Virtual Library
Start Your Search
V.N. Tran
Moderator of
-
+
MS11 - Next Generation Technology for Detection and Treatment of Lung Cancer (ID 28)
- Event: WCLC 2013
- Type: Mini Symposia
- Track: Biology
- Presentations: 4
- Moderators:S. Toyooka, V.N. Tran
- Coordinates: 10/29/2013, 14:00 - 15:30, Bayside Auditorium B, Level 1
-
+
MS11.1 - Next Generation Sequencing (ID 506)
14:05 - 14:25 | Author(s): R. Govindan
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MS11.2 - Validating Platforms for Routine Clinical Use (ID 507)
14:25 - 14:45 | Author(s): P. Waring
- Abstract
- Presentation
Abstract
In this presentation, we describe the process that our laboratory followed that led to successful accreditation, by the Australian National Association of Testing Authorities (NATA), for medical use of NGS in clinical practice. First, we evaluated three different amplicon –based MPS technologies in order to choose the platform of choice for clinical use. We compared the performance of two commercial somatic mutation panels (Life Technology’s AmpliSeq cancer panel and Illumina’s TruSeq amplicon panel) and a customized panel (Agilent’s HaloPlex). The panels shared 31 genes in common. The AmpliSeq panel was sequenced using the Ion Torrent platform and the TruSeq and HaloPlex panels were sequenced using the Illumina MiSeq platform. In-house bioinformatics and variant annotation and reporting pipeline were developed to allow data from all three panels to be compared. A training set of 28 FFPET samples with known missense or deletion mutations in EGFR, KRAS, BRAF, NRAS, PIK3CA and KIT were tested by all three panels. These samples, previously tested using NATA - accredited Sanger sequencing, SNaPshot and fragment analysis performed on an ABI3730, were used to empirically determine the parameters required for accurate mutation detection by MPS. Sample acceptance criteria included samples with at least 1mg of extractable DNA following macrodissection from tumour areas with at least 70% purity. Library quality was assessed by BioAnalyser and libraries were sequenced to a median depth of 2000x. The panels and platforms were compared for % aligned reads, % on - target reads, median and range of coverage, input DNA quantity and quality requirements, data quality and variability, cost, turn around time, ease of use, and accuracy of mutation detection. There was marked variation in the number and types of variants identified across the three panels. With minimum variant calling criteria of depth >50x, variant depth >20x, variant frequency >5% and base quality >15, we identified 18557 variants with AmpliSeq, 15064 variants with TruSeq and 3326 variants with Haloplex. 14229 of the TruSeq variants were SNPs (9319 were C>T), indicating DNA polymerase errors, while 12370 of the AmpliSeq variants were small indels (mostly in homopolymeric tracts) indicating errors in calling repetitive sequences. In total, 59 variants were identified by all three panels. The TruSeq and Ampliseq panels detected all 31 known somatic mutations, where as the HaloPex panel missed four mutations due to patchy on - target coverage. In panels with adequate coverage of regions of interest, the assay sensitivity was 100%. The TruSeq panel was chosen for clinical use, despite the requirement for higher DNA input (150ng compared to 10ng for AmpliSeq), primarily due to ease of use and less hands - on time by laboratory staff. We then performed reproducibility, repeatability, robustness and limit of detection experiments using the TruSeq panel. Initially, there was poor reproducibility of all variants, particularly SNVs, especially in samples with low input DNA (<50ng) or poor quality DNA (fragment size <250 bp). Most of the identified variants were random and present at low frequency, most being present at <1-2% allele frequency. These showed characteristics suggestive sequencing and polymerase errors, formalin – induced artifacts and misaligned repetitive sequences. To reduced the great excess of false positives, we restricted variant calling by establishing minimum allele frequencies, by eliminating unreported variants and by limiting alignment to clinically- relevant or actionable mutations. Variant reproducibility was increased to 38% by only calling SNVS >5% and indels >1% allele frequency that were contained within the COSMIC database. This was further increased to 92% by restricting variant calling to known clinically - relevant mutations listed on the www.mycancergenome web site. Reproducibility was increased further by strict adherence to sample and library quality control criteria (DNA amount 150ng DNA fragment size at least 250bp, minimum library concentration of 1nM, and minimum of 400,000 reads per sample) and by only calling mutations if present with allele frequency above 5% for FFPET samples and 1% for AML samples. Notabily, non - reproducible “mutations” in clinically relevant genes (eg KRAS G12A) were not infrequently encountered below these cut off values. A second independent test set of 82 FFPET samples with known missense and deletion mutations in EGFR, KRAS, BRAF, NRAS, PIK3CA, KIT and PDGFRA were analysed by the TruSeq panel. By strict adherence to the above criteria and restricting variant calling to clinically relevant mutations, 100% sensitivity and 100% specificity was achieved in the samples that met the criteria. In all, only 71% of the samples tested passed all quality control criteria. 12% of the samples failed the library preparation and were not processed. 17% of the samples passed the library QC criteria but failed the sample QC criteria. In each case the known mutation was identified. In conclusion, by strict adherence to sample and library QC and by restricting analysis to clinically-relevant mutations, the TruSeq amplicon cancer panel was able to detect common somatic missense and deletion mutations with an allele frequency >5% in FFPET samples with 100% specificity and sensitivity without the need for confirmation by an orthogonal method. However, confirmation by an orthogonal methods would be required for suspected mutations present at an allele frequency <5%, for mutations not known to be of clinical – relevance and for samples with low tumour purity, low DNA input or poor quality DNA. This study showed that deep sequencing of tumour tissue from FFPETs generated many low frequency artifacts due to sequencing, polymerase, formalin – induced chemical modifications and well as frequent mapping and variant calling errors. These artifacts and errors mostly occur at low allele frequency and can be difficult to distinguish from low frequency somatic mutations. Strict adherence to sample and library quality control criteria, allele frequency thresholds and clinically relevant mutations allows highly accurate mutation calling without the need for confirmation by an orthogonal method.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MS11.3 - The Role of Current Pathologic Techniques in the Next Gen World (ID 508)
14:45 - 15:05 | Author(s): I.I. Wistuba
- Abstract
- Presentation
Abstract
Over the past decade, significant progress has been made in the characterization of molecular and genetic abnormalities tumors from patients with non-small cell carcinoma (NSCLC) that are being used as molecular targets and predictive biomarkers to select patients for targeted therapy. Recent advances in expanding the available NSCLC targeted therapies require the analysis of a broad panel of molecular abnormalities in tumor specimens, including gene mutations, gene amplifications, gene fusions and protein expression by applying different methodologies to tumor tissue (biopsy) and cell (cytology) samples. The rapid development of technologies for large-scale sequencing (next-generation sequencing, NGS) has facilitated high-throughput molecular analysis holding various advantages over traditionally sequencing including the ability to fully sequence large numbers of genes in a single test and simultaneously detect deletions, insertions, copy number alterations, translocations, and exome-wide base substitutions (including known hot-spot mutations) in all known cancer-related genes [1,2]. Currently, NGS platforms, including whole genome, whole exome and targeted gene sequencing, represent emerging diagnostic methodologies for the detection of oncogenes fusions and mutations in tumor tissue specimens, including formalin-fixed and paraffin-embedded (FFPE) samples [3]. Technical challenges include sequencing samples of low quality and/or quality, reliable identification of structural and copy number variation, and assessment of intratumoral heterogeneity. In addition, the clinical use of the NGS sequencing data is not straightforward and there are several challenges related to data analysis, data storage and report generation [4]. There is growing consensus that tumor tissue specimens must represent the setting of the disease to be treated, and increasingly, more tissue samples are being obtained for molecular testing of advanced, metastatic and chemo-refractory NSCLC tumors (e.g., MD Anderson BATTLE Lung Cancer Program) [5]. However, the biopsy and cytology samples available for molecular testing in those metastatic refractory NSCLC tumors are likely to be more challenging samples for molecular testing, including NGS platforms. The role of the pathologist is becoming increasingly important to adequately integrate routine histopathology assessments and molecular testing, including NGS, with clinical pathology for the most accurate tumor diagnosis and subsequent selection of the most appropriate therapy. References: 1. Meyerson M, Gabriel S, Getz G: Advances in understanding cancer genomes through second-generation sequencing. Nat Rev Genet 11:685-96, 2010 2. Mwenifumbo JC, Marra MA: Cancer genome-sequencing study design. Nat Rev Genet 14:321-32, 2013 3. Ross JS, Cronin M: Whole cancer genome sequencing by next-generation methods. Am J Clin Pathol 136:527-39, 2011 4. Ulahannan D, Kovac MB, Mulholland PJ, et al: Technical and implementation issues in using next-generation sequencing of cancers in clinical practice. Br J Cancer 109:827-35, 2013 5. Kim ES, Herbst RS, Wistuba, II, et al: The BATTLE trial: personalizing therapy for lung cancer. Cancer Discov 1:44-53, 2011Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MS11.4 - Reporting and Interpreting Molecular Results (ID 509)
15:05 - 15:25 | Author(s): M.S. Tsao
- Abstract
- Presentation
Abstract
As molecular biomarkers are becoming routine in the clinical management of lung cancer patients, there is an increasing need to establish standards or guidelines for the reporting of molecular results. In the most ideal situation, reporting of tissue based molecular biomarker results should be integrated into the histopathology report of the tissue sample, to provide a more complete genotype-phenotype characterization of the tumor. This is particularly important for lung cancer as molecular profiling to date has clearly shown that many “driver” genomic aberrations are often closely associated with specific tumor histology. In fact, the current CAP/IASLC/AMP guideline on molecular testing in lung cancer recommends the use of histology (adenocarcinoma containing tumors) as a primary criterion to select lung cancer samples for EGFR and ALK testing. However, until reflex molecular testing becomes routine in pathology practice, molecular testing is often conducted at a laboratory that is separate from the one where the original tissue histopathological diagnosis was made. In such cases, it is important that the stand alone molecular report should also include some histopathological data that may be highly relevant to the interpretation of the results, or at the very least, refer to the relevant Pathology report. In the Pathology report, the data should include: (a) type of sample, whether it is paraffin embedded or fresh (e.g. fluid), (b) tumor diagnosis, subtypes and variants when applicable, (c) essential immunohistochemical markers that were assessed to support the diagnosis, (d) use of tissue processing solution or fixative that could adversely affect the quality of DNA for sequencing, e.g. acid and Bouin’s solution, (e) the approximate size of the tissue, (f) whether a tumor cell enrichment strategy was used, and (g) estimated tumor cellularity in the tissue area marked for isolation of DNA for testing. It is of utmost important that molecular reports are written in language that can be understood by the treating physicians and the pathologists, who are the end-users of the report. Typical laboratory reports should include patient identification codes, the date the sample was acquired (biopsy or resection) from the patient, the date the sample is received in the molecular testing laboratory, and the date the report is signed out. All this information provides not only important sample identification information, but also the real turnaround time of the reported results. Aside from a summary of the molecular results themselves, the report should include a concise but reasonable detailed methodological section, which also provides the performance features of the assay platform being used. It should specify the list of genes included in the assay, the type of aberrations that can be reliably detected, e.g. single nucleotide mutations, deletions, insertions, rearrangements, copy number changes, etc, and the sensitivity and specificity of the assay. The methodology section should also include the analytical software used for processing the data and identifying the genomic aberrations and the version of the normal reference sequence used for comparison with the sequence in question. If the methodology used is fairly new or represent emerging technology, such as next generation sequencing (NGS), then information about mutation verification technology or process may also be required (1). While molecular aberrations are integral to the complete pathological diagnosis of a tumor, in lung cancer their main clinical relevance is for their ability to predict patient response to a specific therapeutic agent, or for patient prognosis. In this context, especially if there are a number of genetic changes being reported (as example with NGS); it may be useful if the aberrations (often called variants) are classified into categories, which reflect their clinical utility. Although there is as yet no universally acceptable classification framework for reporting genomic aberrations identified by NGS platforms, broad categories that establish prognostic, biological or treatment relevance to the aberrations have been proposed or used. These variants have been classified into several “Levels” or “Tiers”, depending on the level of evidence for their predictiveness of response to specific drug. These levels have been derived from widely accepted classification schemes, such as those published by the American College of Medial Genetics (ACMG) for use in diseases such as Breast Cancer. The “actionable” aberrations are those demonstrating proven evidence for their association with high response rates to a specific drug or treatment strategy. The “potentially actionable” alterations are those with strong rationale but as yet proven clinical evidence for being associated high response rate to a specific drug. This group also include aberrations that have demonstrated evidence for response to a specific drug in one type of cancer, yet of unproven response pattern in a different tumor being studied. However, as NGS enables the discovery of a large number of genetic aberrations that typically occur in sporadic adult cancers, many aberrations fall into the category of “unknown therapeutic or biological significance”. While some of these could potentially be predictive markers of drugs that are already available for other reasons, most may not even be pharmacologically targetable. An important risk of conducting comprehensive genomic profiling in patient samples is the identification of “incidental” aberrations, which require clinical management that is not originally planned or anticipated (2). These aberrations could involve genes/mutations with known hereditary roles in cancer or non-cancer conditions, with potentially significant implication on patient and/or other family members. For these reasons, the ACMG recently convened a working group of experts to publish recommendations for reporting of incidental findings in clinical exome and genome sequencing. While these recommendations have been provided primarily as educational resources for medical geneticists and other health care providers (and are still quite controversial), the issues discussed should be considered when deciding upon the reporting strategy for profiling cancer samples using NGS technology. References: 1. Rehm HL, Bale SJ, et al. ACMG clinical laboratory standards for next-generation sequencing. Genet Med. 2013 Jul 25. doi: 10.1038/gim.2013.92. [Epub ahead of print] 2. Green RC, Berg JS, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med. 2013 Jul;15(7):565-74. doi: 10.1038/gim.2013.73. Epub 2013 Jun 20.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
