Somatic SNVs Research Articles

Characterization of mutations in HRR genes, TMB, and PD-L1 expression in Chinese patients with ovarian cancer.

e18048 Background: A great number of clinical studies have confirmed the promising efficacy of ICIs and PARPi in multiple solid tumors. Here we aim to investigate the biomarkers related to these emerging therapies, including mutations in DNA homologous recombination repair (HRR) genes, tumor mutation burden (TMB) and PD-L1 expression in Chinese patients (pts) with ovarian cancer. Methods: A total of 209 pts with ovarian cancer were enrolled. Matched tumor-normal NGS of 1021 cancer-related genes was performed in 98 pts (T: 71 pts, T+B: 25 pts, B: 2 pts). Germline variants in ovarian cancer susceptibility genes were analyzed in 209 pts. PD-L1 expression was assessed by immunohistochemistry (Dako 22C3). Pts with TPS≥50 were considered as high expression and pts with 1≤TPS<50 was considered as expression. Tissue TMB was calculated as the number of somatic non-synonymous SNVs and Indels per Mb in the coding region (with VAF ≥0.03). The threshold of TMB-high was 5.76 muts/MB. Results: Germline pathogenetic or likely pathogenetic variants were identified in 28.7% of pts ( BRCA1 15.3%, BRCA2 7.7%, RAD51D 2.0%, PALB2 1.0%, and 0.7% for CHEK2, MSH2, RAD51C and MSH6). Germline variants in HRR genes was identified in 27.3% of pts. The mutation detection rates in 96 tissue samples from surgery or biopsy and 25 ctDNA samples collected at disease recurrence were 100% and 96%, respectively. An average of 34.6% of tissue-derived mutations can be identified in ctDNA, and 55.2% of tissue-derived actionable mutations were observed in ctDNA. Besides, there were 83 private mutations, including PPMID (5), ARID2 (3), CHEK2 (3), PIK3CA (3) and ARID1B (2). Mutations in HRR genes were found in 36.5% (35/96) of tissue samples, and 12 pts harbored more than one HRR mutations. The detection rate of HRR mutations in ctDNA was 40% (10/25), and cases harbored more than one HRR mutations. The status of TMB and PD-L1 was analyzed in 30 pts. Two pts were evaluated as PD-L1 highly expressed and 9 pts were considered as expressed. Eight pts were assessed as TMB-high. Five pts with PD-L1 negative expression were TMB-high, and 8 pts with TMB-low were PD-L1 positive expression. TMB and PD-L1 expression are independent indicators with the spearman r of 0.01 (p = 0.96). Conclusions: The detection rate of germline and somatic mutations in HRR genes were 27.3% and 36.5-40%, respectively, which defined a large number of ovarian cancer pts who can benefit from PARPi. PD-L1 expression and TMB are independent biomarkers, which may be combined to predict the efficacy of immunotherapy in future exploratory trials.

Comprehensive analysis of indels in whole-genome microsatellite regions and microsatellite instability across 21 cancer types.

Microsatellites are repeats of 1- to 6-bp units, and approximately 10 million microsatellites have been identified across the human genome. Microsatellites are vulnerable to DNA mismatch errors and have thus been used to detect cancers with mismatch repair deficiency. To reveal the mutational landscape of microsatellite repeat regions at the genome level, we analyzed approximately 20.1 billion microsatellites in 2717 whole genomes of pan-cancer samples across 21 tissue types. First, we developed a new insertion and deletion caller (MIMcall) that takes into consideration the error patterns of different types of microsatellites. Among the 2717 pan-cancer samples, our analysis identified 31 samples, including colorectal, uterus, and stomach cancers, with a higher proportion of mutated microsatellite (≥0.03), which we defined as microsatellite instability (MSI) cancers of genome-wide level. Next, we found 20 highly mutated microsatellites that can be used to detect MSI cancers with high sensitivity. Third, we found that replication timing and DNA shape were significantly associated with mutation rates of microsatellites. Last, analysis of mutations in mismatch repair genes showed that somatic SNVs and short indels had larger functional impacts than germline mutations and structural variations. Our analysis provides a comprehensive picture of mutations in the microsatellite regions and reveals possible causes of mutations, as well as provides a useful marker set for MSI detection.

Abstract P5-01-03: Ultrasensitive detection of minimal residual disease in patients treated for breast cancer

Abstract Background: Breast cancer (BC) may recur years to decades after initial treatment, leading to incurable, metastatic disease. Prior studies have shown blood biopsy can detect minimal residual disease (MRD), but not in all patients and often with very short lead time. Most efforts thus far have focused on tracking one or a few mutations via ctDNA, which may limit sensitivity. Here we sought to maximize the detection sensitivity of blood biopsies by tracking up to hundreds of individualized tumor mutations in cell-free DNA (cfDNA). Doing so enables us to break the detection ceiling imposed by the limited copies of each gene in the cell-free DNA in a blood draw. Methods: cfDNA was extracted from archival plasma samples (n=271) from 142 patients with stage 0-III BC enrolled prospectively onto two IRB-approved studies. We applied whole-exome sequencing (WES) to define up to several hundred mutations from each patient’s tumor. To limit potential errors, we employed strict criteria to select somatic SNVs to track using duplex sequencing in cfDNA. We required detection of ≥ 2 mutations for a cfDNA sample to be MRD+ and excluded any mutations also found in a patient’s own germline DNA. Results: We identified 142 patients treated for stage 0-III BC, who had postoperative blood and plasma samples available, and were monitored for distant recurrences for up to thirteen years. All patients had biopsy-proven BC, with 86 (61%) having HR+/HER2- BC, 31 (22%) having HR-/HER2-, or “triple negative” BC (TNBC), and 25 (18%) having HER2-positive disease. Three (2%) patients had stage 0 disease, 32 (23%) had stage I, 68 (42%) had stage II, and 39 (27%) had stage III BC at diagnosis. Archived plasma samples were collected post-operatively, at year 1, and at year 4. We created individualized assays targeting a median of 57 mutations (range 2 - 346) per patient. Reasoning that MRD status after completion of all local therapy and chemotherapy might best predict for distant recurrence, we conducted a landmark analysis at one year following surgery, and found all patients (n=6/6) with detectable MRD experienced recurrence (HR=21.2 (7.43-60.35)) in a median of 6.7 (range 3.4 - 15.8) months. Median lead time including all timepoints from first positive blood sample to recurrence was 18.9 (range: 3.4 - 39.2) months. Finally, in a multivariate model, MRD remained highly statistically significant independent of stage, subtype, and age at diagnosis. Conclusions: We present a novel approach to MRD tracking based on the premise that tracking many - rather than one or a handful of - mutations inherently increases sensitivity and that assay personalization maximizes sensitivity in a disease without multiple, recurrent mutations. To our knowledge, the lead time we show here is significantly longer than that seen in prior investigations, and if confirmed, could offer an opportunity to treat MRD long before the development of clinically apparent metastatic disease. Prospective studies are needed to determine whether earlier detection of MRD is clinically meaningful for patients. Citation Format: Heather A Parsons, Justin Rhoades, Sarah C. Reed, Greg Gydush, Priyanka Ram, Pedro Exman, Kan Xiong, Christopher Lo, Tianyu Li, Mark Fleharty, Greg Kirkner, Denisse Rotem, Ofir Cohen, Melissa Hughes, Shoshanna Rosenberg, Laura Collins, Kathy Miller, Brendan Blumenstiel, Carrie Cibulskis, Donna Neuberg, Samuel S. Freeman, Niall Lennon, Nikhil Wagle, Gavin Ha, Daniel G. Stover, Atish D. Choudhury, Gad Getz, Matthew Meyerson, Nancy U. Lin, Ian E. Krop, J. Christopher Love, G. Mike Makrigiorgos, Ann Partridge, Erika Mayer, Todd R. Golub, Viktor A. Adalsteinsson. Ultrasensitive detection of minimal residual disease in patients treated for breast cancer [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-01-03.

Abstract A051: A comprehensive, highly accurate genomics platform for precision immunotherapy: Simultaneously characterize tumors and the TME from a single FFPE sample

Abstract Immunogenomic profiling of the tumor and the tumor microenvironment (TME) is critical for identifying new biomarkers of immunotherapy response, understanding resistance, and enabling the development of personalized immunotherapies. However, running a comprehensive array of biomarker assays for each patient sample is often impractical given limited sample quantity, processing complexity, and prohibitive cost. To address these challenges, we developed a novel, augmented exome and transcriptome-based platform that simultaneously characterizes the tumor and TME from a single FFPE sample. We co-optimized the design of our sequencing assays and analytics to increase performance for the detection of somatic SNVs, indels, CNAs, and fusions across ~20,000 genes, as well as the evaluation of neoantigens, expression signatures, HLA typing and LOH, TCR/BCR repertoires, oncoviruses, tumor-infiltrating lymphocytes (TILs), clinically-actionable mutations, tumor mutational burden (TMB), and MSI status. We developed novel methods to sequence difficult regions of the exome and to extend coverage to key immunogenomic biomarkers. Analytic pipelines were designed to utilize assay optimizations to achieve higher accuracy than with other platforms. We then validated the platform for diagnostic and therapeutic use. With as little as 50ng of DNA per FFPE sample and co-extracted RNA, this platform completely covers between 17% to 40% more genes compared to a non-augmented exome, thus increasing sensitivity to somatic mutations and putative neoantigens. For neoantigen performance, we generated immune-peptidomic data from mono-allelic HLA transfected cell-lines and trained neural networks to predict neoepitope binding to MHC, demonstrating a higher precision (0.88) across alleles than publicly available tools (<0.7). For TCR alpha and beta clonotype profiling in tumor samples, we demonstrate strong correlation with the results from a targeted TCR kit (R2>0.9 and >0.94, respectively). For TILs, we developed signatures for CD4, CD8 T-cells, and other immune cells, demonstrating concordance with synthetic and CyTOF-derived validation sets. For HLA typing, we achieve an accuracy of 99.1% for HLA Class I, and 95% for HLA Class II typing calls, and have developed a novel tool for HLA LOH detection. We demonstrate sensitive detection of HPV, EBV, HCV, HTLV, and KSHV in known samples, and accurate MSI and TMB assessment. Finally, for diagnostic reporting, we achieve high sensitivity and specificity for clinically-reportable mutations comparable with diagnostic cancer panels. With this platform, we have developed a novel immunogenomics platform that can characterize both the tumor and TME from a single sample. By co-optimizing our assay and analytics for immuno-oncology, we enhance biomarker sensitivity compared to non-optimized genomics assays. Validation of the platform extends its use to diagnostics and personalized immunotherapy development. Citation Format: Robert Power, Gabor Bartha, Jason Harris, Sean Michael Boyle, Eric Levy, Pamela Milani, Prateek Tandon, Robin Li, Manjula Chinnappa, Paul McNitt, Rena McClory, Massimo Morra, Sebastian Saldivar, Michael Clark, Christian Haudenschild, Erin Newburn, Christelle Johnson, John West. A comprehensive, highly accurate genomics platform for precision immunotherapy: Simultaneously characterize tumors and the TME from a single FFPE sample [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics; 2019 Oct 26-30; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2019;18(12 Suppl):Abstract nr A051. doi:10.1158/1535-7163.TARG-19-A051

409P - Integrated genomic and transcriptomic analysis revealed mutagenic patterns of dedifferentiated liposarcoma and leiomyosarcoma in Chinese patients

BackgroundL-type sarcoma (liposarcoma + leiomyosarcoma) is the most common subtype of soft tissue sarcoma (STS). Leiomyosarcoma (LMS) and dedifferentiated LPS (DDLPS) are more aggressive with worse outcomes than well-differentiated sarcoma. Efforts have thus been made to decipher the difference of oncogenesis between DDLPS and LMS, and assist in the choice of treatment. MethodsWES was conducted on 32 STS (20 LMS and 12 DDLPS) and matched peripheral blood. Somatic SNVs and indels were identified by GATK tools. Recurrent somatic copy-number alterations (SCNA) were called using Control-FREEC and GISTIC2.0. We further conducted RNA-Seq on 8 DDLPS samples and 8 LMS samples, and use STAR-Fusion to reveal the distinct chromosome rearrangement patterns of LMS and DDLPS samples. ResultsWe identified top 20 highly variable genes between LMS and DDLPS. 55% LMS (n = 11) carry deletions or mutations of TP53. All DDLPS carry wildtype TP53, but chromosome 12q14–15 region which encodes MDM2, CDK4 and HMGA2 are highly amplified in all DDLPS samples, suggesting distinct mechanisms of TP53 pathway inactivation. Hippo pathway is reported to be involved in resistance to cytotoxic drugs and modulate tumour immunogenicity. In our study, 70% LMS (n = 14) and 58% DDLPS (n = 7) carry at least one Hippo pathway alterations, top variable genes in LMS include DCHS1, LLGL1, AJUBA et al. Among 8 LMS and 8 DDLPS who underwent RNA-Seq, we identified 4 fusion transcripts in 3 LMS and 36 fusion transcripts in 8 DDLPS (P = 0.007). Fusion transcripts involving chromosome 12 (Chr12) are only identified in DDLPS, including 10 interchromosomal and 12 intrachromosomal rearrangements. MDM2 and RAB3IP are the most common fusion partners. There was also significant correlation between MDM2/CDK4 co-amplification and Chr12 rearrangement (P < 0.001). All DDLPS with MDM2/CDK4 co-amplification showed elevated MDM2 and CDK4 co-expression. ConclusionsLMS and DDPLS carry distinct mutagenic patterns, ranging from SNVs to gross genomic instability. Chromosome rearrangement of DDLPS, particularly in Chr12, result in both gene amplification and fusion events, and these neoantigens produced by gene fusion may open up alternative avenues for immunotherapy. Legal entity responsible for the studyThe authors. FundingHas not received any funding. DisclosureAll authors have declared no conflicts of interest.

Comprehensive Liquid Profiling of Circulating Tumor DNA and Protein Biomarkers in Long-Term Follow-Up Patients with Hepatocellular Carcinoma

Circulating tumor DNA (ctDNA) provides a novel approach for detecting tumor burden and predicting clinical outcomes of hepatocellular carcinoma (HCC). Here, we performed a thorough evaluation of HCC circulating genetic features and further fully integrated them to build a robust strategy for HCC monitoring and prognostic outcome assessment. We performed target sequencing and low-coverage whole-genome sequencing on plasma samples collected from 34 long-term follow-up patients with HCC to capture tumor somatic SNVs and CNVs, respectively. Clinical information was also obtained to evaluate the prognostic performance of ctDNA comparing with clinically applied protein biomarkers. All plasma samples before surgery showed somatic genetic variations resembling corresponding tumor tissues. During follow-up, SNVs and CNVs dynamically changed correlating to patients' tumor burden. We integrated the comprehensive ctDNA mutation profiles to provide a robust strategy to accurately assess patients' tumor burden with high consistence comparing with imaging results. This strategy could discover tumor occurrence in advance of imaging for an average of 4.6 months, and showed superior performance than serum biomarkers AFP, AFP-L3%, and Des-Gamma-Carboxy Prothrombin (DCP). Furthermore, our strategy could precisely detect minimal residual disease (MRD) in advance and predict patients' prognostic outcomes for both relapse-free survival (P = 0.001) and overall survival (P = 0.001); further combining ctDNA with DCP could increase the sensitivity for MRD detection. We demonstrated that plasma CNV and SNV levels dynamically correlated with patients' tumor burden in HCC. Our strategy of comprehensive mutation profile integration could accurately and better evaluate patients' prognostic risk and detect tumor occurrence in advance than traditional strategies.

Abstract 1371: Assessing the utility of cell-free DNA in identifying prostate cancer and characterizing tumor heterogeneity via whole exome and whole genome, multi-region sequencing

Abstract Early cancer diagnosis, especially while the disease is still localized and before symptoms appear, results in significantly higher survival rates compared to late-stage diagnosis. At the time of diagnosis, it is also common to find multiple foci within a single prostate gland in men with localized disease. Cell-free DNA (cfDNA) may not only reflect underlying disease biology, but also simultaneously allow for the identification of genetically distinct tumor subclones. The objectives of this study are to determine if 1) cfDNA levels are able to distinguish between healthy individuals from patients with localized or metastatic castration-resistant prostate cancer (mCRPC), and 2) if somatic mutations identified in tumor tissue are detectable in cfDNA and representative of the distribution observed in tumor tissue. This study includes samples from 130 individuals at UCSF: 21 healthy donors, 100 patients with localized prostate cancer who underwent radical prostatectomy (RP), and 9 mCRPC patients. Blood samples and matched tissue from adjacent normal seminal vesicles and multiple tumor regions (1-9 samples per patient) were collected from patients undergoing radical prostatectomy. CfDNA was extracted from plasma, and the concentration and fragment length distribution were measured with a Bioanalyzer 2100. Comparisons of cfDNA levels between groups were assessed with a Welch’s t-test due to the potential for unequal variances. Fifty-seven samples from nine patients have been subjected to whole exome sequencing, and 22 samples from five patients have been subjected to whole genome sequencing at ~40x coverage. Somatic variant calling was performed with Broad Institute’s Firecloud platform (GATK4/MuTect2) for tumor tissue and with the Curio platform for cfDNA to build consensus sequences leveraging unique molecular tags. CfDNA levels are able to distinguish between healthy and localized (p = 0.005), as well as healthy and metastatic groups (p = 0.043). Tumor foci within a patient’s prostate gland are genetically heterogeneous, with the majority of somatic mutations private to tumor regions and a subset of mutations at the intersection of these regions. Preliminary analyses result in an average of 72 somatic SNVs and indels per tumor tissue region, with ~10% overlap between regions in the same patient. Additional mCRPC and follow-up blood samples are being collected from all patients. The association between cfDNA levels prior to RP surgery and biochemical recurrence will be investigated when follow-up collection concludes. Further analysis of mutational concordance, clonality, and copy number variation between tumor tissue DNA and cfDNA, along with clinical data, will be performed. Citation Format: Emmalyn Chen, Clinton Cario, Lancelote Leong, Karen Lopez, Patricia Li, Erica Oropeza, Imelda Tenggara, Janet Cowan, Jeffry Simko, Daniel Wells, Robin Kageyama, June Chan, Terence Friedlander, Pamela Paris, Peter Carroll, John Witte. Assessing the utility of cell-free DNA in identifying prostate cancer and characterizing tumor heterogeneity via whole exome and whole genome, multi-region sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1371.

Deep sequencing of a recurrent oligodendroglioma and the derived xenografts reveals new insights into the evolution of human oligodendroglioma and candidate driver genes.

ABSTRACTWe previously reported the establishment of a rare xenograft derived from a recurrent oligodendroglioma with 1p/19q codeletion. Here, we analyzed in detail the exome sequencing datasets from the recurrent oligodendroglioma (WHO grade III, recurrent O2010) and the first-generation xenograft (xenograft1). Somatic SNVs and small InDels (n = 80) with potential effects at the protein level in recurrent O2010 included variants in IDH1 (NM_005896:c.395G>A; p. Arg132His), FUBP1 (NM_003902:c.1307_1310delTAGA; p.Ile436fs), and CIC (NM_015125:c.4421T>G; p.Val1474Gly). All but 2 of these 80 variants were also present in xenograft1, along with 7 new variants. Deep sequencing of the 87 SNVs and InDels in the original tumor (WHO grade III, primary O2005) and in a second-generation xenograft (xenograft2) revealed that only 11 variants, including IDH1 (NM_005896:c.395G>A; p. Arg132His), PSKH1 (NM_006742.2:c.650G>A; p.Arg217Gln), and SNX12 (NM_001256188:c.470G>A; p.Arg157His), along with a variant in the TERT promoter (C250T, NM_198253.2: c.-146G>A), were already present in primary O2005. Allele frequencies of the 11 variants were calculated to assess their potential as putative driver genes. A missense change in NDST4 (NM_022569:c.2392C>G; p.Leu798Val) on 4q exhibited an increasing allele frequency (~ 20%, primary O2005, 80%, recurrent O2010 and 100%, xenograft1), consistent with a selection event. Sequencing of NDST4 in a cohort of 15 oligodendrogliomas, however, revealed no additional cases with potential protein disrupting variants. Our analysis illuminated a tumor evolutionary series of events, which included 1p/19q codeletion, IDH1 R132H, and TERT C250T as early events, followed by loss of function of NDST4 and mutations in FUBP1 and CIC as late events.

Performance and validation of a tumor mutation profiling, based on artificial intelligence annotation, to assist oncology decision making.

e13148 Background: Tumor mutation profiling has become a key component for orienteering the treatment of oncologic patients. A crucial step for this is the correct identification and classification of pathogenic and actionable variants. In the present work we aimed at the development and validation of a tumor mutation profiling panel, based on NGS, which uses artificial intelligence for variant annotation. Methods: We designed a hybrid capture panel, containing 366 genes to evaluate somatic SNVs, INDELs and CNVs, and to calculate TMB, using a customized bioinformatics pipeline. MSI status was determined by fragment analysis using capillary electrophoresis. Analytical performance was determined using reference cell lines. FFPE samples from 70 tumors were accessed and 53 were sequenced. Variant annotation was performed by IBM Watson for Genomics (WfG) platform. Assay performance on clinical samples was defined based on orthogonal assays using Agilent CGH+SNParray 400K (for CNVs only) and Foundation One test (Foundation Medicine) (F1). Results: Breast, colon and lung were the most common tumor origins. Fifty-three samples were successfully sequenced, while 41 of them could also be analyzed by F1 test. A summary of the assay performance is presented in Table 1. Our pipeline detected 1219 variants and 290 (23%) were classified as Pathogenic, Likely Pathogenic or Actionable, according to WfG. Thirty-five samples (66%) presented a variant that could drive the treatment, with 37.7% of samples being sensitive to targeted therapies, while 22.6% were resistant; additionally, 86% had an indication for a clinical trial. Conclusions: The developed assay presented a good overall sensitivity and allele frequency correlation, with TMB and MSI having the best rates. Comparisons with F1 had reduced values of concordance; however, SNVs and INDELs presented a similar frequency. Differences on CNVs identification may rely on distinct thresholds established by the different groups. The high percentage of samples that could benefit from mutational profiling highlights the importance of such approach in the clinical routine. Additionally, the high number of variants features the need for updated information for annotation. [Table: see text]

A novel patient derived orthotopic xenograft model of gastro-esophageal junction cancer: Key platform for translational discoveries.

64 Background: Mouse models of gastroesophageal junction (GEJ) cancer strive to recapitulate the intratumoral heterogeneity and cellular crosstalk within patient tumors to improve clinical translation. Current GEJ models have limited applications in tumor microenvironment, immune oncology and metastatic studies. Methods: A novel patient derived tumor orthotopic xenograft (PDOX) was established from GEJ cancer via surgical implantation. Patient tumor was compared to subcutaneously implanted PDX and PDOX by H&amp;E, IHC, and next generation sequencing (T200.1 panel). Drug efficacy studies of 5-flurouracil with and without radiotherapy are being performed. Results: Mechanical abrasion of mouse GEJ prior to implantation of patient derived tumor in situ promotes tumor engraftment (100%, n = 6). Complete PDOX engraftment was observed with rapid intra and extra luminal tumor growth as evidenced by MRI. Patient derived stroma co-engrafts with tumor cells in GEJ-PDOX. PDOXs contain fibroblasts, immune and inflammatory cells, vascular and lymphatic vessels. Stromal hallmarks of aggressive GEJs are recapitulated in GEJ-PDOX mouse model. PDOXs demonstrates tumor invasion into vasculature. GEJ-PDOXs is a clinically relevant model for metastases and immunological studies. Next generation sequencing with the T200.1 revealed that the loss of heterozygosity of NOTCH3, TGFβ1, EZH2, and MLL3 are maintained with similar allelic frequency between the patient tumor and the xenografts. Additional somatic SNVs such as ARID1A, NSD1 (CDS157-158), NSD1 (CDS158-159), KDM6A, XPO1, MAPK1 and EGFR were found to be acquired in xenograft tumor tissues that were not observed in patient tumor. Drug and radiation efficacy studies are ongoing, tumor response to radiation was observed. Conclusions: A GEJ-PDOX model exhibits remarkable fidelity to human disease and captures the precise tissue microenvironment present within the local GEJ architecture facilitating it as a novel tool in translating to clinics. This model can be applied to address importance of tumor microenvironment in metastatic and immunological studies, and to develop novel therapeutic approaches for the treatment of GEJ cancer.

Dynamical detection and characterization of circulating tumor DNA in patients with hepatocellular carcinoma

Introduction: Circulating tumor DNA (ctDNA) provides a novel approach for detecting tumor burden and predicting clinical outcomes of hepatocellular carcinoma (HCC). Here we performed a thorough evaluation of HCC circulating genetic features and further fully integrated them to build a robust strategy for HCC surveillance and prognostic outcome assessment. Method: We performed target sequencing and low coverage whole genome sequencing on 168 plasma samples collected from 34 long-term follow-up HCC patients to capture tumor somatic SNVs and CNVs, respectively. Clinical information was also obtained to evaluate the prognostic performance of ctDNA comparing with clinically applied protein biomarkers. Result: All plasma samples before surgery showed somatic genetic variations resembling corresponding tumor tissues. During follow-up, SNVs and CNVs dynamically changed correlating to patients’ tumor burden. We integrated the comprehensive ctDNA mutation profiles to provide a robust strategy to accurately assess patients’ tumor burden with high consistence comparing to imaging results. This strategy could discover tumor occurrence in advance of imaging for an average of 4.6 months, and showed superior performance than serum biomarkers AFP, AFP-L3% and DCP. Furthermore, our strategy could precisely detect MRD in advance and predict patients’ prognostic outcomes for both RFS (p=0.001) and OS (p=0.001); further combining ctDNA with DCP could increase the sensitivity for MRD detection. Conclusion: We demonstrated that plasma CNV and SNV levels dynamically correlated with patients’ tumor burden in HCC. Our strategy of comprehensive mutation profile integration could accurately and better evaluate patients’ prognostic risk and detect tumor occurrence in advance than traditional strategies.

SU94 - EXPLORING THE SOMATIC VARIATION IN SCHIZOPHRENIA

Background The presence of somatic variation in the human brain has been elucidated over the past several decades. Most recently, single-cell whole-genome sequencing studies have provided unequivocal evidence that individual human brain cells harbor unique genetic variants. The functional significance of this phenomenon is unclear, but some studies have suggested that it contributes to risk of neuropsychiatric disease. Currently, cost and labor requirement render difficult the sequencing of individual genomes at the scale that will likely be necessary to fully uncover the extent and role of somatic variation in the brain. A scalable alternative, however, has been developed in the field of cancer genomics wherein somatic variants are identified by comparing the frequencies of variant alleles in tumor and normal bulk tissue specimens from the same individual. Here, we apply this approach to investigate the role of somatic variation in schizophrenia in a pilot study of 9 individuals (5 schizophrenia cases and 4 unaffected controls). Methods Specimens from the prefrontal cortex and temporal muscle of 5 schizophrenia cases and 4 controls were obtained from the Mount Sinai NIH Brain and Tissue Repository. Deep whole-exome sequencing (250X coverage) was performed using DNA extracted from neuronal and non-neuronal nuclei, which were isolated from cortical specimens by fluorescent-activated nuclear sorting. We followed standard protocols for mapping and filtering reads, then called somatic Single Nucleotide Variants (sSNVs) using a suite of 6 calling algorithms. For each individual, the calling algorithms were utilized to call variants in the exomes of both the neuronal and non-neuronal cells of the brain by comparing them to a matched “normal” sample from the same individual (i.e., the temporal muscle). Variants were then filtered using an in-house quality control pipeline, and a subset of the remaining calls were validated with digital PCR (dPCR) and parallel sequencing of multiple clones (clone-seq). Results We identified 26 sSNVs in neuronal and non-neuronal cells from the brains of 9 individuals. We chose 10 of 26 for experimental validation using dPCR and clone-seq, observing a validation rate of 80%. We noted that 13 of the 26 sSNVs were found in a single schizophrenia case; as such, 4 of the 10 sSNVs included in validation experiments were from this individual, 3 of which were successfully validated. For 5 sSNVs, only the neuronal exome of the individual had the variant, for 17 only the non-neuronal exome and for 4 both the neuronal and non-neuronal exomes. At least 1 sSNV was observed in all 5 of the cases and in 2 of the 4 controls. We noted that 2 of the 23 somatic SNVs in cases fell within one of the 16 copy number variants implicated in schizophrenia by the largest study of copy number variation in schizophrenia to date. Several of the sSNVs observed in our cohort fall within genes known to function in determining cellular fate and development, including WNT10B, SMAD1 and FGF1. Discussion Here, we present a pipeline suitable for the large-scale investigation of somatic variation in the human brain. We show evidence that somatic variation occurs in a cell-type specific manner in the human brain. Genes affected by sSNVs in the brain have known roles in determining cell fate. Greater number of sSNVs in schizophrenia cases compared to controls was observed, and multiple case sSNVs fall within known schizophrenia CNVs. Larger sample sizes are needed so as to further elucidate the extent and role of somatic variation in the human brain.

Detection of Clinically Relevant Molecular Alterations in Chronic Lymphocytic Leukemia (CLL) By Nanopore Sequencing

IntroductionChronic Lymphocytic Leukaemia (CLL) is the most prevalent leukaemia in the Western world and characterised by clinical heterogeneity. IgHV mutation status, mutations in the TP53 gene and deletions of the p-arm of chromosome 17 are currently used to predict an individual patient's response to therapy and give an indication as to their long-term prognosis. Current clinical guidelines recommend screening patients prior to initial, and any subsequent, treatment. Routine clinical laboratory practices for CLL involve three separate assays, each of which are time-consuming and require significant investment in equipment. Nanopore sequencing offers a rapid, low-cost alternative, generating a full prognostic dataset on a single platform. In addition, Nanopore sequencing also promises low failure rates on degraded material such as FFPE and excellent detection of structural variants due to long read length of sequencing. Importantly, Nanopore technology does not require expensive equipment, is low-maintenance and ideal for patient-near testing, making it an attractive DNA sequencing device for low-to-middle-income countries.MethodsEleven untreated CLL samples were selected for the analysis, harbouring both mutated (n=5) and unmutated (n=6) IgHV genes, seven TP53 mutations (five missense, one stop gain and one frameshift) and two del(17p) events. Primers were designed to amplify all exons of TP53, along with the IgHV locus, and each primer included universal tails for individual sample barcoding. The resulting PCR amplicons were prepared for sequencing using a ligation sequencing kit (SQK-LSK108, Oxford Nanopore Technologies, Oxford, UK). All IgHV libraries were pooled and sequenced on one R9.4 flowcell, with the TP53 libraries pooled and sequenced on a second R9.4 flowcell. Whole genome libraries were prepared from 400ng genomic DNA for each sample using a rapid sequencing kit (SQK-RAD004, Oxford Nanopore Technologies, Oxford, UK), and each sample sequenced on individual flowcells on a MinION mk1b instrument (Oxford Nanopore Technologies, Oxford, UK). We developed a bespoke bioinformatics pipeline to detect copy-number changes, TP53 mutations and IgHV mutation status from the Nanopore sequencing data. Results were compared to short-read sequencing data obtained earlier by targeted deep sequencing (MiSeq, Illumina Inc, San Diego, CA, USA) and whole genome sequencing (HiSeq 2500, Illumina Inc, San Diego CA, USA).ResultsFollowing basecalling and adaptor trimming, the raw data were submitted to the IMGT database. In the absence of error correction, it was possible to identify the correct VH family for each sample; however the germline homology was not sufficient to differentiate between IgHVmut and IgHVunmut CLL cases. Following bio-informatic error correction and consensus building, the percentage to germline homology was the same as that obtained from short-read sequencing and nanopore sequencing also called the same productive rearrangements in all cases.A total of 77 TP53 variants were identified, including 68 in non-coding regions, and three synonymous SNVs. The remaining 6 were predicted to be functional variants (eight missense and two stop-gains) and had all been identified in early MiSeq targeted sequencing. However, the frameshift mutation was not called by the analysis pipeline, although it is present in the aligned reads. Using the low-coverage WGS data, we were able to identify del(17p) events, of 19Mb and 20Mb length, in both patients with high confidence.ConclusionsHere we demonstrate that characterization of the IgHV locus in CLL cases is possible using the MinION platform, provided sufficient downstream analysis, including error correction, is applied. Furthermore, somatic SNVs in TP53 can be identified, although similar to second generation sequencing, variant calling of small insertions and deletions is more problematic. Identification of del(17p) is possible from low-coverage WGS on the MinION and is inexpensive. Our data demonstrates that Nanopore sequencing can be a viable, patient-near, low-cost alternative to established screening methods, with the potential of diagnostic implementation in resource-poor regions of the world. DisclosuresSchuh:Giles, Roche, Janssen, AbbVie: Honoraria.

Clinical cancer genomic profiling by three-platform sequencing of whole genome, whole exome and transcriptome

To evaluate the potential of an integrated clinical test to detect diverse classes of somatic and germline mutations relevant to pediatric oncology, we performed three-platform whole-genome (WGS), whole exome (WES) and transcriptome (RNA-Seq) sequencing of tumors and normal tissue from 78 pediatric cancer patients in a CLIA-certified, CAP-accredited laboratory. Our analysis pipeline achieves high accuracy by cross-validating variants between sequencing types, thereby removing the need for confirmatory testing, and facilitates comprehensive reporting in a clinically-relevant timeframe. Three-platform sequencing has a positive predictive value of 97–99, 99, and 91% for somatic SNVs, indels and structural variations, respectively, based on independent experimental verification of 15,225 variants. We report 240 pathogenic variants across all cases, including 84 of 86 known from previous diagnostic testing (98% sensitivity). Combined WES and RNA-Seq, the current standard for precision oncology, achieved only 78% sensitivity. These results emphasize the critical need for incorporating WGS in pediatric oncology testing.

Abstract 2085: Exploring mutation signatures in pediatric cancers

Abstract Background: Mutation signatures have been catalogued across most common cancers, providing insights into their underlying mutational processes and suggesting strategies for categorization and subphenotyping. However, these analyses have been limited to mostly adult cancers, likely due to the relative rarity of pediatric analogs and a corresponding lower mutation burden of these tumors. Here we attempted to compare mutation signatures across several pediatric cancers, given these constraints, with their corresponding adult versions. Data and Methods: From the ICGC data portal, we downloaded clinical and mutation data for pediatric brain (medulloblastoma, glioblastoma, neuroblastoma) and blood-based (acute lymphoblastic leukemia, acute myeloid leukemia) cancers. We also obtained data for related adult cancers, including glioblastoma and acute myeloid leukemia. For each cancer, we characterized mutation patterns using trinucleotides centered at somatic SNVs and then estimated the composition of underlying putative COSMIC mutation signatures. For cancers with low mutation rates (impeding estimates of mutation signatures), we grouped mutations from similar samples based on clinical and molecular statuses to infer mutation signatures. Results: Each of the pediatric cancer types strongly exhibited the common cancer mutation Signature 1 (spontaneous deamination of 5-methylcytosine), as might be expected since it has been described across all adult cancer types. Further, pediatric acute myeloid leukemia and glioblastoma both harbored mutation signatures seen in their adult counterparts in addition to unexpected patterns of defective DNA mismatch repair and DNA double-strand break repair. Interestingly, for pediatric glioblastoma we identified Signatures 3, 8, 15 and 16, which were previously unidentified in adult glioblastoma and have been recently implicated in pediatric medulloblastoma. Conclusion: Our findings may be influenced by the sparsity of mutations in pediatric cancers, highlighting the need for quantifying precision in measures of mutation signatures. This notwithstanding, from our initial observations, pediatric cancers appear to exhibit a more diverse set of mutation signatures than their adult counterparts. Further, pediatric cancers of different histologies, particularly those of brain cancers, may share molecular mechanisms more than currently believed. Citation Format: Francis A. San Lucas, Philip Lupo, Austin Brown, Michael Scheurer, Paul Scheet. Exploring mutation signatures in pediatric cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2085.

Abstract 636: Accounting for processed pseudogene-related artifacts improves specificity of clinical cancer diagnostic sequencing

Abstract Background: Duplicated genomic regions pose a challenge to accurate variant calling in clinical sequencing applications as duplicate-specific variants may incorrectly be assigned to the target. One source of duplicated coding sequences are processed pseudogenes (PPGs) that can originate from LINE-mediated reverse transcription and genomic integration of processed mRNA, resulting in partial or complete copies of the original gene, lacking intronic sequences. False-positive variants resulting from pseudogenes found in the reference genome, such as those of PIK3CA and PTEN, have been well studied; however, the recent discovery of rare and even individual-specific cancer-related PPGs demonstrates a need for more systematic interrogation and mediation of PPG-related clinical artifacts on a sample-by-sample basis. Methods: We analyzed germline and somatic alterations in over 15,000 clinical circulating cell-free DNA samples sequenced using a 73-gene panel (Guardant Health, CA) focused primarily on clinically-relevant somatic alterations. All targeted genes were interrogated for two key hallmarks of PPGs: 1) split sequencing reads mapping to adjacent exons without intervening intronic sequence and 2) atypically increased sequencing coverage of exons relative to adjacent introns.Results: Applying a strict cutoff of ≥5 unique molecules supporting any putative intron-skipping event, we observed evidence of 94 PPGs in 91 of 15,315 analyzed samples (0.59%), affecting SMAD4 (57 samples), CDK4 (20), GNAS (6), HRAS (2), KRAS (1), TP53 (1), ESR1 (1), STK11 (1), RB1 (1), MAPK3 (1), MAPK1 (1), NF1 (1), and GATA3 (1). To determine the impact of PPGs on specificity, we applied a resampling test to assess whether somatic variant calling without PPG aware detection of SNVs results in over-representation of calls near splice junctions in samples harboring PPGs. Indeed, SMAD4 somatic SNVs were identified in 30% of samples containing SMAD4 PPGs as compared to an expected 0.7% of non-PPG samples (two-tailed p &amp;lt; 10E-4). Similarly, 10% of samples containing CDK4 PPGs also showed somatic CDK4 SNVs, whereas only 0.9% would be expected by chance (two-tailed p = 0.03). These results strongly suggest that PPG-related artifacts may be misclassified as true somatic events in these contexts. Algorithmically restricting variant-calling only to those reads containing intronic sequence can maintain high specificity with only minimal cost to sensitivity in most exons. Conclusions: Our analysis reveals the existence of rare PPGs derived from over a dozen common clinical cancer sequencing targets and provides strong evidence that these may produce false positive SNV calls in an appreciable number of cases. Importantly, we have also shown that PPGs can be detected in a sample-specific manner based on the presence of intron-skipping split reads, enabling algorithmic detection and elimination of artifacts. Citation Format: Carlo G. Artieri, Marcin Sikora, Alex Artyomenko, Elena Helman, Darya Chudova, Richard Lanman, AmirAli Talasaz. Accounting for processed pseudogene-related artifacts improves specificity of clinical cancer diagnostic sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 636.

Abstract 922: Access, visualize and analyze 5,000 whole-genomes from pediatric cancer patients on St. Jude Cloud

Abstract While whole-genome (WGS), whole-exome (WES), and RNA-Seq data of patient samples are key resources for the development of precision medicine, major computing infrastructure is typically required to use them effectively. The St Jude Cloud (SJCloud, https://stjude.cloud), built in collaboration with DNAnexus and Microsoft, aims to remove this barrier by sharing genomic sequencing data generated at St Jude Children's Research Hospital, making complex bioinformatics pipelines easily accessible, and providing intuitive visualizations for data mining in the cloud. Over 5000 WGS, 6000 WES and 1500 RNA-Seq from &amp;gt;5,000 pediatric cancer patients mapped to the latest reference genome are securely available in SJCloud. These data were generated from three St Jude-funded genomic initiatives: the Pediatric Cancer Genome Project (PCGP), the St Jude Life Genome Project, and the Genomes for Kids Clinical Trial. SJCloud hosts BAM files, coding and non-coding somatic and germline SNVs and indels, copy number (CNV) and structural alterations (SV). Non-identifiable data (e.g. somatic alterations, genotype frequency, cancer diagnosis and demographics) can be viewed immediately using our interactive genome browser, while raw data and individual genotype access requires a simple online approval. Data synchronization and visualization enables novel discoveries by non-bioinformaticians. For example, a genomic view of the TERT locus shows enrichment of CNVs and SVs in neuroblastoma (NBL), consistent with reports of activation via rearrangement. The same view also shows a somatic promoter mutation, C228T, in one NBL; such mutations have not been reported in primary samples to our knowledge. This integrated view across somatic mutation types enables evaluation of the diverse genetic mechanisms deregulating cancer genes. SJCloud also facilitates data re-analysis. We ported the “MutationalPatterns” R package (Blokzijl et al. 2017) to the cloud to elucidate major mutational signatures in &amp;gt;500,000 PCGP WGS somatic variants. Inclusion of non-coding mutations was critical as the low number of exonic mutations in some pediatric cancers is insufficient for robust analysis. A surprising finding was a signature consistent with ultraviolet-induced DNA damage in a subset of B-acute lymphoblastic leukemia. End-to-end workflows to detect gene fusions, predict neoepitopes, classify mutations, process ChIP-seq, and identify differentially expressed genes are also freely accessible. By integrating analytic tools with the world's largest set of pediatric genomics data, SJCloud enables data sharing and mining, innovative genomic analysis, and development of new analytic methods. We anticipate that in 2019 we will host data from over 10,000 pediatric cancer patients, and we are actively exploring approaches to make this a federated data repository capable of interchange with the global pediatric cancer research community. Citation Format: Scott Newman, Xin Zhou, Clay McLeod, Michael Rusch, Gang Wu, Edgar Sioson, Shuoguo Wang, J. Robert Michael, Aman Patel, Michael N. Edmonson, Andrew Frantz, Ti-Cheng Chang, Yongjin Li, Robert I. Davidson, Singer Ma, Irina McGuire, Nedra Robison, Xing Tang, Lance Palmer, Ed Suh, Leigh Tanner, James McMurry, Keith Perry, Zhaoming Wang, Carmen Wilson, Yong Cheng, Mitch Weiss, Leslie L. Robison, Yutaka Yasui, Kim E. Nichols, David W. Ellison, James R. Downing, Jinghui Zhang. Access, visualize and analyze 5,000 whole-genomes from pediatric cancer patients on St. Jude Cloud [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 922.

Abstract LB-138: Zero Childhood Cancer: A comprehensive precision medicine platform for children with high-risk cancer

Abstract Molecular genomics analyses aim to identify the subset of patients harbouring actionable mutations as a pathway to better targeted treatment selection. Low mutation rates and the paucity of clinical data linking targeted treatments with mutations in paediatric cancer suggests genomic analysis alone has limitations for translation into clinical benefit. The Zero Childhood Cancer program aims to assess the feasibility of a precision medicine platform to identify targeted therapeutic agents for high-risk (HR) paediatric malignancies (expected survival &amp;lt;30%). We combine molecular genomic analysis (WGS (tumour, germline DNA), deep sequencing of a panel of cancer associated genes, and whole transcriptome (RNASeq)) with in vitro high-throughput drug screening, and patient-derived xenograft (PDX) drug efficacy testing, followed by assessment of and recommendations made by a national Multidisciplinary Tumour Board (MTB). The Pilot Feasibility Study enrolled 59 patients from June 2015 to October 2017 with a range of tumour types (47% central nervous system tumours, 20% sarcoma, 12% leukaemia, 9% neuroblastoma, 12% other rare cancers). The median age was 9 years (range 1-21 years) and 49% of patients were enrolled at diagnosis and 51% at relapse. In 54 curated cases, the complete molecular platform identified reportable somatic SNVs, fusions, and CNVs in 56%, 24% and 39% of patients, respectively. 5 patients had a reportable germline cancer predisposition variant and in 3 patients, the genomic findings changed the primary diagnosis. Fresh tissue collection in 48 of cases resulted in successful in vitro high-throughput drug screening (112 compound library single agent) in 23 cases and 23 successful patient-derived xenograft model engraftments. Overall, 57% of patients received a personalised medicine recommendation (targeted therapy, change in diagnosis, germline mutation referral), and 10 patients with a therapy recommendation went on to receive the recommended therapy. The Zero Childhood Cancer Pilot demonstrates the likelihood that oncologists will more frequently use a recommendation to benefit individual patients on the clinical trial who have typically exhausted other therapeutic approaches. Now, a national multicentre prospective study of the feasibility and clinical value of identifying therapeutic targets and recommending personalised treatment for children and adolescents with high-risk cancer (PRISM) opened in late 2017 for Australians up to 21 years with HR cancer. Citation Format: Emily Mould, Loretta Lau, Greg Arndt, Paulette Barahona, Mark Cowley, Paul Ekert, Tim Failes, Jamie Fletcher, Andrew Gifford, Michelle Haber, Alvin Kamili, Amit Kumar, Richard Lock, Glenn Marshall, Chelsea Mayoh, Scott Mead, Murray Norris, Tracey O'Brien, Mark Pinese, Dong Anh Khuong Quang, Toby Trahair, Maria Tsoli, Katherine Tucker, Meera Warby, Marie Wong, Jinhan Xie, David Ziegler, Vanessa Tyrrell. Zero Childhood Cancer: A comprehensive precision medicine platform for children with high-risk cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr LB-138.

Germline cytoskeletal and extra-cellular matrix-related single nucleotide variations associated with distinct cancer survival rates

BackgroundHuman mutagenesis has a large stochastic component. Thus, large coding regions, especially cytoskeletal and extra-cellular matrix protein (CECMP) coding regions are particularly vulnerable to mutations. Recent results have verified a high level of somatic mutations in the CECMP coding regions in the cancer genome atlas (TCGA), and a relatively common occurrence of germline, deleterious mutations in the TCGA breast cancer dataset. MethodsThe objective of this study was to determine the correlations of CECMP coding region, germline nucleotide variations with both overall survival (OS) and disease-free survival (DFS). TCGA, tumor and blood variant calling files (VCFs) were intersected to identify germline SNVs. SNVs were then annotated to determine potential consequences for amino acid (AA) residue biochemistry. ResultsGermline SNVs were matched against somatic tumor SNVs (i.e., tumor mutations) over twenty TCGA datasets to identify 23 germline-somatic matched, deleterious AA substitutions in coding regions for FLG, TTN, MUC4, and MUC17. ConclusionsThe germline-somatic matched SNVs, in particular for MUC4, extensively implicated in cancer development, represented highly, statistically significant effects on OS and DFS survival rates. The above results contribute to the establishment of what is potentially a new class of inherited cancer-facilitating genes, namely dominant negative tumor suppressor proteins.

Discovery of targetable genetic alterations in advanced non-small cell lung cancer using a next-generation sequencing-based circulating tumor DNA assay

Next-generation sequencing (NGS)-based circulating tumor DNA (ctDNA) assays have provided a new method of identifying tumor-driving genes in patients with advanced non-small cell lung carcinoma (NSCLC), especially in those whose cancer tissues are unavailable or in those that have acquired treatment resistance. Here, we describe a total of 119 patients with advanced EGFR-TKI-naive NSCLC and 15 EGFR-TKI-resistant patients to identify somatic SNVs, small indels, CNVs and gene fusions in 508 tumor-related genes. Somatic ctDNA mutations were detected in 82.8% (111/134) of patients in the total cohort. Of the 119 patients with advanced NSCLC, 27.7% (33/119) were suitable for treatment with National Comprehensive Cancer Network (NCCN) guideline-approved targeted drugs. Actionable genetic alterations included 25 EGFR mutations, 5 BRAF mutations, and 1 MET mutation, as well as 1 EML4-ALK gene fusion and 1 KIF5B-RET gene fusion. In 19.3% (23/119) of the patients, we also identified genomic alterations with that could be targeted by agents that are in clinical trials, such as mTOR inhibitors, PARP inhibitors, and CDK4/6 inhibitors. Additionally, the EGFR T790M mutation was found in 46.7% (7/15) of the patients with EGFR-TKI-resistant NSCLC, suggesting that the NGS-based ctDNA assay might be an optional method to monitor EGFR-TKI resistance and to discover mechanisms of drug resistance.