RNA-based NGS methods are preferred for the detection of NTRK gene fusions^1-3

• RNA-based NGS sequences transcribed and spliced mature mRNAs (with introns already removed)³
• Thus, RNA-based NGS avoids the difficulties associated with sequencing through challenging features of NTRK gene fusions and is able to identify in-frame fusion products^2,3,5

Library prep is a critical step in the NGS process¹

Library prep generates a pool of sequencing targets representing areas of interest ready for sequencing on an appropriate NGS platform. Target enrichment is a key component of this step. During target enrichment, DNA or RNA isolated from a sample is enriched for regions of interest and formatted for sequencing on the chosen platform.¹

Choice of target enrichment method during library prep is a key consideration when detecting NTRK gene fusions^2,3

Two potential methods for target enrichment are amplification and hybridization. Both methods have unique attributes that can influence their ability to optimally detect various genomic alterations.¹

Amplification vs Hybridization-based Capture

	Amplification	Hybridization-Based Capture
Variants Detected	Best suited for detecting known variants of interest (SNVs or indels)4 Can include known, specific gene fusions (eg, EML4-ALK) and CNVs1	Potential for broad detection of all actionable mutations (CNVs, SNVs, indels, and fusions) depending on assay design, potentially including novel gene fusion partners1
Benefits	Commonly used for small panels; ideal for SNVs, indels, and known fusions1 Requires low input of nucleic acid1	Can detect larger number of genes and genomic regions1 Sensitivity for detecting fusions with novel fusion partners1
Drawbacks	Requires complex amplicon design to sufficiently cover all genomic alterations including gene fusions1	Requires more input of nucleic acid than amplification-based methods1

Clonal amplification generates required copies of templates of interest for accurate sequencing^1,6

Depending on NGS platform, clonal amplification may be required to ensure sufficient material for successful sequencing.¹

Clonal amplification approach with Thermo Fisher Ion Personal Genome Machine, Thermo Fisher Proton, or Qiagen GeneReader^6-8

Clonal amplification approach with Illumina MiSeq or Illumina NextSeq^6,9

No clonal amplification needed with PacBio RS 1/II, Oxford nanopore, or GridlON^6,10,11

Current NGS technologies employ different sequencing methods⁶

Sequencing by synthesis using serial DNTP flow (Thermofisher Scientific): Semiconductor sequencing⁶

1 Template-enriched beads are carefully arrayed into a microtitre plate chip⁶

2

3

4

5

Sequencing by synthesis using REVERSIBLE TERMINATORS (ILLUMINA AND QIAGEN)⁶

1

2 Nucleotide addition

• All 4 uniquely fluorophore-labeled, terminally blocked nucleotides added and hybridize to complementary base. Each cluster incorporates a different base (Illumina)⁶ or
• All 4 uniquely fluorophore-labeled, terminally blocked and unlabeled, blocked nucleotides added and hybridize to complementary base. Each bead incorporates a different base (Qiagen)⁶

3 Imaging

• Slides are imaged with 2 or 4 laser channels. Each cluster emits color corresponding to base incorporated during cycle (Illumina)⁶ or
• Slides are imaged with 4 laser channels. Each bead emits color corresponding to base incorporated during cycle (Qiagen)⁶

4 Cleavage

• Fluorophores cleaved and washed away. A new cycle begins with the addition of new nucleotides⁶

NGS platform considerations^12-18

MANUFACTURER	METHOD	INSTRUMENT	RUN TIME	READ LENGTH	OUTPUT (Gb)
Illumina	Bridge PCR Chemistry: hybridization Sequence by synthesis Paired end sequencing	MiSeq v2 MiSeq v3 NextSeq HiSeq	~39 hours ~56 hours ≤35 hours 6 days	2x250 2x300 2x150 2x125	7.5-8.5 13.2-15 ≥90 450-500
ThermoFisher	Emulsion PCR Semiconductor sequencing Single end sequencing	Ion PGM (CE/IVD) Ion Proton Ion S5	5 hours 2.5 hours 12 hours	400 200 400	0.6-1 15 1.2-2
Qiagen	Emulsion PCR Sequencing by synthesis Single end sequencing	GeneReader	30 hours	134	1

Bioinformatics: Analyzing, interpreting, and reporting NGS data

Multiple bioinformatic tools available from sequencing providers and third-party vendors are evaluated, selected, and then integrated to create a validated bioinformatic process for NGS that allows for the discovery and identification of gene fusions.¹

Bioinformatics workflow: Key processes for accurate gene fusion detection¹

• Sequencing data: Platform-specific DNA sequences determined and demultiplexed to produce pool of patient specific-sequence reads¹
• Alignment: Patient-specific sequence reads assembled or mapped by comparing to a control (reference) DNA genome sequence¹
• Variant calling: DNA base pairs in the reads or segments of the reads different from the reference are noted as variants¹
• Variant annotation: Variants are compared against public or proprietary databases to determine biological relevance¹
• Variant interpretation: Relevant variants compared against public or proprietary databases to determine and add clinical context to the annotation¹
• Sequencing report: Provides key findings and actionable steps to end-user (ie, triage patients to a clinical trial or prescribe a targeted therapy)¹

Experimental and bioinformatic considerations for carrying out RNA-based NGS

Number of reads

• Total number of RNA transcript sequence reads should be optimized for system under study to support gene fusion detection³

Length of reads

• Paired-end sequencing and longer sequence reads provide data that accumulate overlapping coverage with higher specificity to fused genes during alignment³
• This can be problematic with FFPE samples that generally contain fragmented DNA/RNA³

Filtering of reads

• Reads originating from highly expressed genes that are unlikely to be involved in gene fusions, such as ribosomal RNA, must be anticipated and filtered out appropriately³

Reads ratio

• Low ratio of chimeric to wild type reads in the vicinity of the fusion boundary may indicate spurious mismapping of reads³

How to detect gene fusions

• Highly promising fusion-sequence predictions evident if sequences in the read strongly align to sequences in 1 of the genes that constitute the gene fusion³
• Gene fusions usually follow splicing patterns of wild-type genes; thus, their detection at exon boundaries that match those of the wild-type gene(s) is another powerful indicator of gene fusion sequence identification³

NGS test reporting

A clear and concise report is crucial to ensure immediate clinical use to the ordering physician⁴

The report should contain all the information required for the ordering clinician to know what was tested and the results obtained from the test. Incomplete or unclear representation of the data can lead to clinical errors and incorrect patient management.⁴

Detected genomic alterations should be classified under a 4-tiered system⁴

Tier I	Variants of strong clinical significance Therapies approved by the FDA, and professional guidelines, or on well-powered clinical studies with expert consensus Diagnostic or prognostic biomarker in specific tumor types, based on professional guidelines or well-powered clinical studies with expert consensus
Tier II	Variants of potential clinical significance Therapies based on FDA-approved therapies for different tumor types or investigational therapies Based on multiple small studies with some consensus Preclinical trials or a few case reports without consensus
Tier III	Variants of unknown clinical significance None or no convincing evidence to determine clinical/biological significance
Tier IV	Variants of known insignificance No existing published evidence of cancer association Observed at significant allele frequencies in the general population or a specific subpopulation

• Tier I–III alterations must be reported in descending order of clinical importance⁴
• Clinically useful interpretations should be provided for tier I and II variants to inform disease management decisions⁴
• Interpretations for tier III variants should be brief, bearing in mind the goal to keep critical information clear and concise⁴
• Tier IV alterations are benign or likely benign; therefore, it is recommended that they not be reported⁴

Treatment decisions are multifaceted and often based on several factors that are unknown to the molecular pathologist. Therefore, recommendations in an NGS test report should be relevant to the patient’s cancer diagnosis and contain language clarifying that the recommendations are based on data available to the laboratory, but that additional factors need to be considered to create a treatment plan for each individual.⁴

Nomenclature

Detected genomic alterations should be annotated and reported by the HUGO Gene Nomenclature Committee. Gene fusions should be reported listing both gene partners separated by a dash. For example, a gene fusion between the NTRK1 gene and the partner gene TPM3 should be reported as TPM3-NTRK1 ⁴

Use of standard nomenclature does not outweigh the goal of clear communication. Use colloquial nomenclature in addition to standard nomenclature, as needed, to clearly communicate results to the clinical provider.⁴

Other reporting elements

The report should also contain any information that can help provide a comparison with other results from the patient over time or provide additional context for the reported results, including⁴

• Genomic coordinates
• Genome build
• Transcript reference sequence
• Pertinent negatives
• Sequencing coverage cutoff for the NGS assay used
• Genes and/or hotspots that did not meet the sequencing coverage criteria should be declared as failed

Role of the Multidisciplinary Molecular Tumor Board

With rapid advancements in sequencing technology, implementing multidisciplinary molecular tumor board may be crucial to help translate genomic alteration findings into evidence-based recommendations.⁶ Diverse perspectives provide a multidisciplinary evaluation of the patient, as well as ensure expert interpretation of complex genomic data, especially when using whole-exome or whole-genome sequencing.^6,7 Thus, the molecular tumor board should invite collaboration from multiple disciplines, including clinicians with varying specialties (eg, oncologists, hematologists, pathologists, geneticists, molecular biologists, and bioinformaticians).^6,7

Molecular tumor boards can provide education, facilitate interpretation, and implementation of precision medicine by⁶

• Improving clinician’s understanding of assay strengths, limitations, and results
• Increasing clinician’s confidence in and efficiency of utilizing molecular diagnostics
• Providing information about laboratory updates, analysis software, and challenges in data interpretation and utilization