN was delineated and validated on both series. The PFS and OS of these 2 series were significantly longer in patients with a favorable SUMSCAN. Importantly, SUMSCAN predicted exclusively efficacy to MTKI: no correlation was observed with the previous response to widely used irinotecan and/or oxaliplatin containing regimens in mCRC treatment.
[0120] The generalizability of SUMSCAN in other tumor types and other MTKI has been tested. All 33 patients included in ProfilER and pretreated with MTKI were analyzed using the same strategy. These included a variety of histological subtypes and a six different MTKIs (sorafenib, sunitinib, pazopanib, axitinib, vandetanib and cabozantinib). The presence of gains and deletions of target genes of each of these six MTKIs was determined for each patient and TTC as well as the SUMSCAN score could be identified for each patient and each line of treatment. Again, the sum of gains of genes encoding for targets of the MTKIs were higher in responding patients across this variety of histological types and MTKIs. The SUMSCAN score was predictive of response and tumor control, and was a powerful prognostic factor for PFS and OS in this series.
[0121] This was further confirmed when the response to the second line MTKI treatment was assessed in the subgroup of 22 of these 33 patients. In these patients, response and tumor control to the second line MTKI was significantly better when a favorable SUMSCAN score was observed, with a short PFS (median=2.8 months) in patients with an unfavorable SUMSCAN score across histological subtypes and MTKI.
[0122] Finally, in all the series, all patients with more than 4 gains in genes encoding for targets of the MTKI derived clinical benefit.
[0123] Genomic alterations of several of the individual genes encoding for MTKI targets were found predictive of clinical benefit, but none were as discriminant as the numerical combination of TTC and TTL described in the SUMSCAN model. In an attempt to understand the false positive of the SUMSCAN score in the series of patients treated regorafenib, the presence of specific mutations within the NGS panel of the ProfiLER trial was investigated: only PIK3CA mutations were found exclusively observed in non-responders to regorafenib, although 2 of these patients had a favorable SUMSCAN score. No other mutations were associated with response, including KRAS, Tp53. These findings suggest that PIK3CA mutation could further refine the SUMSCAN prediction for regorafenib resistance. It was also opbserved that the mutation frequencies inversely correlate with the copy number alterations in mCRC. This trend is referred as cancer genome hyperbola which is initially to describe the fact that tumors at the extremes of genomic instability had either a large number of somatic mutations or a large number of copy number alterations, never both.
[0124] An important observation is that the overall number of gene gains in the whole genome was higher in responding patients, while conversely the number of gene losses tended to be higher in progressive patients. Even though the difference was less significant when target genes of individual MTKI were excluded from the comparison, this observation suggest that tumors which respond to these MTKIs are characterized by a global “gain profile”, with more proto-oncogene copies, including those coding for target proteins of these
[0125] MTKIs. This was not observed for oxaliplatin or irinotecan response in mCRC (not shown). These tumors with a “gain profile” may therefore be better candidate for therapies targeting multiple oncogenes in general.
[0126] These results also challenge the antiangiogenic role of these MTKIs as a major component of their antitumor activity. Indeed, while prolonged clinical benefit was observed in patients whose tumor do not present gains of the target genes, the analysis of the SUMSCAN score and TTC of individual tumor suggests that the antitumor activity of these agents is exerted primarily on the tumor cells which gained additional copies of genes encoding for MTKI targets during the process of acquisition of genomic alteration. This question is also of importance for patients treated with regorafenib or sunitinib for a GIST, with well-identified key molecular. This question is currently explored in a large dataset of patients treated in 2nd or more line with these MTKI is currently explored. It is important to note that the antitumor activity of regorafenib was observed at the same level regardless of the nature of the KIT/or PDGFRA mutation.sup.18.In conclusion, these results point to a novel concept that the response to any MTKI in human solid tumors is influenced by the sum of gains and losses of the genes encoding for the protein targets of these MTKI in the tumor. A predictive model for the selection of patients is presented and proposed for future evaluation in other series. These results could have important consequences for a better selection of patient candidate for these treatments in routine clinical settings. In addition, these results and the GSH method disclosed herein may be useful to identify candidate patients for these MTKIs outside the approved indications enabling registration of an already approved agent in additional indications, or enabling a non-approved agent to be registered. Finally, this concept that the sum of gains and losses of genes coding for target proteins is predictive for treatment efficacy has broader application beyond MTKI targeting VEGFR: the identification of responders and refractory patient to multitargeted inhibitors of ALK MET, SRC, mTOR/PI3KCA/AKT, Src family of kinases for instance is made using this method.
REFERENCES
[0127] 1. Demetri, G. D., et al. Proc Am Soc Clin Oncol, 2005, Vol. 23, p. 4000a. [0128] 2. Motzer, R. J. et al. New England Journal of Medicine, 2007, 356(2), 115-124. [0129] 3. Llovet, J. M. et al. New England Journal of Medicine, 2008, 359(4), 378-390. [0130] 4. Cheng, A. L. et al. The lancet oncology, 2009, 10(1), 25-34. [0131] 5. Grothey, A. et al. The Lancet, 2013, 381(9863), 303-312. [0132] 6. Carr, L. L. et al. Clinical Cancer Research, 2010, 16(21), 5260-5268. [0133] 7. Sternberg, C. N. et al. Journal of Clinical Oncology, 2010, 28(6), 1061-1068. [0134] 8. Cohen, E. E. et al. Journal of Clinical Oncology, 2008, 26(29), 4708-4713. [0135] 9. Demetri, G. D. et al. The Lancet, 2006, 368(9544), 1329-1338. [0136] 10. Hanahan, D., & Weinberg, R. A. et al. Cell, 2011, 144(5), 646-674. [0137] 11. Davoli, T. et al. Cell, 2013, 155(4), 948-962. [0138] 12. Druker, B. J. et al. New England Journal of Medicine, 2001, 344(14), 1031-1037. [0139] 13. Heinrich, M. C. et al. Journal of Clinical Oncology, 2007, 21(23), 4342-4349. [0140] 14. Debiec-Rychter, M. et al. European Journal of Cancer, 2004, 40 (5), 689-695. [0141] 15. Miller, A. J., & Mihm Jr, M. C. New England Journal of Medicine, 2006, 355(1), 51-65. [0142] 16. Flaherty, K. Tet al. New England Journal of Medicine, 2010, 363(9), 809-819. [0143] 17. Kelleher, F. C., & McDermott, R. European Journal of Cancer, 2010, 46(13), 2357-2368. [0144] 18. Demetri, G. D. et al. The Lancet, 2013, 381(9863), 295-302. [0145] 19. George, S., et al. J Clin Oncol, 29(15 May 2011 Supplement), 10007. [0146] 20. Pena, C. et al. Clinical cancer research, 2010, 16(19),4853-4863. [0147] 21. Zhang, Z et al. BMC medicine, 2009, 7(1), 41. [0148] 22. Porta, C. et al. Kidney international, 2010, 77(9), 809-815. [0149] 23. Gruenwald, V. et al. BMC cancer, 2010, 10(1), 695. [0150] 24. Xu, C. F et al. Journal of clinical oncology, 2011, JCO-2010. [0151] 25. Tran, H. T. et al. The lancet oncology, 2012, 13(8), 827-837. [0152] 26. Lenz, H. J. et al. J Clin Oncol, 2013, 31, abstract-3514. [0153] 27. Meyerson, M., Gabriel, S. & Getz, G. (2010. Nature Reviews Genetics, 2010, 11(10), 685-696. [0154] 28. Stratton, M. R., Campbell, P. J., & Futreal, P. A. Nature, 2009, 458(7239), 719-724. [0155] 29. Vogelstein, B. et al. Science, 2013, 339(6127), 1546-1558. [0156] 30. Eisenhauer, E. A., Therasse, P., Bogaerts, J., Schwartz, L. H., Sargent, D., Ford, R., & Verweij, J. (2009). New response evaluation criteria in solid tumours: revised RECIST guideline. [0157] 31. http://www.drugbank.ca/ [0158] 32. Ciriello, G. et al. Nature genetics, 2013, 45(10), 1127-1133.