Editorial
Next-generation sequencing in precision oncology: challenges and opportunities Expert Review of Molecular Diagnostics Downloaded from informahealthcare.com by Georgian Court University on 02/23/15 For personal use only.
Expert Rev. Mol. Diagn. 14(6), 635–637 (2014)
Kristina M Kruglyak Illumina, Inc., 5200 Illumina Way, San Diego, CA 92122, USA
Erick Lin Illumina, Inc., 5200 Illumina Way, San Diego, CA 92122, USA
Frank S Ong Author for correspondence: Illumina, Inc., 5200 Illumina Way, San Diego, CA 92122, USA [email protected]
High throughput gene sequencing is transforming the utilization of genomics in patient care by providing physicians with a powerful tool to aid the diagnosis and management of disease, particularly in precision oncology. As next-generation sequencing (NGS)-based diagnostic assays are developed, significant hurdles such as assessing tumor heterogeneity, characterizing ‘driver’ and ‘passenger’ mutations, typing molecular signatures of individual cancers and determining limits of detection pose significant challenges for clinical laboratories and downstream bioinformatics analyses. Despite these challenges, NGS has the potential to affect all facets of cancer treatment, including early detection and diagnosis through cancer screening in at-risk populations and assessing therapeutic efficacy by detection of circulating tumor DNA via noninvasive blood draws. As the utilization of NGS in precision oncology matures, NGS-based laboratory tests could be used throughout the evolution of cancer in patients and allow for cancers to be monitored and managed as a chronic disease, rather than an acute condition.
While it has long been understood that cancers are characterized by the uncontrolled growth of abnormal cells, more recent insights that these abnormal cells are unique at the genetic level [1] has ushered in the era of precision oncology. Over time, somatic mutations accumulate in the cells of a patient, due to replication errors or DNA damage that remain unchecked by traditional cellular mechanisms. When one of these somatic mutations confers a growth advantage to a particular population of cells, either by promoting cellular division or by inhibiting apoptosis, this clonal population proliferates and manifests as disease. As the number of somatic mutations increases, the number of ‘driver’ mutations that confer a selective advantage to the population also increases. With the advent of next-generation sequencing (NGS), the promise of targeted, personalized treatment draws closer. The throughput and cost of NGS has now reached a point where a whole human genome can be sequenced in 1 day for less than US$1000 [2]. Targeted assays such as whole exome sequencing or
multigene panels have become commonplace in the clinical research setting, and the performance of these assays combined with their cost are gradually pushing traditional Sanger sequencing to the periphery [3–5]. Despite the huge potential of NGS in this area, significant hurdles must be overcome to make these assays commonplace and truly effective. Although cancer is known to be a genetic disease, the complex heterogeneity among the abnormal cells in a single tumor is remarkable, making detection of low-level mutations difficult. Additionally, although a tumor may have many somatic mutations, not all of them are driver mutations that should be considered in guiding treatment decisions; the majority in fact are ‘passenger’ mutations that do not contribute to tumor pathogenicity [6]. Differentiation of these two classes is critical as treatment is more likely to be effective when applied at an early stage, and each round of treatment also has a significant cost and impact on the patient’s quality of life. However, as targeted therapeutics are administered, selective pressures and
KEYWORDS: bioinformatics • ctDNA • next-generation sequencing • NGS • personalized medicine • precision oncology • tumor heterogeneity
informahealthcare.com
10.1586/14737159.2014.916213
Ó 2014 Informa UK Ltd
ISSN 1473-7159
635
Expert Review of Molecular Diagnostics Downloaded from informahealthcare.com by Georgian Court University on 02/23/15 For personal use only.
Editorial
Kruglyak, Lin & Ong
clonal evolution may provide passenger mutations opportunities to play additional roles in tumor biology and confer additional survival advantages to the population against selective pressures of therapy, adding to the complexity of tumor heterogeneity. Technical difficulties persist as well. Because of tumor heterogeneity, it is necessary to sequence both primary and secondary tumors to understand the range of somatic mutations present. However, in the case of metasynchronous or synchronous metastases where limited tumor samples are available for genomic profiling, results from primary tumors still provide relevant genomic information to guide treatment or recurrence [7]. Even in the case of a single primary tumor, it is preferable to sequence from multiple areas [8]. To add an additional layer of complexity, subclones may comprise minor, equal or major components of both primary and secondary tumors and may be affected by the tumor microenvironments present. In many cancers, and particularly in the case of lung cancer, this requirement poses difficulties as biopsy is a painful procedure and may be difficult to perform based on tumor location. Additionally, requiring continuous biopsies to monitor the effect of treatment is not feasible in the long term. Similar to previous studies on circulating tumor cells (CTCs) where CTCs represent an attractive target for assessing prognosis, monitoring response to therapy, pharmacodynamic studies and rational selection of therapies in cancer patients [9], one promising NGS-based alternative is to sequence circulating tumor DNA (ctDNA) [10] via a noninvasive blood draw. For CTCs and ctDNA, ongoing research to assess the clinical validity and clinical utility of these surrogate markers of tumors show promise. Therefore, methods for noninvasive testing of cancer, such ctDNA testing, have the potential to transform cancer detection and monitoring into a standard laboratory assay. Another issue relating to the clinical presentation of cancer is the fact that there are many disease subtypes that are not currently differentiable to the oncologist – each presentation of cancer in an individual is effectively unique. As more research is performed for assessing the molecular profiles of cancers, the results should provide additional granularity to identify the distinctiveness of individual cancers. For example, in colorectal cancer, it has recently been shown using NGS technologies that two subtypes exist: hypermutated and nonhypermutated. In the former case, the genetic signature is marked by microsatellite instability and in the latter case, the genetic signature was differentiated by mutation of the TP53 and APC genes. In nonhypermutated colorectal cancer, 60% of patients have a mutation in TP53, while only 20% of patients with the nonhypermutated form have such mutations [11]. One can envision future work further classifying both groups into finer subtypes to guide treatment decisions. Ultimately, however, as the era of personalized genomics and personalized medicine evolves, the paradigm of tumor subtype stratification should naturally shift toward approaching cancer as an individual disease with unique molecular signatures rather than as an organ-based disease treated based on population-level patterns and outcomes statistics. 636
Bioinformatics and analytical obstacles are not new in reviewing NGS data but can be particularly challenging in the context of tumor heterogeneity. Tumor heterogeneity manifests itself as low-level somatic variants that are seen only in a small number of sequencing reads. For rare variants, identification may be difficult or impossible because they may not be differentiable from errors due to instrument noise. For example, if the genome is sequenced to a depth of 100, then a variant present at a level of 1% would be expected to appear in only one read. Currently, sequencing accuracy is very high, generally reported at 99.9% or above for the majority of bases [12]. However, this implies that one base per 1000 sequenced will have an error. At a depth of 100, this means that 300 million out of 300 billion bases sequenced across the genome would report an alternate base, given a 0.1% sequencing error rate. In the context of the genome, the 300 million alternate base calls would translate into 1 in 10 positions reporting an alternate base. Clearly, a cancer signature does not encompass 300 million sites, so a 1% somatic frequency is effectively impossible to differentiate from noise. By contrast, a 5% somatic frequency is much more unique, and setting such a threshold would reduce the number of false positive calls substantially. Other variant types such as copy number variants, structural variants and epigenomic variants pose additional difficulties. Regardless of the variant type, once a set of variants is identified from a cancer sample, a further challenge is to classify each as driver versus passenger. Current methods for such classification are generally based on calculation of risk scores, such as SIFT or PolyPhen [13,14]; determination of the protein coding effect of each mutation, either sense, missense or nonsense; or identification within a relevant database, such as dbSNP, COSMIC, OMIM or HapMap. All of these methods are largely qualitative in nature, and the result of such methods is not a final classification of each variant, but rather a ranking of the identified variants according to likelihood of pathogenicity. Some pipelines endeavor to identify drivers versus passengers at the gene level rather than at the variant level, with the rationale that many variants within a single gene are more likely to affect its function and are thus related to cancer progression [15–17]. Many studies have been published across cancer types that identify risk genes rather than specific variants. More recently however, some groups have raised concerns on the use of this method because, as the sample size increases, the list of putatively significant genes balloons and is dominated by false positives, which then need to be extensively filtered to enable identification of the true positives [15]. Future of NGS in precision oncology
Although challenges exist, NGS in precision oncology has already achieved much more than was ever imagined possible even a few years ago. NGS also has the same potential to transform all facets of cancer treatment, from early detection and treatment decisions, to various monitoring aspects, including minimal residual disease. As more biomarkers are associated with specific forms of cancer, it will be easier to design targeted assays that effectively act as ‘cancer screens’ that can be applied Expert Rev. Mol. Diagn. 14(6), (2014)
Expert Review of Molecular Diagnostics Downloaded from informahealthcare.com by Georgian Court University on 02/23/15 For personal use only.
NGS in precision oncology
quickly and cheaply to an entire at-risk population, comparable to today’s blood panel tests. Notably, as the cost of whole genome sequencing decreases, it will be most effective to sequence the entire genome, and by utilizing bioinformatics, report only those variants or genes requested by the treating physician. Such a noninvasive, broad screening option will increase the rate of early detection across all cancer types, ideally allowing treatments to begin prior to metastasis. These treatments will also be guided by the results from whole genome sequencing. Therapeutics tailored to an individual’s mutational spectrum will have a much greater success rate than a general therapy that may not target the specific driver mutations. After a treatment option is chosen, NGS methods can be used to monitor the efficacy of that treatment over time. For example, if a patient’s lung cancer is shown to be dominated by a specific mutation in EGFR, then regular screenings for this variant can show the success or failure of the chosen treatment in near to real-time. Similarly, in the case of surgical interventions, a ctDNA-based NGS-assay following surgery will be used to identify whether tumor resection was successful by monitoring decreasing levels of ctDNA over time. Given that the half-life of circulating DNA in the blood is less References 1.
Stratton MR, Campbell PJ, Futreal PA. The cancer genome. Nature 2009;458:719-24
2.
Winnick E. Illumina launches two new platforms at JP Morgan Conference; claims $1,000 genome. GenomeWeb 2014. Available from: www.genomeweb.com/ sequencing/illumina-launches-two-newplatforms-jp-morgan-conference-claims1000-genome [Last accessed 27 January 2014]
3.
Rehm HL. Disease-targeted sequencing: a cornerstone in the clinic. Nat Rev Genet 2013;14:295-300
4.
Park JY, Kricka LJ, Fortina P. Next-generation sequencing in the clinic. Nat Biotechnol 2013;31:990-2
5.
Strom SP, Lee H, Das K, et al. Assessing the necessity of confirmatory testing for exome-sequencing results in a clinical molecular diagnostics laboratory. Genet Med 2014. [Epub ahead to print]
6.
Vogelstein B, Papadopoulos N, Velculescu VE, et al. Cancer genome landscapes. Science 2013;339: 1546-58
informahealthcare.com
than half an hour [18], ctDNA could theoretically be applied to assess successful tumor resection in the short term and also to monitor for minimal residual disease or cancer recurrence in the long term. Even with ctDNA methods currently in their infancy, NGS assays based on ctDNA have already been shown to be more sensitive compared to traditional imaging technologies used to monitor recurrence [10]. Eventually, the NGS-based option may become the norm across the entire spectrum in the evolution of cancer. The hope is that cancer will eventually be monitored and managed as a chronic disease rather than as a serious, acute condition it is currently. Financial & competing interests disclosure
The authors are employees of Illumina, Inc. and own stocks in the company. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript.
7.
Vignot S, Frampton GM, Soria JC, et al. Next-generation sequencing. reveals high concordance of recurrent somatic alterations between primary tumor and metastases from patients with non-small-cell lung cancer. J Clin Oncol 2013;31:2167-72
8.
Kosmidou V, Oikonomou E, Vlassi M, et al. Tumor heterogeneity revealed by KRAS, BRAF, and PIK3CA pyrosequencing: KRAS and PIK3CA intratumour mutation profile differences and their therapeutic implications. Hum Mutat 2014;35(3): 329-40
9.
10.
11.
Editorial
Urtishak S, Alpaugh RK, Weiner LM, Swaby RF. Clinical utility of circulating tumor cells: a role for monitoring response to therapy and drug development. Biomark Med 2008;2(2):137-45 Dawson SJ, Rosenfeld N, Caldas C, et al. Analysis of circulating tumor DNA to monitor metastatic breast cancer. N Engl J Med 2013;368:1199-209 The Cancer Genome Atlas Network. Comprehensive molecular characterization of human colon and rectal cancer. Nature 2013;487:330-7
12.
Harismendy O, Ng PC, Strausberg RL, et al. Evaluation of next generation sequencing platforms for population targeted sequencing studies. Genome Biol 2009;10:R32
13.
Ng PC, Henikoff S, SIFT: predicting amino acid changes that affect protein function. Nucleic Acids Res 2003;31:3812-14
14.
Adzhubei IA, Schmidt S, Peshkin L, et al. A method and server for predicting damaging missense mutations. Nat Methods 2010;7:248-9
15.
Lawrence MS, Stojanov P, Polak P, et al. Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature 2013;499:214-18
16.
Gonzalez-Perez A, Mustonen V, Reva B, et al. Computational approaches to identify functional genetic variants in cancer genomes. Nat Methods 2013;10:723-9
17.
Hodis E, Watson IR, Kryukov GV, et al. A landscape of driver mutations in melanoma. Cell 2012;150:251-63
18.
Lo YM, Zhang J, Leung TN, et al. Rapid clearance of fetal DNA from maternal plasma. Am J Hum Genet 1999;64:218-24
637
Next-generation sequencing in precision oncology: challenges and opportunities Expert Review of Molecular Diagnostics Downloaded from informahealthcare.com by Georgian Court University on 02/23/15 For personal use only.
Expert Rev. Mol. Diagn. 14(6), 635–637 (2014)
Kristina M Kruglyak Illumina, Inc., 5200 Illumina Way, San Diego, CA 92122, USA
Erick Lin Illumina, Inc., 5200 Illumina Way, San Diego, CA 92122, USA
Frank S Ong Author for correspondence: Illumina, Inc., 5200 Illumina Way, San Diego, CA 92122, USA [email protected]
High throughput gene sequencing is transforming the utilization of genomics in patient care by providing physicians with a powerful tool to aid the diagnosis and management of disease, particularly in precision oncology. As next-generation sequencing (NGS)-based diagnostic assays are developed, significant hurdles such as assessing tumor heterogeneity, characterizing ‘driver’ and ‘passenger’ mutations, typing molecular signatures of individual cancers and determining limits of detection pose significant challenges for clinical laboratories and downstream bioinformatics analyses. Despite these challenges, NGS has the potential to affect all facets of cancer treatment, including early detection and diagnosis through cancer screening in at-risk populations and assessing therapeutic efficacy by detection of circulating tumor DNA via noninvasive blood draws. As the utilization of NGS in precision oncology matures, NGS-based laboratory tests could be used throughout the evolution of cancer in patients and allow for cancers to be monitored and managed as a chronic disease, rather than an acute condition.
While it has long been understood that cancers are characterized by the uncontrolled growth of abnormal cells, more recent insights that these abnormal cells are unique at the genetic level [1] has ushered in the era of precision oncology. Over time, somatic mutations accumulate in the cells of a patient, due to replication errors or DNA damage that remain unchecked by traditional cellular mechanisms. When one of these somatic mutations confers a growth advantage to a particular population of cells, either by promoting cellular division or by inhibiting apoptosis, this clonal population proliferates and manifests as disease. As the number of somatic mutations increases, the number of ‘driver’ mutations that confer a selective advantage to the population also increases. With the advent of next-generation sequencing (NGS), the promise of targeted, personalized treatment draws closer. The throughput and cost of NGS has now reached a point where a whole human genome can be sequenced in 1 day for less than US$1000 [2]. Targeted assays such as whole exome sequencing or
multigene panels have become commonplace in the clinical research setting, and the performance of these assays combined with their cost are gradually pushing traditional Sanger sequencing to the periphery [3–5]. Despite the huge potential of NGS in this area, significant hurdles must be overcome to make these assays commonplace and truly effective. Although cancer is known to be a genetic disease, the complex heterogeneity among the abnormal cells in a single tumor is remarkable, making detection of low-level mutations difficult. Additionally, although a tumor may have many somatic mutations, not all of them are driver mutations that should be considered in guiding treatment decisions; the majority in fact are ‘passenger’ mutations that do not contribute to tumor pathogenicity [6]. Differentiation of these two classes is critical as treatment is more likely to be effective when applied at an early stage, and each round of treatment also has a significant cost and impact on the patient’s quality of life. However, as targeted therapeutics are administered, selective pressures and
KEYWORDS: bioinformatics • ctDNA • next-generation sequencing • NGS • personalized medicine • precision oncology • tumor heterogeneity
informahealthcare.com
10.1586/14737159.2014.916213
Ó 2014 Informa UK Ltd
ISSN 1473-7159
635
Expert Review of Molecular Diagnostics Downloaded from informahealthcare.com by Georgian Court University on 02/23/15 For personal use only.
Editorial
Kruglyak, Lin & Ong
clonal evolution may provide passenger mutations opportunities to play additional roles in tumor biology and confer additional survival advantages to the population against selective pressures of therapy, adding to the complexity of tumor heterogeneity. Technical difficulties persist as well. Because of tumor heterogeneity, it is necessary to sequence both primary and secondary tumors to understand the range of somatic mutations present. However, in the case of metasynchronous or synchronous metastases where limited tumor samples are available for genomic profiling, results from primary tumors still provide relevant genomic information to guide treatment or recurrence [7]. Even in the case of a single primary tumor, it is preferable to sequence from multiple areas [8]. To add an additional layer of complexity, subclones may comprise minor, equal or major components of both primary and secondary tumors and may be affected by the tumor microenvironments present. In many cancers, and particularly in the case of lung cancer, this requirement poses difficulties as biopsy is a painful procedure and may be difficult to perform based on tumor location. Additionally, requiring continuous biopsies to monitor the effect of treatment is not feasible in the long term. Similar to previous studies on circulating tumor cells (CTCs) where CTCs represent an attractive target for assessing prognosis, monitoring response to therapy, pharmacodynamic studies and rational selection of therapies in cancer patients [9], one promising NGS-based alternative is to sequence circulating tumor DNA (ctDNA) [10] via a noninvasive blood draw. For CTCs and ctDNA, ongoing research to assess the clinical validity and clinical utility of these surrogate markers of tumors show promise. Therefore, methods for noninvasive testing of cancer, such ctDNA testing, have the potential to transform cancer detection and monitoring into a standard laboratory assay. Another issue relating to the clinical presentation of cancer is the fact that there are many disease subtypes that are not currently differentiable to the oncologist – each presentation of cancer in an individual is effectively unique. As more research is performed for assessing the molecular profiles of cancers, the results should provide additional granularity to identify the distinctiveness of individual cancers. For example, in colorectal cancer, it has recently been shown using NGS technologies that two subtypes exist: hypermutated and nonhypermutated. In the former case, the genetic signature is marked by microsatellite instability and in the latter case, the genetic signature was differentiated by mutation of the TP53 and APC genes. In nonhypermutated colorectal cancer, 60% of patients have a mutation in TP53, while only 20% of patients with the nonhypermutated form have such mutations [11]. One can envision future work further classifying both groups into finer subtypes to guide treatment decisions. Ultimately, however, as the era of personalized genomics and personalized medicine evolves, the paradigm of tumor subtype stratification should naturally shift toward approaching cancer as an individual disease with unique molecular signatures rather than as an organ-based disease treated based on population-level patterns and outcomes statistics. 636
Bioinformatics and analytical obstacles are not new in reviewing NGS data but can be particularly challenging in the context of tumor heterogeneity. Tumor heterogeneity manifests itself as low-level somatic variants that are seen only in a small number of sequencing reads. For rare variants, identification may be difficult or impossible because they may not be differentiable from errors due to instrument noise. For example, if the genome is sequenced to a depth of 100, then a variant present at a level of 1% would be expected to appear in only one read. Currently, sequencing accuracy is very high, generally reported at 99.9% or above for the majority of bases [12]. However, this implies that one base per 1000 sequenced will have an error. At a depth of 100, this means that 300 million out of 300 billion bases sequenced across the genome would report an alternate base, given a 0.1% sequencing error rate. In the context of the genome, the 300 million alternate base calls would translate into 1 in 10 positions reporting an alternate base. Clearly, a cancer signature does not encompass 300 million sites, so a 1% somatic frequency is effectively impossible to differentiate from noise. By contrast, a 5% somatic frequency is much more unique, and setting such a threshold would reduce the number of false positive calls substantially. Other variant types such as copy number variants, structural variants and epigenomic variants pose additional difficulties. Regardless of the variant type, once a set of variants is identified from a cancer sample, a further challenge is to classify each as driver versus passenger. Current methods for such classification are generally based on calculation of risk scores, such as SIFT or PolyPhen [13,14]; determination of the protein coding effect of each mutation, either sense, missense or nonsense; or identification within a relevant database, such as dbSNP, COSMIC, OMIM or HapMap. All of these methods are largely qualitative in nature, and the result of such methods is not a final classification of each variant, but rather a ranking of the identified variants according to likelihood of pathogenicity. Some pipelines endeavor to identify drivers versus passengers at the gene level rather than at the variant level, with the rationale that many variants within a single gene are more likely to affect its function and are thus related to cancer progression [15–17]. Many studies have been published across cancer types that identify risk genes rather than specific variants. More recently however, some groups have raised concerns on the use of this method because, as the sample size increases, the list of putatively significant genes balloons and is dominated by false positives, which then need to be extensively filtered to enable identification of the true positives [15]. Future of NGS in precision oncology
Although challenges exist, NGS in precision oncology has already achieved much more than was ever imagined possible even a few years ago. NGS also has the same potential to transform all facets of cancer treatment, from early detection and treatment decisions, to various monitoring aspects, including minimal residual disease. As more biomarkers are associated with specific forms of cancer, it will be easier to design targeted assays that effectively act as ‘cancer screens’ that can be applied Expert Rev. Mol. Diagn. 14(6), (2014)
Expert Review of Molecular Diagnostics Downloaded from informahealthcare.com by Georgian Court University on 02/23/15 For personal use only.
NGS in precision oncology
quickly and cheaply to an entire at-risk population, comparable to today’s blood panel tests. Notably, as the cost of whole genome sequencing decreases, it will be most effective to sequence the entire genome, and by utilizing bioinformatics, report only those variants or genes requested by the treating physician. Such a noninvasive, broad screening option will increase the rate of early detection across all cancer types, ideally allowing treatments to begin prior to metastasis. These treatments will also be guided by the results from whole genome sequencing. Therapeutics tailored to an individual’s mutational spectrum will have a much greater success rate than a general therapy that may not target the specific driver mutations. After a treatment option is chosen, NGS methods can be used to monitor the efficacy of that treatment over time. For example, if a patient’s lung cancer is shown to be dominated by a specific mutation in EGFR, then regular screenings for this variant can show the success or failure of the chosen treatment in near to real-time. Similarly, in the case of surgical interventions, a ctDNA-based NGS-assay following surgery will be used to identify whether tumor resection was successful by monitoring decreasing levels of ctDNA over time. Given that the half-life of circulating DNA in the blood is less References 1.
Stratton MR, Campbell PJ, Futreal PA. The cancer genome. Nature 2009;458:719-24
2.
Winnick E. Illumina launches two new platforms at JP Morgan Conference; claims $1,000 genome. GenomeWeb 2014. Available from: www.genomeweb.com/ sequencing/illumina-launches-two-newplatforms-jp-morgan-conference-claims1000-genome [Last accessed 27 January 2014]
3.
Rehm HL. Disease-targeted sequencing: a cornerstone in the clinic. Nat Rev Genet 2013;14:295-300
4.
Park JY, Kricka LJ, Fortina P. Next-generation sequencing in the clinic. Nat Biotechnol 2013;31:990-2
5.
Strom SP, Lee H, Das K, et al. Assessing the necessity of confirmatory testing for exome-sequencing results in a clinical molecular diagnostics laboratory. Genet Med 2014. [Epub ahead to print]
6.
Vogelstein B, Papadopoulos N, Velculescu VE, et al. Cancer genome landscapes. Science 2013;339: 1546-58
informahealthcare.com
than half an hour [18], ctDNA could theoretically be applied to assess successful tumor resection in the short term and also to monitor for minimal residual disease or cancer recurrence in the long term. Even with ctDNA methods currently in their infancy, NGS assays based on ctDNA have already been shown to be more sensitive compared to traditional imaging technologies used to monitor recurrence [10]. Eventually, the NGS-based option may become the norm across the entire spectrum in the evolution of cancer. The hope is that cancer will eventually be monitored and managed as a chronic disease rather than as a serious, acute condition it is currently. Financial & competing interests disclosure
The authors are employees of Illumina, Inc. and own stocks in the company. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript.
7.
Vignot S, Frampton GM, Soria JC, et al. Next-generation sequencing. reveals high concordance of recurrent somatic alterations between primary tumor and metastases from patients with non-small-cell lung cancer. J Clin Oncol 2013;31:2167-72
8.
Kosmidou V, Oikonomou E, Vlassi M, et al. Tumor heterogeneity revealed by KRAS, BRAF, and PIK3CA pyrosequencing: KRAS and PIK3CA intratumour mutation profile differences and their therapeutic implications. Hum Mutat 2014;35(3): 329-40
9.
10.
11.
Editorial
Urtishak S, Alpaugh RK, Weiner LM, Swaby RF. Clinical utility of circulating tumor cells: a role for monitoring response to therapy and drug development. Biomark Med 2008;2(2):137-45 Dawson SJ, Rosenfeld N, Caldas C, et al. Analysis of circulating tumor DNA to monitor metastatic breast cancer. N Engl J Med 2013;368:1199-209 The Cancer Genome Atlas Network. Comprehensive molecular characterization of human colon and rectal cancer. Nature 2013;487:330-7
12.
Harismendy O, Ng PC, Strausberg RL, et al. Evaluation of next generation sequencing platforms for population targeted sequencing studies. Genome Biol 2009;10:R32
13.
Ng PC, Henikoff S, SIFT: predicting amino acid changes that affect protein function. Nucleic Acids Res 2003;31:3812-14
14.
Adzhubei IA, Schmidt S, Peshkin L, et al. A method and server for predicting damaging missense mutations. Nat Methods 2010;7:248-9
15.
Lawrence MS, Stojanov P, Polak P, et al. Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature 2013;499:214-18
16.
Gonzalez-Perez A, Mustonen V, Reva B, et al. Computational approaches to identify functional genetic variants in cancer genomes. Nat Methods 2013;10:723-9
17.
Hodis E, Watson IR, Kryukov GV, et al. A landscape of driver mutations in melanoma. Cell 2012;150:251-63
18.
Lo YM, Zhang J, Leung TN, et al. Rapid clearance of fetal DNA from maternal plasma. Am J Hum Genet 1999;64:218-24
637
