1

Course Syllabus

Clinical Application of Next Generation Sequencing for the Management of Patients with Solid Tumors

Faculty:

Director: Jeffrey S. Ross, MD Albany Medical Center, Albany, NY Foundation Medicine, Inc., Cambridge, MA Co-Directors: Neal Lindeman, MD Brigham and Women’s Hospital Boston, MA Stephen Yip, MD PhD University of British Columbia Vancouver, BC

2

Introduction

Cancer develops due to an accumulation of mutations in DNA. For many years, it has been widely

accepted that the development and progression of cancer is associated with alterations in the DNA

sequence of the cancer cell genome, such as base substitutions, short insertions and deletions,

homozygous deletions, amplifications and fusions (translocations) of genetic material [1]. Improved

understanding of the genetic mechanisms that initiate or drive cancer progression has set the stage for

the development of personalized cancer treatment [2–5]. Although the total catalogue of critical

“driver” alterations known to promote oncogenesis of a given cancer type may be large, the number of

driver alterations contributing to an individual patient’s solid tumor is typically low and unpredictable

[2,6,7]. Direct sequencing of the tumor cell DNA is necessary to identify which alterations drive an

individual patient’s disease. The identification and targeting of specific mutations that have arisen in a

tumor continues to show great promise as a means to increase the efficacy of cancer therapy.

History of DNA sequencing and the human genome

Severeal decades passed after the structure of DNA was discovered before the sequence of

human DNA began to be elucidated. It was not until the 1970s that Frederick Sanger developed the

“Sanger method” of rapid DNA sequencing in 1977 [8]. In 1986, Leroy Hood introduced the first semi-

automated DNA sequencing machine [9], and in 1987 the first fully automated sequencing machine, the

ABI 370, was introduced. Shortly thereafter, the sequencing of human cDNA ends, known as “expressed

sequence tags”, began in the laboratory of Craig Venter. These technical advances culminated with the

first publications of the human genome sequence in 2001 [10–12].

At the time of its publication in 2001, the initial near-complete draft of the human genome had

required more than 12 years of sequencing at multiple laboratories at a cost of more than $3 billion

USD. Since then, a continuous demand for more rapid and low-cost sequencing has driven the

3

development of novel approaches designed to parallelize the sequencing process. These new

“massively parallel” or “next generation” strategies, in comparison with traditional Sanger and other

methods, have increased sequencing rates by orders of magnitude and driven down the cost per base

significantly [13–15].

Table 1. Clinically available DNA sequencing techniques

Methodology Variant Types Detected Clinical Considerations

SNVs Indels CNVs Rearr.

High

-thr

ough

put N

GS T

echn

ique

s

Whole genome sequencing (WGS)

WGS can be performed using any of the NGS technologies described above.

Y Y Y Y Captures all types of genomic variants, including unanticipated structural variants. Fresh or frozen samples needed FFPE samples not appropriate Complex analysis, matched normal sample important Relatively long turn-around time

Chain termination

DNA synthesis terminated by random incorporation of fluorescently labeled bases, as in Sanger dideoxy sequencing. Optical detection systems capture nucleotide incorporation after each cycle. Example: Illumina

Y Y Y Yǂ Highest sensitivity for mutation detection Can simultaneously detect all classes of genomic alterations High cost of instruments Generates large amounts of data (optical images) High bioinformatics requirements for analyzing subsequent data

4

HiSeq

Ion semiconductor detection

Relies on the detection of hydrogen ions released during DNA synthesis. Is unique among NGS technologies because it does not rely on optical measurements. Example: Ion Torrent

Y Y Y Y* Low cost Less bioinformatics expertise required Often used on limited regions of exomes as a hotspot test Reduced sensitivity for mutation detection High error rate in homopolymer sequences

Low

thro

ughp

ut P

CR-b

ased

Tec

hniq

ues

Sanger dideoxy sequencing

DNA synthesis terminated by random incorporation of fluorescently labeled bases. DNA fragments separated by capillary electrophoresis to determine sequence.

Y Y N N Provides complete sequence of interest Captures unknown mutations within targeted region Historical gold standard Very time consuming Cannot detect large deletions, translocations, or copy number changes

Allele-specific PCR

Primers span codon of interest and probes detect specific mutation.

Y N N N Very high sensitivity Widely used for clinical testing of oncogenic mutations in CRC, NSCLC Limited to hotspots Cannot detect large indels, translocations, or copy number changes

Mass spectrometry

Single nucleotide primer extension assays followed by

Y Y* N N Readily identifies somatic point mutations and germline base substitutions Limited to hotspots

5

analysis of DNA product using a mass spectrometer.

High initial investment costs for instruments Requires substantial operator expertise

Real-time melting curve PCR

Melting curve of DNA measured to identify mutated PCR products, which melt at lower temperatures than wild-type DNA.

Y Y* N N Very high sensitivity Provides percentage of mutated vs wild-type DNA Does not provide absolute percentage of mutated DNA

Pyrosequencing Measures the release of pyrophosphate during nucleotide incorporation.

Y Y* N N Fast Greater sensitivity than Sanger Provides percentage of mutated DNA Works well with fragmented DNA from FFPE samples Captures unknown short variant mutations within targeted region High reagent costs High error rates Cannot detect large deletions, translocations, or copy number changes

Y - Variants detected. * - Variant detected with difficulty. N – Variants not detected. ǂIt is straightforward to identify rearrangements with specific breakpoints, but variants resulting from unforeseen breakpoints not covered within the bait or primer set would not be detected. Indels = small insertions or deletions; NGS = next-generation sequencing; PCR = polymerase chain reaction; SNVs = single nucleotide variants; CNVs = copy number alterations; Rearr. = Rearrangements, Fusions, Translocations

6

DNA Sequencing Technologies (Table 1)

Traditional DNA sequencing methods that have been used to characterize clinical cancer

specimens and impact treatment decisions are highly sensitive, but are often limited in their scope to

known mutational hot spots. Although targeted and quick, the rate of false negatives, limitations in the

type of alterations that can be identified, and the missed opportunities for identifying other potential

drivers are disadvantages. Next-generation methods have the capability to sequence a much larger set

of alleles simultaneously, providing scale and breadth of analysis that was not previously possible [16–

18].

Sanger Sequencing. The chain termination sequencing method developed by Sanger and colleagues [8]

was the cornerstone procedure used in the original sequencing of the human genome. Contemporary

Sanger sequencing uses automated instruments that detect the insertion of fluorescently labeled

dideoxynucleotide chain terminators and determine their position in the sequenced product following

capillary electrophoresis.

Pyrosequencing. Pyrosequencing differs from the chain termination method by relying on the detection

of pyrophosphate after it is released when the complementary DNA (cDNA) strand is synthesized using

as a template the single DNA strand to be sequenced [19,20]. It is also known as the sequencing by

synthesis method. As each new base is added to the cDNA strand by a chemiluminescent DNA

polymerase, the sequencing system determines the nucleotide of the original template DNA strand

being sequenced. This method is limited in the length of the template DNA strand that can be

sequenced, which is significantly shorter than that for Sanger chain termination sequencing [14].

However, it is considered more sensitive than Sanger sequencing and provides a percentage of the initial

DNA that harbors the specific mutation. Thus, pyrosequencing is most often applied in clinical settings

for short-length “hot-spot” sequencing of specific codons within the gene of interest [21,22].

7

First introduced in 2005, the 454 NGS platform encapsulates a single DNA template strand in an

oil droplet emulsion along with a primer coated bead [23]. The sequencing instrument is organized into

picoliter wells in which an individual oil coated bead is placed. Pyrosequencing is performed and the

nucleotides are visualized in a luciferase system with a fiber-optic coupled imaging camera. The system

provides longer read lengths, which is considered a strength. The 454 NGS platform features the

strengths of relative fast instrument run times and long read length, but is limited by the high cost of

reagents and problems with high error rates in genetic regions rich in homopolymer repeats.

Allele-specific real-time PCR. A variety of closed commercial PCR-based systems have been developed

to perform DNA sequencing and have shown high sensitivity for mutation detection with a reduced risk

of sample contamination [3]. Allele-specific real-time PCR determines the sequence of pre-identified

hot-spots in the cancer cell genome. Primer and probe sets are designed to detect the mutations of

clinical interest, and will not detect other mutations, deletions, or translocations involving related genes.

This method is reported to detect a KRAS mutation present in as little as 1% of the total DNA extracted

from the FFPE specimen [24,25].

Analysis of Melting Curve qPCR. Analysis of the melting curve observed for the DNA products produced

by PCR amplification can determine the presence of a specific DNA mutation of interest [26]. This

method is based on the principle that wild-type DNA will melt at a higher temperature than mutated

DNA and that the system will show 2 lower temperature melting peaks for heterozygous mutations and

a single lower temperature peak for homozygous mutations. A variation of this method is the PCR Clamp

method, which uses a peptide nucleic acid probe to block amplification of the wild-type DNA within a

sample to detect a specific mutation [27]. Although this method has high sensitivity, it cannot calculate

the percentage of mutated DNA.

8

MALDI-TOF Sequencing. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry

(MALDI-TOF MS) has been applied to clinical samples for DNA sequencing with very high resolution and

sensitivity, especially for the detection of somatic point mutations in cancer samples and single

nucleotide polymorphisms in germline DNA [28,29]. This type of DNA genotyping is the backbone of the

The Sequenom MassARRAY® system.

Hybrid-capture based comprehensive genomic profiling. The first organization to develop a second, or

next generation, approach to DNA sequencing was Lynx Therapeutics. Their “massively parallel signature

sequencing” (MPSS) platform was a microsphere (bead)-based system that read nucleotides in groups of

4 via an adapter ligation and adapter decoding strategy. Through a merger with the Solexa Corporation,

which was subsequently acquired by Illumina, Inc., this bead-based approach using reversible dye

terminators and short read lengths was adapted for use in a flow cell with 8 individual lanes, the

surfaces of which are coated with oligonucleotide anchors. In this approach, unincorporated nucleotides

are washed away after each cycle, with the remaining DNA extended one nucleotide at a time.

Subsequent system cycles take place after digital images are captured of the fluorescently labeled

nucleotides and the terminal 3' blocker is chemically removed from the DNA. By a process called

“bridge amplification,” DNA templates are amplified in the flow cell by “arching” over and hybridizing to

an adjacent anchor oligonucleotide [30]. A number of technical issues, particularly those involving

aberrant nucleotide incorporation rates, place major responsibility on the bioinformatics systems and

computational biologists to correctly interpret the raw sequencing data produced by the Illumina

systems. The Illumina technique is currently the most widely used NGS platform, and Illumina currently

markets three major clinical instruments, the HiSeq 2500, the HiSeq 3000/4000, and the MiSeq. HiSeq

platforms can sequence up to 1 trillion bases in about 3 days or approximately 10 billion bases in a rapid

run mode that takes as few as 7 hours. The MiSeq is a much cheaper, lower capacity instrument used for

rapid turnaround (it can sequence 500 million bases in 4 hours).

9

Semiconductor Sequencing. This method utilizes ion semiconductor sequencing based on the detection

of hydrogen ions that are released during the polymerization of DNA, and is the basis for the Ion Torrent

system [31]. It has been widely adapted for use in clinical molecular diagnostics laboratories.

Incorporation of nucleotides into the growing complementary DNA strand causes the release of a

hydrogen ion that triggers a hypersensitive ion sensor. This approach is now owned by Life

Technologies, which claims that PostLightTM sequencing technology has the major strength of being the

first of its kind to eliminate the cost and complexity associated with the extended optical detection

currently used in all other sequencing platforms. The uses of this system appear to be focused on rapid

and affordable short sequence determination of exons containing hotspot mutations.

Comparison of Traditional and NGS Strategies for Cancer Cell Genomic Sequencing (Table 2)

Prior to the launch of NGS testing platforms, traditional hotspot DNA sequencing had reached the

bedside for the treatment of a variety of tumors including non-small cell lung cancer (NSCLC), colorectal

cancer (CRC), hematological malignancies and melanoma (Table 2) [5,32]. Next generation technologies

are also capable of testing for each of these known driver mutations[33], and have expanded the

repertoire of genetic abnormalities that can be evaluated to include copy number changes, such as HER2

gene amplification in the context of breast or upper gastrointestinal tumors, and a wider array of

variable fusion or rearrangement events, such as those affecting ROS1 or RET in NSCLC [34–36].

Table 2 Comparison of Traditional and Next Generation Sequencing of the Cancer Cell Genome

Parameter Traditional NGS Advantage Cost (per base) High Low NGS Cost (per multiplex multi-gene “test”) High Moderate Uncertain Equipment cost Moderate High Traditional Expertise required for sequencing and data analysis Moderate High Traditional Can be performed on FFPE samples Yes Yes - Challenged by small samples, necrotic tumor, tumoral heterogeneity and very low percent of tumoral DNA in sample

Yes Yes -

Generally restricted to one gene at a time Yes No NGS

10

Can easily sequence 100’s of cancer-related genes in one sample

No Yes NGS

Generally restricted to hot spots only Yes No NGS Can easily detect deletions No Yes NGS Can easily detect translocations No Yes NGS Can easily detect gene copy number alterations No Yes NGS Sensitivity Low High NGS Turn around Time (single gene) Shorter Longer Traditional Turn around Time (per multiplex multi-gene “test”) Longer Shorter NGS

The traditional approaches to cancer cell DNA sequencing are compared with the NGS approach

in Table 2. The relative cost of the two approaches is of great importance to current and future test

providers, consumers and payers. Although the cost per base sequenced for the traditional approaches

is high, these narrow approaches focused on one gene or a few hotspots are often less expensive overall

than the cost of an NGS assay that evaluates many hundreds of genes with more expensive reagents and

equipment. Without question, the expertise, especially in computational biology, required to perform

clinical NGS testing for cancer patients is significantly higher than for traditional sequencing. In daily

clinical pathology practice, both traditional and NGS sequencing approaches are challenged by several

concerns: what is the best sample to test (e.g., primary versus metastatic tumor tissue or tumor tissue

versus circulating tumor cells); small sample size, as from fine needle aspiration biopsies (FNAs); tumoral

heterogeneity with respect to genetic abnormalities; and extensive necrosis or samples that feature a

very low percent of tumoral DNA compared with noncancerous tissue.

The restriction of traditional sequencing to analysis of one gene at a time, and within that gene

typically focused on hot spots (eg. codons 12 and 13 of exon 2 in the KRAS oncogene), is a significant

drawback. NGS platforms allow for large-scale gene sequencing that can both determine the status of

mutational hotpots “expected” in a given clinical situation and discover “unexpected” sequence

abnormalities that could significantly alter the treatment plan. Novel mutations with clinical impact

11

continue to be discovered, even for established cancer genes, but are undetectable using traditional

platforms as designed today. By analyzing read counts at given loci, NGS sequencing can provide

information on gene copy number, identifying homozygous and heterozygous deletions and gene

amplifications when traditional sequencing approaches cannot. NGS can also detect translocations that

drive therapy selection, such as the EML4-ALK translocation that is the key indication for crizotinib

treatment in NSCLC. Furthermore, the sensitivity of NGS can match or exceed traditional approaches

when the mutation is present in only a small percentage of the total DNA extracted from the specimen.

The rapid analysis of many genes in parallel, as made possible with NGS technology, also

facilitates the identification of potentially relevant clinical trials [18,34,37]. Knowing a patient’s

comprehensive genomic profile can allow for the selection of either a selective trial investigating

therapeutic strategies in the limited context of one biomarker and/or disease [34,37], or indicate that

patients could benefit from enrollment into a basket trial with potentially fewer restrictions on tumor

type or molecular profile [37].

Table 3. NGS Sequencing Platforms*

Parameter Illumina HiSeq Illumina MiSeq 454 Pyrosequencing

Ion Torrent

Basic Technique

Bead-based Dye Termination

Bead-based Dye Termination

Oil droplet based pyrosequencing

Semiconductor-based Sequencing

Major Uses SNP Genotyping CNV Analysis Targeted and Whole Exome Sequencing Whole Genome Resequencing

Clinical Molecular Diagnostics Sequencing

Whole Exome Sequencing Whole Genome Resequencing SNP Detection

Medium sized sequencing projects

Speed/TAT Intermediate 1 day to 3 days

Fastest 5-65 hours

Fastest < 1 Day

Fast

Instrument Cost

~$500-900K ~$125K for MiSeq

$500-700K, ~$108K for 454GS Junior

~$50K

Length of Reads

Intermediate 2 X 150 bp

2 X 300 bp Longest > 300 bases, up

Intermediate

12

to 1000 bases Advantages Highest

sensitivity for mutation detection Can detect all classes of genomic alterations simultaneously (mutation, copy number, rearrangement)

Low cost of instrument Rapid TAT

Fast Long reads useful for mapping

Low cost Lower bioinformatics expertise required

Disadvantages High cost Requires substantial operator expertise High bioinformatics requirements

Low depth of coverage Lack of sensitivity for low mutant allele frequency tumors

High reagent costs High error rates

Generally used as a “hotspot” test and does not cover entire exome of selected genes Lower sensitivity for mutation detection Cannot detect gene amplification or homozygous deletions Cannot detect gene fusions High error rate in homopolymer sequences

* Adapted and modified from Metzker ML. Sequencing technologies - the next generation. Nat Rev Genet. 2010;11:31-46.

Although the turn-around time for NGS of a multiplex (>100 gene) cancer genome panel is

currently longer (4-7 days) than traditional single gene hot spot sequencing, it is anticipated that this

difference will rapidly narrow as NGS technology continues to evolve Information on the patient’s

germline DNA sequences may be needed in a variety of clinical settings to make sense of the tumor cell

sequence or to distinguish rare, harmless germline polymorphisms from possibly significant somatic

mutations [38,39]. Finally, in an era of growing demand for a more personalized approach to oncology

practice, it is likely that other traditional and emerging cancer cell diagnostics, including slide-based

13

assays (IHC and FISH), analysis of the epigenome using methylation-specific RT-PCR and microRNA

profiling, will be combined with tumor cell DNA sequencing to create some form of a unified laboratory

report.

Table 4. Hot-Spot Genotyping vs. Targeted Exome NGS Vs. Whole Genome Analysis

Hot-Spot Genotyping Targeted Next Generation Sequencing

Whole Genome Analysis

Starting Material FFPE FFPE Fresh/Frozen Can be performed on small samples

Yes Yes Unknown

Exhausts the sample when multiple individual tests are ordered

Yes No No

Focused on cancer-related genomic alterations

Yes Yes No “creates the haystack and surrounds the needle”

Detects all relevant classes of genomic alterations (copy number, mutation, rearrangement)

No Yes Yes

Requires Matched Normal Tissue Sample

No Not absolutely necessary but it is preferred.

Yes

Flexible to add new genes to test

No Yes NA

Analysis Complexity and TAT

Low Intermediate High

Challenges for Delivering NGS Results for Cancer Patients

In order to deliver NGS results as a clinical assay for patient management a number of barriers

must be overcome [32], from specimen requirements and cost to turn around time and ensuring proper

analysis and interpretation of the sequencing data.

Obtaining an Adequate Sample to Sequence. Clinical NGS performed for solid tumors generally

uses formalin-fixed paraffin-embedded material [33,40], although many other tissue samples can be

14

analyzed [41]. Major resection specimens almost always provide an adequate sample, but small needle

biopsies, fine needle aspiration biopsies and fluid cell block samples may be limiting. In general, a

sample approximately 15 mm2 with a minimal depth of 40 microns is adequate for NGS [33]. For assay

systems that measure gene copy number in addition to other mutations, tumor nuclei must account for

at least 20% of the total tissue nuclei present. Contamination with non-cancerous tissue or high levels of

necrosis can affect detection sensitivity. When tumor nuclei proportions are below 20%, the risk of

missing a copy number gain or homozygous loss increases rapidly. Macro-dissection can often be used

on larger specimens to enrich the sample for tumor nuclei.

Detecting all classes of alterations. Cancer growth and progression can be driven by many

different alteration types, all of which can dysregulate the checks and balances that normally preserve

cellular homeostasis: point mutations that selectively alter enzyme activity, genomic rearrangements

that create novel oncogenic molecules, copy number gains or losses that dramatically change transcript

levels, and small insertions and deletions (indels) with various effects depending on the gene and

location of the alteration [1]. Given the variety of driver mutations possible, the collection of oligo baits

used for hybrid capture and the algorithms processing the data must be designed to probe for and

detect multiple alteration types [7,42]. Concordance between complimentary assays such as

immunohistochemistry and the measurement of copy number gains or losses provided by DNA

sequencing illustrates the power of NGS techniques for tumor profiling [33].

System Validation. Validating the sensitivity and specificity of an NGS assay is a major challenge

for test providers[33]. One approach has relied on the use of HapMap cell lines known to have specific

genomic alterations that can be diluted to low mutant allele frequencies (MAF) and run in parallel with

clinical samples. The more traditional approach is to obtain sets of samples with known mutations (as

15

defined by another method or another lab) in each of the genes of interest. However, this approach is

generally feasible only for the most commonly mutated genes.

Bioinformatics Requirements. Although the proper management of an NGS system requires

technical expertise in many areas, the bioinformatics expertise required for proper analysis and

interpretation is key [43–46]. Statistical analysis of system performance, including depth and uniformity

of sequencing coverage is typically performed by the bioinformatics team. The software identifying

alterations and determining which are clinically significant often requires local algorithm construction

and modifications needed to bring the system to full performance in sensitivity and specificity. The lack

of trained bioinformaticians capable of managing NGS data systems software is a major impediment to

the development of NGS testing services in many clinical laboratories.

Identifying Actionable Genomic Alterations. Rapid growth and cell division, coupled with

dysregulated quality control, leads tumors to quickly accumulate mutations, without these changes

necessarily conferring an advantage. These passenger mutations can arise in any gene, including well

characterized oncogenes or tumor suppressors, but are not expected to predict response to treatment

or prognosis. Distinguishing driver mutations from passengers relies on clinical or experimental

observations indicating significance. Many databases have been developed to aid the understanding and

identification of significant mutations found in cancer. Two major initiatives have been designed to map

out all the somatic intragenic mutations in cancer: The Cancer Genome Atlas of the National Cancer

Institute (http://cancergenome.nih.gov/) and the Sanger Centre’s Cancer Genome Project

(http://cancer.sanger.ac.uk/cosmic/) [47]. The COSMIC database displays the data generated from all

published or otherwise publicly available human cancer sequencing efforts, whereas The Cancer

Genome Atlas Project has banked information on cancer genomes, as well as transcriptomes and

16

proteomes. The International Cancer Genome Consortium is so far the biggest project to collect human

cancer genome data, and is accessible through the ICGC website (https://dcc.icgc.org/) [48].

The term “actionable” has been applied to somatic cancer genotyping to indicate when the

results of sequencing can direct a specific action on the part of the oncologist. Significant discussion

surrounds the general definition of actionable genomic alterations as well as the potential

“actionability” of alterations in individual genes [5,32]. There is universal agreement that alterations

associated with a specific approved therapy in that tumor type are actionable. Most investigators also

agree that an alteration indicating an approved therapy for a different tumor type is also actionable,

although it would require off-label drug use. The most controversial actionability definition concerns

alterations directly listed in entry criteria for registered anti-cancer clinical trials. In some of these

alteration-driven trials investigating targeted therapies, the association of the sequence result with the

proposed mechanism of action or clinical responses is straight forward and well-accepted, in others the

alteration and link with drug efficacy is not as well-established. Given that the FDA continues to approve

anti-cancer drugs based on their site of origin, a careful pathology review is required to assign the

correct diagnosis to the sequence results. Although curation for well-known alterations may be

relatively straightforward, as the list of anti-cancer drugs linked to genomic alterations continues to

grow, having a skilled curation team capable of searching the current literature and databases is a

critical component of NGS reporting.

NGS and The Liquid Biopsy

Liquid Biopsy Definitions: CTC: circulating tumor cells; cfDNA: cell-free DNA; ctDNA: circulating

tumor DNA. The sensitivity, specificity and overall capabilities increasing as technology advances. There

are many organizations working in this space including Guardant; Adaptive Biotechnologies; Trovagene;

Foundation Medicine and Academic Hospitals and Major Cancer Centers. Without dubt, the liquid

17

biopsy is a significant “threat” to the future role of general anatomic pathology in the management of

cancer patients.

Liquid biopsies cannot achieve the sensitivity and accuracy of a tissue based test with a blood-

based specimen due to low total tumor-derived DNA in circulation in most tumor types. Current blood-

based tests trying to “replace” tissue-based tests are both insensitive (miss critical alterations) and non-

specific (have a high false positive rate). This may change in the future as more sensitive and specific

tests emerge that can detect all classes of genomic alterations. The inaccuracies of current tests are

driven by low tumor DNA in circulation, false amplification of patient’s germline SNPs and VUS derived

from circulating WBC. That said, test approach today appears to be a focused panel for defined clinical

situations. The current indications for NGS performed on a liquid biopsy include: 1) Unavailable tissue in

any location, 2) Recurrent disease requiring a new biopsy that could be harmful to the patient,

3)Detection of specific therapy resistance associated genomic alterations, 4) EGFR T790M test now FDA

approved for AZ’s osimertinib and 5) Monitoring disease based on specific genomic alterations.

II. Clinical Utility of NGS Testing in Clinical management of patients with Solid Tumors

Roles of NGS in the Management of Patients with Cancer

• Prediction of response to conventional therapy: Limited to date but genomic signatures for

platinin and taxane drugs are emerging

• Prediction of Response to Targeted Therapies: The Main role

• Prediction of Response to Immunotherapies: Total Mutational Burden (TMB); MSI; gInterferon

Signature; PD-L1 Amplification (rare)

• Prediction of Therapy Toxicity (Germline Status): Cytochrome p450 (2D6, others) ; UGT1A1

(others)

• Determining Site of Origin in Metastatic Disease/CUP: Not a primary purpose: Many examples

(TMPRSS-ERG, Sarcoma fusions, others

18

Prior to the launch of NGS testing platforms, hotspot traditional DNA sequencing had reached the

bedside for the treatment of a variety of tumors including non-small cell lung cancer, colorectal cancer

and melanoma (Table 5).

Selected Examples of Cancer Genome Sequencing and Anti-Cancer Drug Selection

Genetic Event Disease Drug KRAS Mutation CRC Cetuximab/Panitumumab

(contraindicated by KRAS mutation) BRAF Mutation Melanoma Vemurafenib/Dabrafenib EGFR Mutation NSCLC Gefitinib/Erlotinib/Afatinib EML4_ALK Translocation NSCLC Crizotinib KIT Mutation GIST/melanoma Imatinib/Sunitinib/Regorafenib/Pazopanib BCR-ABL Translocation CML Imatinib/Dasatinib/Nilotinib/Bosutinib PML-RARA Translocation t(15;17)

APL ATRA

HER2 Gene Amplification* Breast and Upper GI Cancer Trastuzumab/Lapatinib ROS1 Fusion NSCLC Cabozantinib (investigational) RET Fusion NSCLC Cabozantinib (investigational)

Furthermore, if we look at the many targets and drugs in clinical trials, the list grows substantially. It is now believed that more than 700 compounds and 150 targets are in some phase of clinical development for cancer.

19

As seen in the table below (Simon and Roychowdry), the number of target genes, pathways impacted, aberrations detected, tumor types involved and drugs on the market or in clinical trials has become formidable:

20

The final section of this course consists of a series of clinical case studies based on tumor type designed illustrate the potential uses of NGS in the management of patients with solid tumors.

NGS and the Prediction of Response to Immunotherapies

NGS has started to emerge as a major approach to identify patients more likely to respond to immunotherapies especially the immune checkpoint inhibitors. The following biomarkers can be identified by NGS using sensitive and deep sequencing: Total Mutational Burden (TMB); MSI; γInterferon Signature; and PD-L1 Amplification (rare). Of this group of biomarkers, the total mutational burden appears to be the most predictive. Cances with high mutation burdens appear to create the greatest number of tumor protein antigens and activate T cell function.

Case Studies of NGS in the Clinic including therapy selection and disease outcome

Case 1 Non-small Cell Lung Cancer (Lindeman). Genomic alterations and their significance to be discussed include:

General genomic features of lung cancers—high rates of mutations, rearrangements, and copy number alterations

Alterations in receptor tyrosine kinase/Ras/Raf pathways in lung adenocarcinoma: EGFR amplifications and mutation; ALK fusions; ROS1 fusions; RET fusions; MET alterations including amplifications; KRAS mutations

21

Genomic alterations in other pathways in lung adenocarcinoma: p53 pathway (TP53 and MDM2); PI3 kinase pathway (PIK3CA and STK11); Chromatin modification (SMARCA4, ARID1A, and SETD2) and RNA splicing pathways; Lineage oncogenes (TTF1/NKX2-1)

Cell cycle pathways

Genomic alterations in squamous cell lung carcinomas: Receptor tyrosine kinase pathways (FGFR mutations/fusions/gains, EGFR/ERBB2, and DDR2); Cell cycle (CDKN2A); Reduction-oxidation pathways (NFE2L2, KEAP1)

Summary: need for multi-gene, multi-modality testing in NSCLC

Break (20 minutes)

Case 2 Breast Cancer (Ross). Genomic alterations and their significance to be discussed include:

ERBB2 (HER2) amplifications and mutations

ESR1 (ER alpha) mutations

Other targetable alterations (EGFR, FGFR, etc.) focused on triple negative disease

Genomic features of uncommon subtypes of breast cancer

Biomarkers of Immunotherapy Response

Case 3 Melanoma (Yip). Genomic alterations and their significance to be discussed include:

BRAF V600E

BRAF V600 D and K

BRAF Fusions

BRAF non-V600 mutations

NRAS mutations

KIT mutations

Others

22

Case 4 Brain Tumors (Yip) Genomic alterations and their significance to be discussed include:

Genomic drivers of common gliomas

Genomic drivers of rare entities

BRAF substitution vs BRAF fusion driven tumors

ATRX driven brain tumors

IDH1/2 driven brain tumors

Case 5 Carcinoma of Unknown Primary Site (Ross) Topics and issues to be discussed include:

How does the frequency of actionable genomic alterations detected by NGS in CUP compare to that found when the primary site is known?

Is finding the primary site important for patient management if NGS performed on a metastatic lesion can guide therapy selection?

Does NGS-based detection of actionable alterations in CUP have potential to improve patient outcomes for this disease?

Can NGS directed therapy for CUP be more cost effective than IHC, serum tumor markers, mRNA profiling and imaging used to search for the primary site in the management of CUP patients?

Case 6 Myeloid Sarcoma (Lindeman) Genomic alterations and their significance to be discussed include:

Sarcoma fusion gene detection

Targetable genomic alterations in sarcomas

23

References:

[1] L.A. Garraway, E.S. Lander, Lessons from the cancer genome, Cell. 153 (2013) 17–37. doi:10.1016/j.cell.2013.03.002.

[2] L.A. Garraway, Genomics-driven oncology: framework for an emerging paradigm, J. Clin. Oncol. 31 (2013) 1806–1814. doi:10.1200/JCO.2012.46.8934.

[3] J.M. Heuckmann, R.K. Thomas, A new generation of cancer genome diagnostics for routine clinical use: overcoming the roadblocks to personalized cancer medicine, Ann Oncol. (2015) mdv184. doi:10.1093/annonc/mdv184.

[4] S. Roychowdhury, A.M. Chinnaiyan, Translating Genomics for Precision Cancer Medicine, Annual Review of Genomics and Human Genetics. 15 (2014) 395–415. doi:10.1146/annurev-genom-090413-025552.

[5] R.L. Schilsky, Implementing personalized cancer care, Nat Rev Clin Oncol. 11 (2014) 432–438. doi:10.1038/nrclinonc.2014.54.

[6] M. Arnedos, P. Vielh, J.-C. Soria, F. Andre, The genetic complexity of common cancers and the promise of personalized medicine: is there any hope?, J. Pathol. 232 (2014) 274–282. doi:10.1002/path.4276.

[7] J.S. Ross, K. Wang, L. Gay, G.A. Otto, E. White, K. Iwanik, et al., Comprehensive Genomic Profiling of Carcinoma of Unknown Primary Site: New Routes to Targeted Therapies, JAMA Oncol. 1 (2015) 40–49. doi:10.1001/jamaoncol.2014.216.

[8] F. Sanger, S. Nicklen, A.R. Coulson, DNA sequencing with chain-terminating inhibitors, Proc. Natl. Acad. Sci. U.S.A. 74 (1977) 5463–5467.

[9] L.M. Smith, J.Z. Sanders, R.J. Kaiser, P. Hughes, C. Dodd, C.R. Connell, et al., Fluorescence detection in automated DNA sequence analysis, Nature. 321 (1986) 674–679. doi:10.1038/321674a0.

[10] International Human Genome Sequencing Consortium, Finishing the euchromatic sequence of the human genome, Nature. 431 (2004) 931–945. doi:10.1038/nature03001.

[11] E.S. Lander, L.M. Linton, B. Birren, C. Nusbaum, M.C. Zody, J. Baldwin, et al., Initial sequencing and analysis of the human genome, Nature. 409 (2001) 860–921. doi:10.1038/35057062.

[12] J.C. Venter, M.D. Adams, E.W. Myers, P.W. Li, R.J. Mural, G.G. Sutton, et al., The Sequence of the Human Genome, Science. 291 (2001) 1304–1351. doi:10.1126/science.1058040.

[13] E.L. van Dijk, H. Auger, Y. Jaszczyszyn, C. Thermes, Ten years of next-generation sequencing technology, Trends Genet. 30 (2014) 418–426. doi:10.1016/j.tig.2014.07.001.

24

[14] M.L. Metzker, Sequencing technologies - the next generation, Nat. Rev. Genet. 11 (2010) 31–46. doi:10.1038/nrg2626.

[15] M. Salto-Tellez, D. Gonzalez de Castro, Next-generation sequencing: a change of paradigm in molecular diagnostic validation, J. Pathol. 234 (2014) 5–10. doi:10.1002/path.4365.

[16] J. Gagan, E.M. Van Allen, Next-generation sequencing to guide cancer therapy, Genome Medicine. 7 (2015). doi:10.1186/s13073-015-0203-x.

[17] E.R. Mardis, Next-generation sequencing platforms, Annu Rev Anal Chem (Palo Alto Calif). 6 (2013) 287–303. doi:10.1146/annurev-anchem-062012-092628.

[18] D.G. Stover, N. Wagle, Precision Medicine in Breast Cancer: Genes, Genomes, and the Future of Genomically Driven Treatments, Curr Oncol Rep. 17 (2015) 1–11. doi:10.1007/s11912-015-0438-0.

[19] M. Ronaghi, M. Uhlén, P. Nyrén, A sequencing method based on real-time pyrophosphate, Science. 281 (1998) 363, 365.

[20] M. Ronaghi, S. Shokralla, B. Gharizadeh, Pyrosequencing for discovery and analysis of DNA sequence variations, Pharmacogenomics. 8 (2007) 1437–1441. doi:10.2217/14622416.8.10.1437.

[21] C.R. King, S. Marsh, Pyrosequencing of clinically relevant polymorphisms, Methods Mol. Biol. 1015 (2013) 97–114. doi:10.1007/978-1-62703-435-7_6.

[22] S. Marsh, Pyrosequencing applications, Methods Mol. Biol. 373 (2007) 15–24. doi:10.1385/1-59745-377-3:15.

[23] M. Margulies, M. Egholm, W.E. Altman, S. Attiya, J.S. Bader, L.A. Bemben, et al., Genome sequencing in microfabricated high-density picolitre reactors, Nature. 437 (2005) 376–380. doi:10.1038/nature03959.

[24] S.J. Clayton, F.M. Scott, J. Walker, K. Callaghan, K. Haque, T. Liloglou, et al., K-ras point mutation detection in lung cancer: comparison of two approaches to somatic mutation detection using ARMS allele-specific amplification, Clin. Chem. 46 (2000) 1929–1938.

[25] F.A. Monzon, S. Ogino, M.E.H. Hammond, K.C. Halling, K.J. Bloom, M.N. Nikiforova, The role of KRAS mutation testing in the management of patients with metastatic colorectal cancer, Arch. Pathol. Lab. Med. 133 (2009) 1600–1606. doi:10.1043/1543-2165-133.10.1600.

[26] P.S. Bernard, C.T. Wittwer, Real-time PCR technology for cancer diagnostics, Clin. Chem. 48 (2002) 1178–1185.

[27] H. Orum, PCR clamping, Curr Issues Mol Biol. 2 (2000) 27–30.

[28] J.R. Edwards, H. Ruparel, J. Ju, Mass-spectrometry DNA sequencing, Mutat. Res. 573 (2005) 3–12. doi:10.1016/j.mrfmmm.2004.07.021.

25

[29] I.G. Gut, DNA analysis by MALDI-TOF mass spectrometry, Hum. Mutat. 23 (2004) 437–441. doi:10.1002/humu.20023.

[30] D.R. Bentley, S. Balasubramanian, H.P. Swerdlow, G.P. Smith, J. Milton, C.G. Brown, et al., Accurate whole human genome sequencing using reversible terminator chemistry, Nature. 456 (2008) 53–59. doi:10.1038/nature07517.

[31] J.M. Rothberg, W. Hinz, T.M. Rearick, J. Schultz, W. Mileski, M. Davey, et al., An integrated semiconductor device enabling non-optical genome sequencing, Nature. 475 (2011) 348–352. doi:10.1038/nature10242.

[32] R. Simon, S. Roychowdhury, Implementing personalized cancer genomics in clinical trials, Nat Rev Drug Discov. 12 (2013) 358–369. doi:10.1038/nrd3979.

[33] G.M. Frampton, A. Fichtenholtz, G.A. Otto, K. Wang, S.R. Downing, J. He, et al., Development and validation of a clinical cancer genomic profiling test based on massively parallel DNA sequencing, Nat. Biotechnol. 31 (2013) 1023–1031. doi:10.1038/nbt.2696.

[34] J.M. Jürgensmeier, J.P. Eder, R.S. Herbst, New Strategies in Personalized Medicine for Solid Tumors: Molecular Markers and Clinical Trial Designs, Clin Cancer Res. 20 (2014) 4425–4435. doi:10.1158/1078-0432.CCR-13-0753.

[35] M. Marrone, K.K. Filipski, E.M. Gillanders, S.D. Schully, A.N. Freedman, Multi-marker Solid Tumor Panels Using Next-generation Sequencing to Direct Molecularly Targeted Therapies, PLoS Curr. 6 (2014). doi:10.1371/currents.eogt.aa5415d435fc886145bd7137a280a971.

[36] J.S. Ross, M. Cronin, Whole cancer genome sequencing by next-generation methods, Am. J. Clin. Pathol. 136 (2011) 527–539. doi:10.1309/AJCPR1SVT1VHUGXW.

[37] S. Kummar, P.M. Williams, C.-J. Lih, E.C. Polley, A.P. Chen, L.V. Rubinstein, et al., Application of Molecular Profiling in Clinical Trials for Advanced Metastatic Cancers, JNCI J Natl Cancer Inst. 107 (2015) djv003. doi:10.1093/jnci/djv003.

[38] D.V.T. Catenacci, A.L. Amico, S.M. Nielsen, D.M. Geynisman, B. Rambo, G.B. Carey, et al., Tumor genome analysis includes germline genome: Are we ready for surprises?, Int. J. Cancer. 136 (2015) 1559–1567. doi:10.1002/ijc.29128.

[39] M.E. Robson, A.R. Bradbury, B. Arun, S.M. Domchek, J.M. Ford, H.L. Hampel, et al., American Society of Clinical Oncology Policy Statement Update: Genetic and Genomic Testing for Cancer Susceptibility, JCO. 33 (2015) 3660–3667. doi:10.1200/JCO.2015.63.0996.

[40] S.Q. Wong, J. Li, R. Salemi, K.E. Sheppard, Hongdo Do, R.W. Tothill, et al., Targeted-capture massively-parallel sequencing enables robust detection of clinically informative mutations from formalin-fixed tumours, Scientific Reports. 3 (2013). doi:10.1038/srep03494.

26

[41] G. Young, K. Wang, J. He, G. Otto, M. Hawryluk, Z. Zwirco, et al., Clinical next-generation sequencing successfully applied to fine-needle aspirations of pulmonary and pancreatic neoplasms, Cancer Cytopathol. 121 (2013) 688–694. doi:10.1002/cncy.21338.

[42] J. Chmielecki, J.S. Ross, K. Wang, G.M. Frampton, G.A. Palmer, S.M. Ali, et al., Oncogenic alterations in ERBB2/HER2 represent potential therapeutic targets across tumors from diverse anatomic sites of origin, Oncologist. 20 (2015) 7–12. doi:10.1634/theoncologist.2014-0234.

[43] C. Cantacessi, A.R. Jex, R.S. Hall, N.D. Young, B.E. Campbell, A. Joachim, et al., A practical, bioinformatic workflow system for large data sets generated by next-generation sequencing, Nucleic Acids Res. 38 (2010) e171. doi:10.1093/nar/gkq667.

[44] L. Ding, M.C. Wendl, D.C. Koboldt, E.R. Mardis, Analysis of next-generation genomic data in cancer: accomplishments and challenges, Hum. Mol. Genet. 19 (2010) R188–196. doi:10.1093/hmg/ddq391.

[45] Y. Erlich, P.P. Mitra, M. delaBastide, W.R. McCombie, G.J. Hannon, Alta-Cyclic: a self-optimizing base caller for next-generation sequencing, Nat. Methods. 5 (2008) 679–682. doi:10.1038/nmeth.1230.

[46] S. Schwartz, R. Oren, G. Ast, Detection and removal of biases in the analysis of next-generation sequencing reads, PLoS ONE. 6 (2011) e16685. doi:10.1371/journal.pone.0016685.

[47] S.A. Forbes, D. Beare, P. Gunasekaran, K. Leung, N. Bindal, H. Boutselakis, et al., COSMIC: exploring the world’s knowledge of somatic mutations in human cancer, Nucl. Acids Res. 43 (2015) D805–D811. doi:10.1093/nar/gku1075.

[48] J. Zhang, J. Baran, A. Cros, J.M. Guberman, S. Haider, J. Hsu, et al., International Cancer Genome Consortium Data Portal—a one-stop shop for cancer genomics data, Database. 2011 (2011) bar026. doi:10.1093/database/bar026.