TruSight Oncology Comprehensive

P230011 · Illumina, Inc. · PQP · Aug 21, 2024 · Pathology

Device Facts

• Record ID: P230011
• Device Name: TruSight Oncology Comprehensive
• Applicant: Illumina, Inc.
• Product Code: PQP · Pathology
• Decision Date: Aug 21, 2024
• Decision: APPR
• Device Class: Class 3

Intended Use

TruSight Oncology Comprehensive is a qualitative in vitro diagnostic test that uses targeted next-generation sequencing to detect variants in 517 genes using nucleic acids extracted from formalin-fixed, paraffin embedded (FFPE) tumor tissue samples from cancer patients with solid malignant neoplasms using the Illumina® NextSeq™ 550Dx instrument. The test can be used to detect single nucleotide variants, multi-nucleotide variants, insertions, and deletions from DNA, and fusions in 24 genes and splice variants in one gene from RNA. The test also reports a Tumor Mutational Burden (TMB) score. The test is intended to be used as a companion diagnostic to identify cancer patients who may benefit from treatment with the targeted therapies listed in Table 1, in accordance with the approved therapeutic product labeling. In addition, the test is intended to provide tumor profiling information for use by qualified health care professionals in accordance with professional guidelines in oncology for patients with solid malignant neoplasms. Genomic findings other than those listed in Table 1 of the intended use statement are not conclusive or prescriptive for labeled use of any specific therapeutic product.

Device Story

Distributed IVD test using targeted NGS to profile solid tumor FFPE samples. Input: DNA/RNA extracted from FFPE tissue. Process: Library preparation, enrichment for 517 cancer-related genes, sequencing on NextSeq 550Dx. Output: Variant calls (SNVs, MNVs, indels, fusions, splice variants) and TMB score. Used in clinical laboratories by trained personnel. Software performs secondary analysis (alignment, error correction, variant calling) and tertiary analysis (annotation, CDx calling, tumor profiling). Results assist clinicians in identifying patients for targeted therapies (larotrectinib, selpercatinib) and provide tumor profiling information for oncology decision-making.

Clinical Evidence

Clinical bridging studies used retrospective testing of banked samples from larotrectinib (LOXO-TRK-14001, NAVIGATE, SCOUT) and selpercatinib (LIBRETTO-001) trials. NTRK fusion detection: PPA 89.1%, NPA 96.3% (vs. local tests). RET fusion detection: PPA 94.2%, NPA 100%. Efficacy (ORR) in TSO-positive patients was 85.3% for larotrectinib and 74.3% for selpercatinib, supporting clinical validity.

Technological Characteristics

Targeted NGS enrichment-based assay. Materials: FFPE-extracted DNA/RNA. Sequencing: NextSeq 550Dx (SBS chemistry). Panel: 517 genes. Connectivity: Local Run Manager software on instrument. Software: Local Run Manager TSO Comprehensive (US) analysis module. Algorithm: BWA-MEM (DNA alignment), STAR (RNA alignment), Nirvana (annotation).

Indications for Use

Indicated for cancer patients with solid malignant neoplasms. Used to detect genomic variants (SNVs, MNVs, indels, fusions, splice variants) in 517 genes and report TMB score from FFPE tumor tissue. Companion diagnostic for larotrectinib (NTRK1/2/3 fusions) and selpercatinib (RET fusions).

Regulatory Classification

Identification

A next generation sequencing (NGS) oncology panel is a device used for the qualitative detection of germline or somatic variants in one or more cancer-related genes. The device is intended to be used on DNA or RNA isolated from human clinical specimens.

Reference Devices

• FoundationOne® CDx (P170019/S017)
• OncomineDx Target Test (P160045/S019 and P160045/S031)
• FoundationOne® CDx (P170019/S043)

Submission Summary (Full Text)

{0}

# SUMMARY OF SAFETY AND EFFECTIVENESS DATA (SSED)

## I. GENERAL INFORMATION

Device Generic Name: Next generation sequencing oncology panel, somatic or germline variant detection system

Device Trade Name: TruSight™ Oncology Comprehensive

Device Procode: PQP

Applicant’s Name and Address: Illumina Inc., 5200 Illumina Way, San Diego, CA 92122

Date(s) of Panel Recommendation: None

Premarket Approval Application (PMA) Number: P230011

Date of FDA Notice of Approval: August 21st, 2024

Breakthrough Device: Granted breakthrough device status [previously Expedited Access Pathway (EAP)] on January 17, 2019.

## II. INDICATIONS FOR USE

TruSight Oncology Comprehensive is a qualitative in vitro diagnostic test that uses targeted next-generation sequencing to detect variants in 517 genes using nucleic acids extracted from formalin-fixed, paraffin embedded (FFPE) tumor tissue samples from cancer patients with solid malignant neoplasms using the Illumina® NextSeq™ 550Dx instrument. The test can be used to detect single nucleotide variants, multi-nucleotide variants, insertions, and deletions from DNA, and fusions in 24 genes and splice variants in one gene from RNA. The test also reports a Tumor Mutational Burden (TMB) score.

The test is intended to be used as a companion diagnostic to identify cancer patients who may benefit from treatment with the targeted therapies listed in Table 1, in accordance with the approved therapeutic product labeling.

In addition, the test is intended to provide tumor profiling information for use by qualified health care professionals in accordance with professional guidelines in oncology for patients with solid malignant neoplasms. Genomic findings other than those listed in Table 1 of the intended use statement are not conclusive or prescriptive for labeled use of any specific therapeutic product.

PMA P230011: FDA Summary of Safety and Effectiveness Data

{1}

Table 1. Companion Diagnostic Indications

| Tumor Type | Biomarker(s) Detected | Therapy |
| --- | --- | --- |
| Solid Tumors | NTRK1/2/3 fusions | VITRAKVI® (larotrectinib) |
| Non-Small Cell Lung Cancer (NSCLC) | RET fusions | RETEVMO® (selpercatinib) |

## III. CONTRAINDICATIONS

There are no known contraindications.

## IV. WARNINGS AND PRECAUTIONS

The warnings and precautions can be found in the TruSight Oncology Comprehensive labeling.

## V. DEVICE DESCRIPTION

TruSight Oncology (TSO) Comprehensive is a distributed *in vitro* diagnostic (IVD) test. The assay is for use as part of a test system with the NextSeq 550Dx instrument and sequencing reagents. The TSO Comprehensive assay contains reagents, software, and procedures for testing DNA and RNA extracted from formalin-fixed, paraffin-embedded (FFPE) tumor samples.

The assay employs qualitative next-generation sequencing (NGS) comprehensive genomic profiling (CGP) that assesses genomic variants in a large panel of cancer-related genes listed in Appendix 1. It is an enrichment based targeted NGS test comprised of library preparation and enrichment reagents to enable DNA and/or RNA sequencing from FFPE tissue from solid tumors encompassing multiple cancer types. DNA and/or RNA extracted from FFPE tissue is used to prepare libraries, which are then enriched for cancer-related genes and sequenced on the NextSeq 550Dx instrument. TSO Comprehensive user facing software allows sequencing run set up, secondary analysis, and annotation of detected variants from the sequencing results generated on the NextSeq 550Dx instrument.

TSO Comprehensive can be run with DNA and RNA, DNA only, or RNA only samples. The reported results will be reflective of the input nucleic acid.

The test detects single nucleotide variants (SNVs), multi-nucleotide variants (MNVs) and insertions and deletions from DNA, and gene fusions and a splice variant from RNA. The assay detects small DNA variants in 517 genes, RNA fusions in 24 genes and RNA splice variants in one gene. The TSO Comprehensive assay also reports a Tumor Mutational Burden (TMB) score.

## Test Output

The output of the test includes results from the following levels:

PMA P230011: FDA Summary of Safety and Effectiveness Data

{2}

Level 1: Companion Diagnostic (CDx) Claims noted in Table 1 of the Intended Use
Level 2: Cancer Mutations with Evidence of Clinical Significance
Level 3: Cancer Mutations with Evidence of Potential Clinical Significance

Genomic findings other than those listed in Table 1 of the Companion diagnostic indications table of the intended use statement (i.e., Levels 2 and 3) are not prescriptive or conclusive for labeled use of any specific therapeutic product.

## Test Kit Contents

The TSO Comprehensive assay is for use as part of a test system with the NextSeq 550Dx instrument and associated sequencing reagents. Components of the TSO Comprehensive assay are listed in Table 2. Illumina provides reagent kits and software (Local Run Manager TSO Comprehensive (US) analysis module), which supports both tumor profiling and companion diagnostic (CDx) claims (please refer to the "Software" section below for additional details). The Knowledge Base (updated monthly) is also required to be used with the assay for tumor profiling and is available for download on the Illumina Lighthouse Portal. The assay contains reagents with sufficient volume to generate 24 DNA and 24 RNA libraries inclusive of patient samples and controls (sold separately). Materials required but not provided are described in the text below Table 2. A detailed list of required instruments, software, reagents, consumables, and storage conditions is further described in the product labeling (TSO Comprehensive Package Insert).

Table 2. Reagent Components of the TSO Comprehensive Assay

| Storage Temp (°C) | Component Name | Volume (uL) | Quantity |
| --- | --- | --- | --- |
| TSO Comprehensive RNA Library Prep | | | |
| -25 to -15 | First Strand Synthesis Mix (FSM) | 260 | 1 |
| -25 to -15 | Second Strand Mix (SSM) | 720 | 1 |
| -25 to -15 | Elution Primer Frag Mix (EPH3) | 250 | 1 |
| -25 to -15 | Reverse Transcriptase (RVT) | 70 | 1 |
| TSO Comprehensive Library Prep | | | |
| -25 to -15 | End Repair A-tailing A (ERA1-A) | 85 | 2 |
| -25 to -15 | End Repair A-tailing B (ERA1-B) | 210 | 2 |
| -25 to -15 | Adapter Ligation Buffer 1 (ALB1) | 1730 | 2 |
| -25 to -15 | DNA Ligase 3 (LIG3) | 190 | 2 |
| -25 to -15 | Short Universal Adapters 1 (SUA1) | 290 | 1 |
| -25 to -15 | UMI Adapters v1 (UMI) | 290 | 1 |
| -25 to -15 | Stop Ligation Buffer (STL) | 480 | 2 |
| -25 to -15 | Enhanced PCR Mix (EPM) | 550 | 2 |
| 2 to 8 | Resuspension Buffer (RSB) | 12400 | 1 |
| 2 to 8 | Sample Purification Beads (SPB) | 6110 | 2 |
| 2 to 8 | TE Buffer (TEB) | 10000 | 1 |
| TSO Comprehensive Index Primers | | | |
| -25 to -15 | UP Index Primers | 24 | 16 |
| -25 to -15 | CP Index Primers | 20 | 16 |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{3}

| Storage Temp (°C) | Component Name | Volume (uL) | Quantity |
| --- | --- | --- | --- |
| TSO Comprehensive Enrichment | | | |
| 2 to 8 | Target Capture Buffer 1 (TCB1) | 870 | 2 |
| 2 to 8 | Streptavidin Mag Beads (SMB) | 7780 | 2 |
| 2 to 8 | 2N NaOH (HP3) | 400 | 2 |
| 2 to 8 | Elute Target Buffer 2 (ET2) | 290 | 2 |
| 2 to 8 | Library Normalization Beads 1 (LNB1) | 1040 | 1 |
| 2 to 8 | Library Normalization Wash 1 (LNW1) | 4800 | 2 |
| 2 to 8 | Library Normalization Storage Buffer 1 (LNS1) | 3500 | 2 |
| 2 to 8 | Resuspension Buffer (RSB) | 12400 | 1 |
| 2 to 8 | Sample Purification Beads (SPB) | 6110 | 2 |
| -25 to -15 | Target Capture Additives 1 (TCA1) | 521 | 2 |
| -25 to -15 | Enhanced Enrichment Wash (EEW) | 50400 | 1 |
| -25 to -15 | Enrichment Elution 2 (EE2) | 1650 | 3 |
| -25 to -15 | Enhanced PCR Mix (EPM) | 550 | 2 |
| -25 to -15 | PCR Primer Cocktail 3 (PPC3) | 150 | 2 |
| -25 to -15 | Library Normalization Additives 1 (LNA1) | 4600 | 1 |
| -25 to -15 | PhiX Internal Control (PX3 or PhiX) | 10 | 1 |
| TSO Comprehensive Content Set | | | |
| -25 to -15 | Oncology RNA Probe Pool (OPR1) | 290 | 1 |
| -25 to -15 | Oncology DNA Probe Pool 2 (OPD2) | 290 | 1 |

## Reagents/Consumables Required but not Provided

For a detailed list of required, but not provided reagents and consumables, please refer to the product labeling (TSO Comprehensive Package Insert).

- DNA/RNA Extraction and Purification Reagents
- DNA/RNA Quantification Reagents
- TruSight Oncology DNA Control
- TruSight Oncology RNA Control
- Ethanol (EtOH) 100% (200 proof), molecular biology grade
- RNase/DNase-free water
- NextSeq 550Dx High-Output Reagent Kit v2.5 (300 cycles)

## Required Equipment/Materials, Not Provided

For a detailed list of required, but not provided equipment, please refer to the product labeling (TSO Comprehensive Package Insert).

- NextSeq 550Dx Instrument
- Ultrasonicator
- Thermal cycler
- Vortexer
- Microsample incubators
- Microcentrifuge

PMA P230011: FDA Summary of Safety and Effectiveness Data

{4}

- Plate Centrifuge
- Plate shaker
- Sealing wedge or roller
- Magnetic Stand
- Precision pipettes
- Pipette-aid
- Ice or cold block
- 10 mL serological pipettes
- Adhesive seals for 96-well plates
- Microcentrifuge tubes (1.7 mL and 2 mL), nuclease-free
- Nuclease-free reagent reservoirs
- 15 mL and 50 mL conical tubes
- Aerosol-resistant pipette tips
- 96-well storage plates, 0.8 mL
- 96-well PCR plates compatible with thermal cycler, 0.2 ml (polypropylene wells)
- Dry heat block

## Software

The Local Run Manager TSO Comprehensive (US) analysis module resides on the NextSeq 550Dx instrument as part of the Local Run Manager software to facilitate TSO Comprehensive assay run setup and to perform the secondary analysis of sequencing results. It contains the following components:

- TSO Comprehensive Claims Packages
- TSO Comprehensive Software Suite
- TSO Comprehensive USB Kit

Knowledge Base: Updated regularly and available for download on the Illumina Lighthouse Portal.

## Instruments

TSO Comprehensive assay is validated for use on the Illumina NextSeq 550Dx instrument as part of a test system. An Illumina service representative installs the appropriate version of the TSO Comprehensive (US) analysis module on the Local Run Manager NextSeq 550Dx instrument, including the Knowledge Base, prior to use of the TSO Comprehensive assay and conducts training for the end user. Other required equipment and the specifications for the specific equipment for use with the TSO Comprehensive assay are described in the "Required Equipment/Materials, Not Provided" section.

## Principles of Operation

The TSO Comprehensive assay involves the processes as described below.

## Specimen Requirements, Collection, and Preparation

The TSO Comprehensive assay requires nucleic acid (DNA and/or RNA) isolated from FFPE tissue specimens. Before performing the TSO Comprehensive assay, tissue samples should be examined by a pathologist to ensure that it is appropriate for this test. Tissue cannot be decalcified and should be fixed using formalin fixative suitable for molecular

PMA P230011: FDA Summary of Safety and Effectiveness Data
5 of 176

{5}

analyses (for example, 10% neutral-buffered formalin). Recommended tissue volume is ≥ 1.0 mm³, which is equivalent to a cumulative viable tissue area (the sum of the viable tissue area in all sections submitted for extraction) of ≥ 200 mm² using 5 μm thick sections, or ≥ 100 mm² using 10 μm thick sections. Tumor content for RNA variants depends on the extent of expression. A minimum of 20% tumor content (by area) is required to detect somatic driver mutations. A high amount of necrotic tissue (≥25%) can interfere with the TSO Comprehensive assay's ability to detect RNA fusions. If sample sections contain more than 25% necrosis in total tissue area, then the necrotic tissue must be macrodissected.

## Nucleic Acid Extraction

The TSO Comprehensive assay requires nucleic acid (DNA and/or RNA) isolated from FFPE tissue using appropriate extraction methods. The required input for the TSO Comprehensive assay is 40 ng RNA and/or 40 ng DNA. For DNA, extraction kits should be able to yield this input amount at a minimum concentration of 3.33 ng/uL. Shearing requires a final volume of 52 μl (0.77 ng/μl) with a minimum of 40 μl TEB (provided) used as the diluent. For RNA, extraction kits should be able to yield this input amount in a final volume of 8.5 uL (minimum concentration of 4.7 ng/uL). Proteinase K or equivalent enzyme should not be increased during extraction from the standard concentration provided (interference was tested using 0.04 mg/mL, as well as 2.6 mg/mL and 5.2 mg/mL, 2 times and 4 times the excess from a concentration of 1.3 mg/mL per assay reaction, respectively). This assay has been validated with extracted DNA stored at -25°C to -15°C for up to 28 days, and extracted RNA stored at -85°C to -65°C for up to 28 days.

## Library Preparation

Library preparation is performed using the TSO Comprehensive Library Prep Kit (PN 20031118, PN 20031119) for DNA and the RNA Library Prep Kit (PN 20031127) for RNA. For RNA, 40 ng total is converted to double-stranded complementary DNA (cDNA). For genomic DNA (gDNA), 40 ng of gDNA is sheared into small fragments. Universal adapters for sequencing are ligated onto the cDNA and gDNA fragments. P5 and P7 adapter sequences are incorporated into each library to enable the capture of library fragments onto the surface of the flow cell during sequencing. The adapters include i5 and i7 index sequences to identify each individual sample and, in the case of libraries from gDNA samples, individual molecules with the use of Unique Molecular Identifiers (UMIs).

## Enrichment

Resultant libraries are enriched for specific genes of interest using a capture-based method. Biotinylated probe sequences that span gene regions of interest targeted by the assay are hybridized to the libraries. The probes and hybridized targeted libraries are isolated from non-targeted libraries by capture with streptavidin magnetic particles. The targeted enriched libraries are washed and amplified. The quantity of each enriched library is normalized using a bead-based method to ensure equal representation in the pooled libraries for sequencing.

## Sequencing

PMA P230011: FDA Summary of Safety and Effectiveness Data
6 of 176

{6}

Normalized, enriched libraries are pooled and clustered onto a flow cell, and then sequenced using sequencing by synthesis (SBS) chemistry on the NextSeq 550Dx instrument. SBS chemistry uses a reversible terminator method to detect single, fluorescently labeled deoxynucleotide triphosphate (dNTP) bases as they are incorporated into growing DNA strands. During each sequencing cycle, a single dNTP is added to the nucleic acid chain. The dNTP label serves as a terminator for polymerization. After each dNTP incorporation, the fluorescent dye is imaged to identify the base, and then cleaved to allow incorporation of the next nucleotide. Four reversible terminator-bound dNTPs (A, G, T, and C) are present as single, separate molecules. As a result, natural competition minimizes incorporation bias.

## Data Analysis

### Primary Analysis

During the primary analysis, base calls are made directly from signal intensity measurements during each sequencing cycle, resulting in base-by-base sequencing. A quality score is assigned to each base call, and the sequencing data is stored in binary base call (BCL) format.

### Secondary Analysis

Secondary analysis includes the following sequential steps:

- Validation of run processing and quality control

Sequencing run quality metrics are evaluated to determine if they are within an acceptable range. Please refer to Table 3 for additional information about run-level quality metrics.

- FASTQ File Generation

Sequencing data stored in BCL format is demultiplexed using index sequences unique to each sample added during the library preparation step to assign clusters to the library from which they originated. Each cluster contains two indexes (i5 and i7 sequences, one at each end of the library fragment). The combination of those index sequences is used to demultiplex the pooled libraries. After demultiplexing, FASTQ files are generated, which are files that contain the sequencing reads for each individual sample library and the associated quality scores for each base call, excluding reads from any clusters that did not pass filter.

- Alignment and Error Correction

Sequencing reads derived from DNA sample libraries are aligned to a reference genome using the Burrows-Wheeler Aligner (BWA-MEM) to align DNA sequences to the hg19 reference genome, generating BAM files (*.bam), and BAM index files (*.bam.bai). These files are further processed to remove errors (including errors introduced during PCR amplification or sequencing). Reads derived from the same unique DNA molecule are collapsed into a single representative sequence using their unique molecular identifier (UMI). Indel realignment is performed to recover signal for insertions and deletions that may have been lost during initial alignment.

PMA P230011: FDA Summary of Safety and Effectiveness Data

{7}

Simultaneously, overlapping read pairs are bioinformatically combined into a single consensus read. All reads are then output as a final set of BAM files with corresponding BAM index files.

Sequencing reads derived from RNA sample libraries are downsampled to approximately 30 million reads per RNA sample library by randomly selecting reads from the input FASTQ files following a probability distribution. The ends of the RNA sequences are then trimmed to a maximum length of 76 base pairs and are aligned to the hg19 reference genome. Candidate splice junctions are identified, generating BAM files and BAM index files for aligned reads, and a tab-delimited text file for candidate splice junctions, using STAR aligner. Duplicate reads are then marked in the BAM files so they can be excluded from downstream steps.

- Variant Calling

Small variant calling is performed for DNA sample libraries to detect small variants [single-nucleotide variants (SNVs), multi-nucleotide variants (MNVs) up to 3 base pairs (bp) in length, insertions, and deletions. TSO Comprehensive performance has been demonstrated for insertions up to 24 base pairs and deletions up to 25 base pairs. The error-corrected BAM files (collapsed and insertions and deletions realigned) are used as input by an initial variant calling algorithm to detect small variants. This creates an unfiltered genome Variant Call Format (gVCF) file, which contain reference or variant calls for each targeted locus. Candidate variants are then filtered for assay-specific and sample processing artifacts by using adjusted quality scores calculated by comparing the observed variant frequency against a baseline noise distribution for the same site (assay-specific artifacts) and by comparing the evidence for the candidate variant against the baseline level of noise associated with that base change in the sample (sample-specific artifacts). This results in filtered gVCF files, which then uses a phased variant caller to identify MNVs, indels, and deletions in the EGFR and RET genes that are candidates for phasing. Overlapping reads are clustered in the neighborhood into a minimal set of clusters that contain the same variants. Variants are detected by examining the Concise Idiosyncratic Gapped Alignment Report (CIGAR) strings in the BAM file and comparing read sequences to the reference genome sequence. Finally, MNVs, indels, and deletions detected by the phased variant caller are merged into the filtered gVCF files resulting in merged gVCF files.

Fusion calling and splice variant calling are performed for RNA sample libraries. For fusion calling, candidate fusions are identified from supplemental alignments and anomalous read pairs (reads aligning to different chromosomes or in unexpected orientations) in the BAM files for the fusion genes targeted by the TSO Comprehensive assay. These reads are assembled into candidate fusion contigs, which are then aligned back to the reference genome (hg19). These candidate fusion contigs are then evaluated against various filters before being reported as detected (please refer to 'RNARNA Fusion Calling' in the Local Run Manager TruSight Oncology Comprehensive (US) analysis module Workflow Guide for additional information). For splice variant calling, candidate splice variants (junctions) from the alignment stage are compared against a database of known transcripts and a splice variant baseline of non-tumor junctions. Any splice variants that match the database or

PMA P230011: FDA Summary of Safety and Effectiveness Data
8 of 176

{8}

baseline are filtered out unless they are in a set of junctions with known oncological function. If there is sufficient read support, the candidate splice variant is kept.

## Tertiary Analysis

Tertiary analysis, performed by the Local Run Manager TSO Comprehensive (US) analysis module, consists of annotation of the different variant types followed by the calling/score calculation specific to the intended variants.

- **Annotation**

Detected small DNA variants are annotated using the Nirvana annotation engine with information from the RefSeq database and various population databases (COSMIC, ClinVar, dbSNP, 1000 Genomes, and gnomAD). For TMB, using the gVCF from the small variant phasing step as input, Nirvana annotates filtered small variant calls with static (not updatable) annotation databases for use by the downstream TMB calculation. For tumor profiling variants, Nirvana annotates filtered small variant calls with an updatable RefSeq database (included as part of the KB and may be updated periodically).

RNA fusions identified during RNA fusion calling are merged with fusions from proximal genes identified during RNA splice variant calling. The merged fusions are then annotated with gene symbols or names corresponding to a static database of transcripts (GENCODE Release 19).

Detected RNA splice variants are annotated using the Nirvana annotation engine with information from the RefSeq database. Nirvana annotates detected RNA splice variant calls with a static (not updatable) RefSeq database for use by downstream Companion Diagnostic calling. Splice variants are annotated with transcript-level changes (affected exons in the gene transcript) with respect to RefSeq, the same as the static RefSeq database used by the small variant annotation process. For tumor profiling variants, Nirvana annotates detected RNA splice variant calls with an updatable RefSeq database (included as part of the KB and may be updated periodically). Splice variants are annotated with transcript-level changes (affected exons in the gene transcript) with respect to RefSeq.

- **TMB Score Calculation**

Calculation of the TMB score [which includes SNVs, insertion, and deletion variants and is derived from the count of non-driver somatic variants per megabase (evaluable region)] is generated from the gVCF file generated by the Small Variant Filter step and the annotations generated during small variant Annotation by calculating the number of somatic non-hotspot variants with VAF ≥ 5% divided by the evaluable region size. Driver mutations are identified and filtered based on COSMIC count. Variants are flagged as likely germline for calculating the TMB score, applying a combination of population database and post-database filtering strategies. After database filtering, the proximity filter labels variants as germline if they are surrounded by database-labeled germline variants, and variants identified as likely germline are excluded from the TMB score calculation. The evaluable region is the

PMA P230011: FDA Summary of Safety and Effectiveness Data
9 of 176

{9}

dynamically adjusted per sample based on sequencing depth, and genomic regions with a high background noise level are excluded from the TMB calculation.

- **Companion Diagnostic (CDx) Calling**
For each installed companion diagnostic (CDx) intended use, the TSO Comprehensive (US) analysis module determines the applicability of the CDx intended use for each patient sample based on the tumor type of the patient sample. If the tumor type of the patient is an exact match or a descendant of the tumor type for a CDx intended use, it is considered applicable to that CDx intended use. If the tumor type of the patient is not applicable to a CDx intended use, then the CDx intended use is not evaluated for that sample.

If a required nucleic acid type for a CDx intended use is not sequenced or fails QC, or required controls in the run fail, then the patient sample is not evaluated for that CDx intended use. If a variant type or biomarker required for a CDx intended use fails QC, then the patient sample is not evaluated for that CDx intended use. When it is determined that a CDx intended use is applicable for a patient sample, the required libraries are sequenced, required QC measures pass, and required controls pass, the companion diagnostic intended use is evaluated for the patient sample. Detected variants and/or biomarkers in the patient sample are evaluated to determine the result for the CDx intended use. The evaluation is done through an algorithm specific to the CDx intended use, which assesses the presence and/or absence of variants/biomarkers that match the CDx intended use.

- **Tumor Profiling Variant Calling**
After companion diagnostic results are determined, all passing, detected variants in a patient sample are matched against the installed Knowledge Base Base to determine the cancer mutations that have evidence of clinical significance (Level 2) or have potential clinical significance (Level 3). A genomic finding is either a single variant with evidence of clinical significance or potential clinical significance, or a grouping of variants that, when detected together, have evidence of clinical significance or potential clinical significance.

- When multiple variants are listed together as a genomic finding, it means that there is evidence for clinical significance or potential clinical significance for those variants together, in at least one of the sources listed in the Informatics Details of the report. If there are multiple genomic findings, and a variant is included in more than one of these findings, then that variant may be listed more than one time on a report. A genomic finding with a single variant will only be listed at the highest level, including the "Companion Diagnostics Results" section, where it meets criteria for reporting. Refer to "Positive CDx Results" section in the Local Run Manager TruSight Oncology Comprehensive (US) analysis module Workflow Guide for additional details.

**Results Report Generation**
The interpreted variant results and the TMB score are summarized in the TSO Comprehensive assay results report generated for the testing laboratory, which consists of

PMA P230011: FDA Summary of Safety and Effectiveness Data
10 of 176

{10}

a "Companion Diagnostics Results" section, an "Other Alterations and Biomarkers Identified" section for reporting "Cancer Mutations with Evidence of Clinical Significance" (Level 2) or "Cancer Mutations with Potential Clinical Significance" (Level 3), a "Companion Diagnostics QC" section, and a "Companion Diagnostics Intended Use Evaluated" section. Please refer to the LRM TSO Comp Workflow Guide for additional information about the test reports.

# Controls

# Positive Controls

The TruSight Oncology DNA Control and TruSight Oncology RNA Control are required for use as positive controls. These controls consist of a multiplexed blend of biosynthetic DNA/RNA with multiple verified sequence mutations in a background of genomic DNA/RNA. The controls should be included for each DNA and RNA sequencing run (including combined runs) within a given library preparation event. The controls will be checked for quality of library preparation after sequencing. Failure of the control to meet the pre-defined quality metrics will result in all test samples on the run related to that specific control (DNA Control for DNA samples, RNA Control for RNA samples) to be invalidated.

# No Template Control (NTC)

A no template control (NTC) consists of Tris-EDTA Buffer solution (provided in TSO Comprehensive kit as reagent "TEB") for DNA samples, and DNase/RNase-free water (not included in the TSO Comprehensive kit) for RNA samples to determine if the library preparation is free of contamination by evaluating for the presence of sequencing reads (median exon coverage for DNA, number of genes with median deduplicated coverage for RNA) that map to the genes targeted by the assay. The NTC should be included for each DNA and RNA sequencing run (including combined runs) within a given library preparation event. If a NTC control fails, the assay instructions for use will require that the entire library preparation event is repeated for all samples.

# Quality Metrics

Quality metrics (Table 3) are assessed across the following categories:

- Run-level: Metrics that are quantified per sequencing run; if the external control fails these criteria, no results are reported for the entire batch of samples.
- Sample-level: Metrics that are quantified per sample; no device results are generated for samples failing these metrics.
- Analyte-level: Metrics that are quantified for individual alteration types and positions. Variants passing analyte-level metrics are reported.

Table 3. Summary of TSO Comprehensive post-sequencing library quality control metrics

| Variant Type | Metric | Run/Sample/ Analyte | Required Value | Description | Impact of Failure |
| --- | --- | --- | --- | --- | --- |
| All variants | PCT_RF_READ | Run-Level | ≥ 80.0 | Percentage of | Sequencing run |
| All variants | PCT_RF_REV | Run-Rev | ≥ 80.0 | Percentage of | Sequencing run |
| All variants | PCT_RF_SD | Run-SD | ≥ 80.0 | Percentage of | Sequencing run |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{11}

| Variant Type | Metric | Run/Sample/ Analyte | Required Value | Description | Impact of Failure |
| --- | --- | --- | --- | --- | --- |
| | S (%) | | | reads passing filter (PF). | invalidated. No results reported for any sample in the run. |
| | PCT_Q30_R1 (%) | Run-Level | ≥ 80.0 | Average percent of base calls with quality score of Q30 or higher for Read 1. | |
| | PCT_Q30_R2 (%) | Run-Level | ≥ 80.0 | Average percent of base calls with quality score of Q30 or higher for Read 2. | |
| DNA variants | CONTAMINATION_SCORE | Sample-Level | ≤ 3106 OR (> 3106 and P_VALUE ≤ 0.049) | A metric assessing the likelihood of contamination using the VAF of common variants. The contamination score is based on VAF distribution of SNPs. The contamination P value is used to assess highly rearranged genomes. It is only applicable when contamination score is above Upper Spec Limit. | No TMB or small variant results called |
| | MEDIAN_INSERT_SIZE (bp) | Sample-Level | ≥ 70 | The median fragment length in the sample. | |
| | MEDIAN_EXON_COVERAGE | Sample-Level | ≥ 150 | Median exon fragment coverage across all exon bases. | |
| | PCT_EXON_50 | Sample-Level | ≥ 90.0 | Percent exon | |

PMA P230011: FDA Summary of Safety and Effectiveness Data
12 of 176

{12}

| Variant Type | Metric | Run/Sample/ Analyte | Required Value | Description | Impact of Failure |
| --- | --- | --- | --- | --- | --- |
| | X (%) | | | bases with 50X fragment coverage. | |
| | Small DNA Variants^{6} | Analyte-Level | AQ ≥ 20 for frequent COSMIC mutations^{1}
AQ ≥ 60 for other mutations^{1} | Metric to determine whether a small DNA variant is called | Small DNA variant not called |
| | Small DNA Variants^{6} | Analyte-Level | LQ ≥ 20 for frequent COSMIC mutations^{2}
LQ ≥ 60 for other mutations^{2} | Metric to determine whether a small DNA variant is called | |
| | TMB | Analyte-Level | N/A^{3} | N/A | N/A |
| RNA variants | Median_CV_Gene_500X | Sample-Level | ≤ 0.93 | For each gene with at least 500x coverage, the coefficient of variation in coverage across the gene body is computed. This metric is the median of these values. A high value indicates a high level of variation and indicates a problem in library preparation such as low sample input and/or probe pulldown issues. This metric is computed using all reads (including reads marked as duplicates). | No fusions or splice variant results called. |
| | Total_On_Target | Sample-Level | ≥ 9,000,000 | The total number | |

PMA P230011: FDA Summary of Safety and Effectiveness Data
13 of 176

{13}

PMA P230011: FDA Summary of Safety and Effectiveness Data
14 of 176

| Variant Type | Metric | Run/Sample/ Analyte | Required Value | Description | Impact of Failure |
| --- | --- | --- | --- | --- | --- |
| | _Reads | | | of reads that map to the target regions. This metric is computed using all reads (including reads marked as duplicates). | |
| | MEDIAN_INSE RT_SIZE (bp) | Sample-Level | ≥ 80 | The median fragment length in the sample. | |
| | RNA Fusions | Analyte-Level | Fusion score ≥ 0.45^{4}
Deduped supporting reads ≥ 5 | Metric to determine whether a RNA fusion gene is called | RNA fusion not called |
| | RNA Splice Variants | Analyte-Level | Splice score = 1^{5} | Metric to determine whether a RNA splice variant is called | RNA splice variant not called |

1. Variant quality score (AQ) is calculated as -10*log₁₀(p-value), where p-value is the binomial test by comparing the observed VAF and depth to a baseline of normal FFPE samples.
2. Likelihood ratio (LQ) is calculated as -10log₁₀ (Likelihood (observed variant is an error | DP error rate/Likelihood (observed variant is a mutation | DP, VAF)
3. Evaluation of TMB does not have a reporting threshold, as the value is quantitative (score), not qualitative (classification)
4. Fusion score is a weighted sum that takes into account the number of unique supporting reads, fusion contig length, breakpoint homology, and other associated metrics.
5. Splice score increments by 0.1 for each supporting reads for the splice junction and is capped at 1.
6. Accuracy of small DNA tumor profiling variants below 5% variant allele frequency (VAF) has not been established.

Table 4. TruSight Oncology Control Metrics

| Output Type | Metric | Specification | Impact of Specification Failure |
| --- | --- | --- | --- |
| Positive Control | DNA External Control | 23 of 24 specified variants detected | No small DNA variant or TMB results reported. |
| | RNA External Control | 12 of 13 specified variants detected | No fusions or splice variant results reported. |
| Non-Template Control (NTC) | DNA Median Exon Coverage | ≤8 | No small DNA variant or TMB results reported |
| | RNA Gene above Median Cutoff | ≤1 | No fusions or splice variant results reported. |

{14}

VI. ALTERNATIVE PRACTICES AND PROCEDURES

There are FDA approved companion diagnostic (CDx) alternatives for the detection of genetic alterations using FFPE tumor specimens as listed in Table 1 of the TSO Comprehensive assay intended use statement (Section II). The approved CDx tests are listed in Table 5 below.

Table 5. Alternative FDA-approved CDx assays for CDx biomarkers identified by the TSO Comprehensive Assay

| Indication | Gene | Device (PMA) | Company | Technology | Therapy |
| --- | --- | --- | --- | --- | --- |
| Solid Tumors | NTRK1/2/3 Fusions | FoundationOne® CDx (P170019/S017) | Foundation Medicine, Inc. | NGS | VITRAKVI® (larotrectinib) |
| Non-Small Cell Lung Cancer (NSCLC) | RET Fusions | OncomineDx Target Test (P160045/S019 and P160045/S031) | Life Technologies Corporation | NGS | RETEVMO® (selpercatinib) |
| Solid Tumors | RET Fusions | FoundationOne® CDx (P170019/S043) | Foundation Medicine, Inc. | NGS | RETEVMO® (selpercatinib) |

For additional details see the FDA List of Cleared or Approved Companion Diagnostic Devices at: https://www.fda.gov/medical-devices/vitro-diagnostics/list-cleared-or-approved-companion-diagnostic-devices-vitro-and-imaging-tools

VII. MARKETING HISTORY

The TSO Comprehensive assay was originally CE-marked and commercially available March 2022, as ‘TruSight™ Oncology Comprehensive (EU)’ for use as a tumor profiling test in the European Union.

In May 2022, the NTRK1/2//3 companion diagnostic (CDx) claim was added to the TruSight Oncology Comprehensive (EU) Intended Use and it became commercially available in the European Union on May 24, 2022.

VIII. POTENTIAL ADVERSE EFFECTS OF THE DEVICE ON HEALTH

Failure of the device to perform as expected or failure to correctly interpret test results may lead to incorrect TSO Comprehensive assay results and subsequently improper patient management decisions. Patients with false positive results may undergo treatment with the therapy listed in the intended use statement without clinical benefit and may experience adverse reactions associated with the therapy. Patients with false negative results may not be

PMA P230011: FDA Summary of Safety and Effectiveness Data
15 of 176

{15}

considered for treatment with the indicated therapy. There is also a risk of delayed results, which may lead to delay of treatment with the appropriate indicated therapy.

No adverse events were reported in connection with the clinical studies used to support this PMA as the studies were performed retrospectively using banked samples. For the specific adverse events related to the approved therapeutics, please see approved FDA therapeutic product labeling.

## IX. SUMMARY OF NONCLINICAL STUDIES

### Laboratory Studies

Performance characteristics for the TSO Comprehensive assay were established using DNA and RNA derived from a wide range of FFPE tissue types. FFPE samples were tested for analysis of small DNA variants, fusions, and splice variants, including CDx and Tumor Profiling variants, and a range of TMB scores.

### 1. Accuracy – Comparison to an Orthogonal Method

#### i. CDx Accuracy

##### NTRK1, NTRK2, NTRK3 Fusion Detection in Solid Tumor Samples

The accuracy of the TSO Comprehensive assay for detecting NTRK fusions (NTRK1, NTRK2, or NTRK3) in patients with solid tumors was evaluated by assessing the concordance of NTRK fusion results between the TSO Comprehensive assay and a validated NGS-based comparator method as one of the studies to support the use of the TSO Comprehensive as a companion diagnostic (CDx) for larotrectinib. A retrospective, noninterventional study was conducted using a total of 516 FFPE samples consisting of both larotrectinib clinical trial samples (See Section X. for study details) and supplemental negative samples. These samples were tested with the TSO Comprehensive assay at one external site and with the NGS-based comparator method at a central laboratory. Accuracy of the TSO Comprehensive NTRK fusion calls by the assay was estimated relative to the comparator method. Positive percent agreement (PPA), negative percent agreement (NPA), and the associated two-sided 95% confidence intervals (CIs) were calculated.

Of the 516 samples, 499 were tested by both methods, with seventeen of the samples not tested with one of the assays due to failed extraction, unknown reason (for the comparator method), or protocol deviation. Of the 499 samples tested by both methods, 170 (34.1%) were larotrectinib trial samples and 329 (65.9%) were supplemental samples. In addition, 85 samples (72 supplemental, 13 clinical trial) had invalid TSO Comprehensive assay results (invalid rate is 17%). Of these 85 samples, 53 also had invalid comparator method results, and the comparator method had 7 additional invalid results (32 total samples with invalid TSO Comprehensive results, but valid comparator results). In total, 407 of the 499 samples had valid results by both methods.

PMA P230011: FDA Summary of Safety and Effectiveness Data
16 of 176

{16}

The concordance analysis is shown in Table 6 (aggregate) and Table 6 (by gene). Based on valid test results, aggregate PPA was $96.6\%$ (114/118; $95\%$ CI: $91.5\% - 99.1\%$ ) and aggregate NPA was $94.5\%$ (273/289; $95\%$ CI: $91.2\% - 96.8\%$ ). Sixteen samples were negative for NTRK fusions by the comparator method but positive by the TSO Comprehensive assay, leading to the current NPA value. Of these 16 samples, 6 were TSO positive due to a putative isolated contamination event that occurred during the study. These 6 samples all contained the same fusion but were in the same library preparation plate as a sample with an identical fusion with high supporting reads. This putative contamination was investigated during an additional Cross-contamination study to understand discordant results described in the paragraph below. Of the remaining 10 samples positive by TSO Comprehensive assay and negative by the comparator method, 7 samples were positive with the local test (LT) used in the original therapeutic clinical trial, and 3 samples were LT negative. Six of these 7 LT positive patients showed clinical trial benefit with larotrectinib.

Table 6. NTRK Accuracy Study: Concordance Between the TSO Comprehensive Assay and Comparator Method for Detection of NTRK Fusions

| | Comparator Method Result | | | |
| --- | --- | --- | --- | --- |
| | | NTRK Fusion Positive | NTRK Fusion Negative | Total |
| TSO Comprehensive Assay Result | NTRK Fusion Positive | 114 | 162 | 130 |
| | NTRK Fusion Negative | 4 | 273 | 277 |
| | Total | 118 | 289 | 407 |
| Agreement Statistics | PPA% (n/N; 95% CI1) | 96.6% (114/118; 91.5%-99.1%) | | |
| | NPA% (n/N; 95% CI1) | 94.5% (273/289; 91.2%-96.8%) | | |

$^{1}$ CI based on Clopper-Pearson (exact) method
$^{2}$ 6 of the 16 samples were TSO Comprehensive positive due to an isolated contamination event during study testing. Out of the remaining 10 samples, 7 were positive by the LT, and 6 of those 7 showed clinical benefit with the therapeutic

PMA P230011: FDA Summary of Safety and Effectiveness Data

{17}

Table 7. NTRK Accuracy Study: Concordance Between the TSO Comprehensive Assay and Comparator Method for Detection of NTRK Fusions by Gene

| | Orthogonal Positive | | | Orthogonal Negative | | | |
| --- | --- | --- | --- | --- | --- | --- | --- |
| | TSO Comp Positive | TSO Comp Negative | TSO Comp Positive | TSO Comp Negative | | | |
| Gene | CP | DN | DP | CN | Total | PPA (n/N), (95% CI)1 | NPA (n/N), (95% CI)1 |
| NTRK1 | 40 | 3 | 2 | 362 | 407 | 93.0% (40/43), (80.9%, 98.5%) | 99.5% (362/364), (98.0%, 99.0%) |
| NTRK2 | 7 | 0 | 4 | 396 | 407 | 100.0% (7/7), (59.0%, 100.0%) | 99.0% (396/400), (97.5%, 99.0%) |
| NTRK3 | 67 | 3 | 10 | 327 | 407 | 95.7% (67/70), (88.0%, 99.1%) | 97.0% (327/337), (94.6%, 98.6%) |
| Total | 114 | 4 | 16 | 273 | 407 | 96.6% (114/118), (91.5%-99.1%) | 94.5% (273/289), (91.2%-96.8%) |

Abbreviations: CN = Concordant Negative; CP = Concordant Positive; DN = Discordant Negative; DP = Discordant Positive
${}^{1}$ CI based on Clopper-Pearson (Exact) method.

Testing of discordant results was performed for samples in both the NTRK Accuracy and Clinical Bridging studies due to low NPA values in both studies, possibly due to a suspected contamination event. A total of 142 RNA samples, comprised of 107 concordant samples (53 concordant negative, 54 concordant positive), 26 discordant samples, and 9 samples with unknown NTRK status from any of the methods used (TSO Comprehensive, comparator method, and LT), were re-tested with the TSO Comprehensive assay. Of the 142 samples, 133 produced valid results. Of the 6 samples that were suspected to be contaminated, presented as positive in the original Accuracy study in Table 6 above, only 4 were available for re-testing. All 4 of those samples were NTRK fusion-negative upon re-testing. Additionally, 6 out of the remaining 126 samples with valid retest results changed NTRK fusion status upon retesting. Of those samples, 3 were concordant NTRK fusion-positives by original testing, changing to negative upon retest. None of the concordant NTRK fusion-negatives from the original testing changed to positives. Further, a sensitivity analysis was performed for the results from the testing of the previously discussed discordant samples. The PPA in the imputed dataset was estimated to be $93.3\%$ (95% CI: $87.9\%$ , $98.7\%$ ) and NPA was estimated at $97.3\%$ (95% CI: $95.4\%$ , $99.3\%$ ). The PPA and NPA from the imputed dataset are within the confidence limits of the original estimates (Tables 6 and 7).

# RET Fusion Detection in NSCLC Samples

The accuracy of the TSO Comprehensive assay for detecting RET fusions in patients with NSCLC was evaluated by assessing the concordance of RET fusion results between the TSO Comprehensive assay and a validated NGS-based comparator method as one of the studies to support the use of the TSO Comprehensive as a companion diagnostic (CDx)

PMA P230011: FDA Summary of Safety and Effectiveness Data

{18}

for selpercatinib. A retrospective, noninterventional study was conducted using a total of 219 FFPE samples consisting of 93 samples obtained from LIBRETTO-001 trial (see Section X. for study details) and 126 supplemental negative samples obtained from commercial sources. These samples were tested with the TSO Comprehensive assay at one external site and with the NGS-based a comparator method at a central laboratory. Accuracy of the TSO Comprehensive assay RET fusion calls by the TSO Comprehensive assay was estimated relative to the comparator method. Positive percent agreement (PPA), negative percent agreement (NPA), and the associated two-sided $95\%$ confidence intervals (CIs) were calculated.

Of the 219 FFPE samples tested, 14 samples had invalid TSO Comprehensive assay results (invalid rate of $6.4\%$ ). Of these 14 (13 supplemental and 1 clinical trial sample), 8 also had invalid comparator method results. The comparator method had 15 additional invalid results and 22 samples with no calls. In addition, there were 6 samples (all supplemental) with invalid TSO Comprehensive results and valid comparator method results. In total, 168 of the 219 samples had valid results by both methods (69 RET-fusion positive clinical trial specimens and 99 supplemental RET-fusion negative samples). The concordance analysis is shown in Table 8. Based on valid test results, the Positive Percent Agreement (PPA) and Negative Percent Agreement (NPA) between TSO Comprehensive assay and comparator method were $100.0\%$ (69/69, $95\%$ CI: $94.8\%$ to $100.0\%$ ), and $97.0\%$ (96/99, $95\%$ CI: $91.4\%$ to $99.4\%$ ), respectively.

For the discordant specimens called RET fusion-positive by TSO Comp and RET fusion negative by the orthogonal method, all were clinical trial samples that tested RET fusion-positive by the LT used for enrollment. One sample contained a fusion partner that could not be interrogated by the orthogonal assay, but clinical data showed that the patient had a verified partial response (PR) to therapy, thus suggesting the presence of a RET fusion. For the other two samples, the amount of supporting reads were low, and may have been below the variant calling threshold for the orthogonal assay. These subjects were unevaluable for response assessment.

Table 8. RET Accuracy Study: Concordance Between the TSO Comprehensive Assay and Comparator Method for Detection of RET Fusions in NSCLC

| | Comparator Method Result | | | |
| --- | --- | --- | --- | --- |
| | | RET Fusion Positive | RET Fusion Negative | Total |
| TSO Comprehensive Assay Result | RET Fusion Positive | 69 | 3 | 72 |
| | RET Fusion Negative | 0 | 96 | 96 |
| | Total | 69 | 99 | 168 |
| Agreement Statistics | PPA% (n/N; 95% CI1) | 100.0% (69/69; 94.8% -100.0%) | | |
| | NPA% (n/N; 95% CI1) | 97.0% (96/99; 91.4% - 99.4%) | | |

$^{1}$ CI based on Clopper-Pearson (exact) method

PMA P230011: FDA Summary of Safety and Effectiveness Data

{19}

ii. Tumor Profiling Accuracy

Detection of small DNA variants, TMB, RNA fusions, and RNA splice variants was evaluated with FFPE specimens (both characterized and uncharacterized) using appropriate comparator methods to assess the accuracy of the TSO Comprehensive assay.

Small DNA Variants

The detection of small DNA variants was assessed by comparison with results obtained using an externally validated NGS targeted panel assay (evNGS). A total of 508 unique samples were tested by the evNGS, of which 19 were invalidated due to insufficient quality or quantity of DNA samples (n=3) or libraries (n=16), leaving 489 unique samples with valid evNGS results. Those 508 samples were also tested with the TSO Comprehensive assay, of which there were 77 invalidated samples due to failing contamination QC (n=8), failing small DNA variant QC (n=50), and samples mistakenly tested by the evNGS after an invalid result (n=21), leaving 431 with valid TSO Comp results. Of these 431 samples, 6 were invalid by the evNGS, leaving 425 samples, of which 11 were excluded due to a sample swap, leaving 414 samples with valid results for both assays. This set of 414 FFPE samples (representing SNVs, MNVs, insertions, and deletions) from 16 different tissue types that had valid results for both TSO Comprehensive assay and evNGS were the evaluable sample set (Table 9).

Table 9. Number of Evaluable Samples (n=414) by Cancer/Tissue Type

| Tissue Type/Cancer Type | Number of DNA Samples included in Final Data Analysis |
| --- | --- |
| Bladder | 19 |
| Brain | 21 |
| Breast | 30 |
| CRC | 52 |
| Gastric | 24 |
| Head and Neck | 1 |
| Liver | 21 |
| Medullary Thyroid | 38 |
| NSCLC | 57 |
| Ovary | 16 |
| Pancreas | 24 |
| Papillary Thyroid | 26 |
| Prostate | 36 |
| Skin | 25 |

PMA P230011: FDA Summary of Safety and Effectiveness Data
20 of 176

{20}

The reported variants from TSO Comprehensive assay were compared with results of the evNGS method and used to calculate the PPA and NPA. A summary of overall PPA and NPA per clinical significance and variant type for small DNA variants are provided in Table 10 below. The observed PPA for all Level 2 variants was 95.5%, which includes an observed PPA of 96.1% for Level 2 SNVs. In addition, there was a lower PPA of 80.7% for all Level 3 variants. Differences in Level 2 and Level 3 variant calls were attributed to the differences in cutoff between the evNGS method and TSO Comprehensive, since the evNGS method does not report any variants below the 5% VAF filter. A number of Level 3 variants in the study were below the calling threshold of the evNGS assay and the lower PPA for Level 3 variants was attributed to detection of deamination artifacts by the comparator method. A limitation in the device labeling addresses lack of accuracy data for samples >5% VAF, since accuracy of small DNA tumor profiling variants below 5% variant allele frequency (VAF) has not been established.

Table 10. Concordance Summary for Small DNA Variants by Variant Type and Clinical Significance

| Variant Category | Clinical Significance^{4} | PPA (n/N) (95% CI^{1}) | NPA (n/N) (95% CI^{1}) |
| --- | --- | --- | --- |
| SNVs | Level 2 | 96.1% (99/103) (90.4%, 98.5%) | >99.9% (9832/9833) (>99.9%, >99.9%) |
| | Level 3 | 76.9% (373/485)^{3} (73.0%, 80.4%) | >99.9% (212277/212311) (>99.9%, >99.9%) |
| | All | 80.3% (472/588) (76.9%, 83.3%) | >99.9% (219211/219246) (>99.9%, >99.9%) |
| MNVs | Level 2 | 100% (5/5) (56.6%, 100.0%) | 100.0% (9931/9931) (>99.9%, 100.0%) |
| | Level 3 | 90.0% (9/10) (59.6%, 98.2%) | >99.9% (212785/212786) (>99.9%, >99.9%) |
| | All | 93.3% (14/15) (70.2%, 98.8%) | >99.9% (219818/219819) (>99.9%, >99.9%) |
| Insertions | Level 2 | 100.0% (1/1) (20.7%, 100.0%) | 100.0% (9935/9935) (>99.9%, 100.0%) |
| | Level 3 | 86.0% (49/57) (74.7%, 92.7%) | >99.9% (212738/212739) (>99.9%, >99.9%) |
| | All | 86.2% (50/58) (75,1%, 92.8%) | >99.9% (219775/219776) (>99.9%, >99.9%) |
| Deletions | Level 2 | 66.7% (2/3)^{2} (20.8% - 93.9%) | >99.9% (9932/9933) (>99.9%, >99.9%) |
| | Level 3 | 90.3% (139/154) (84.6% - 94.0%) | >99.9% (212624/212642) (>99.9%, >99.9%) |
| | All | 89.8% (141/157) (84.1%, 93.6%) | >99.9% (219658/219677) (>99.9%, >99.9%) |
| | Level 2 | 95.5% (107/112) (90.0% - 98.1%) | >99.9% (9822/9824) (99.9%, >99.9%) |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{21}

| All Variants | Level 3 | 80.7% (570/706) (77.7%, 83.5%) | >99.9% (212036/212090) (>99.9%, >99.9%) |
| --- | --- | --- | --- |
| | All | 82.8% (677/818) (80.0%, 85.2%) | >99.9% (218960/219016) (>99.9%, >99.9%) |

$^{1}95\%$ 2-sided confidence interval calculated via the Wilson score method.
$^{2}$ TruSight Oncology Comprehensive assay detected an EGFR clinically significant deletion but with 1 nucleotide difference in the alternate sequence relative to the clinically significant deletion found by the comparator method, which caused the discordant result.
3These were discordant negatives detected only by the comparator assay. Based on root cause analysis, $91\%$ (102/112) of them are due to deamination artifacts, which were either not present in the TruSight Oncology Comprehensive bam file $(n = 49)$ , or were filtered out as having low LQ score $(n = 53)$
4Level 2 = Cancer Mutations with Evidence of Clinical Significance, Level 3 = Cancer Mutations with Evidence of Potential Clinical Significance

Since the evNGS method does not report any variants below the $5\%$ VAF filter, concordance for variants above VAF $\geq 5\%$ has been also calculated, by variant category. When looking at results $\geq 5\%$ VAF, observed PPA was $98.1\%$ (95% CI: $93.4\% - 99.5\%$ ) for all Level 2 variants, which includes Level 2 SNVs with an observed PPA of $99.0\%$ (95% CI: $94.5\%$ , $99.8\%$ ), Level 2 MNVs with an observed PPA of $100.0\%$ (95% CI: $56.6\%$ , $100.0\%$ ), and Level 2 insertions with an observed PPA of $100.0\%$ (95% CI: $20.7\%$ , $100.0\%$ ). Further, an observed PPA of $97.5\%$ (95% CI: $95.8\% - 98.5\%$ ) was seen for all Level 3 variants, including an observed PPA of $98.0\%$ (95% CI: $96.0\% - 99.0\%$ ) for Level 3 SNVs, an observed PPA of $90.0\%$ for Level 3 MNVs (95% CI: $59.6\% - 98.2\%$ ), an observed PPA of $96.1\%$ for Level 3 insertions (95% CI: $86.8\% - 98.9$ ), and an observed PPA of $97.1\%$ (95% CI: $92.7\% - 98.9\%$ ) for Level 3 deletions.

A summary of accuracy (PPA and NPA) of insertions and deletions binned by size is presented in Table 11 below. Aggregate PPA for both variant types are lower (86.2% for insertions, 89.8% for deletions) due to differences in Level 2 and Level 3 variant calls that were expected based on the differences in cutoff between the evNGS method and TSO Comprehensive. Looking at insertions and deletions $\geq 5\%$ VAF, all variants had a PPA of $>95\%$ , with the lone exception of a single deletion in the 11-15 bp range; however, the concordance in the 16-20 bp range was $100\%$ . In addition, results from the concordance analysis across insertions and deletions of all sizes regardless of clinical validity showed detection of 24 bp insertion (1/1; 95% CI: 20.7%-100.0), as well 24 bp deletions (2/2; 95% CI: 34.2%-100.0%).

PMA P230011: FDA Summary of Safety and Effectiveness Data

{22}

Table 11. Concordance Summary of All Levelled (Level 2 and 3) Insertion and Deletion Variants by Size

| Variant Category | Indel Bin Size | Total Unique Variants | True Positives | False Positives | False Negatives | True Negatives | PPA (n/N) (95% CI^{1}) | NPA (n/N) (95% CI^{1}) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| INS | 1-5 | 46 | 47 | 1 | 7 | 219779 | 87.0% (47/54) (75.6%-93.6%) | >99.9% (219779/219780) (>99.9%->99.9%) |
| | 6-10 | 2 | 1 | 0 | 1 | 219832 | 50.0% (1/2) (9.5%-90.5%) | 100.0% (219832/219832) (>99.9%-100.0%) |
| | 11-15 | 2 | 2 | 0 | 0 | 219832 | 100.0% (2/2) (34.2%-100.0%) | 100.0% (219832/219832) (>99.9%-100.0%) |
| | All | 50 | 50 | 1 | 8 | 219775 | 86.2% (50/58) (75.1%-92.8%) | >99.9% (219775/219776) (>99.9%->99.9%) |
| DEL | 1-5 | 129 | 132 | 18 | 15 | 219669 | 89.8% (132/147) (83.8%-93.7%) | >99.9% (219669/219687) (>99.9%->99.9%) |
| | 6-10 | 5 | 5 | 0 | 0 | 219829 | 100.0% (5/5) (56.6%-100.0%) | 100.0% (219829/219829) (>99.9%-100.0%) |
| | 11-15 | 2 | 0 | 1 | 1 | 219832 | 0.0% (0/1) (0.0%-79.3%) | >99.9% (219832/219833) (>99.9%->99.9%) |
| | 16-20 | 3 | 4 | 0 | 0 | 219830 | 100.0% (4/4) (51.0%-100.0%) | 100.0% (219830/219830) (>99.9%->99.9%) |
| | All | 139 | 141 | 19 | 16 | 219658 | 89.8% (141/157) (84.1%-93.6%) | >99.9% (219658/219677) (>99.9%->99.9%) |

$^{1}95\%$ 2-sided confidence interval calculated via the Wilson score method.

Accuracy results (PPA and NPA) for Variants with Evidence of Clinical Significance (Level 2) by gene and variant are included in Table 12 below, and a summary of concordance of all small DNA variants by gene is included in Appendix 2.

PMA P230011: FDA Summary of Safety and Effectiveness Data
23 of 176

{23}

Table 12. Accuracy of Variants with Evidence of Clinical Significance (Level 2) by Gene and Variant

| Gene | Variant | Concordant Positive Calls | Called only in evNGS (Discordant Negative) | Called only in TSO Comp (Discordant Positive) | Concordant Negative Calls | PPA (%) (95% CI) | NPA (%) (95% CI) |
| --- | --- | --- | --- | --- | --- | --- | --- |
| BARD1 | Trp91Ter | 0 | 1 | 0 | 413 | 0.0 (0.0, 79.3) | 100.0 (99.1, 100.0) |
| BRAF | Val600Glu | 41 | 1 | 0 | 372 | 97.6 (87.7, 95.6) | 100.0 (99.0, 100.0) |
| BRIP1 | Arg162Ter | 0 | 1 | 0 | 413 | 0.0 (0.0, 79.3) | 100.0 (99.1, 100.0) |
| CDK12 | Leu453IlefsTer9 | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| CHEK2 | Thr367MetfsTer15 | 1 | 0 | 0 | 413 | 100.0 (20.7,100.0) | 100.0 (99.1, 100.0) |
| EGFR | Leu747_Ala750del | 0 | 0 | 1 | 413 | N/A | 99.8 (98.6, >99.9) |
| | Leu747_Thr751del insSer | 0 | 1 | 0 | 413 | 0.0 (0.0, 79.3) | 100.0 (99.1, 100.0) |
| | Leu858Arg | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| KRAS | Ala146Thr | 2 | 0 | 0 | 412 | 100.0 (34.2, 100.0) | 100.0 (99.1, 100.0) |
| | Gln61Leu | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| | Gly12Ala | 12 | 0 | 0 | 402 | 100.0 (75.8, 100.0) | 100.0 (99.1, 100.0) |
| | Gly12Cys | 18 | 0 | 1 | 395 | 100.0 (82.4, 100.0) | 99.7 (98.6, >99.9) |
| | Gly12Phe | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| | Gly12Ser | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| | Gly12Val | 7 | 0 | 0 | 407 | 100.0 (64.6, 100.0) | 100.0 (99.1, 100.0) |
| | Gly13Ala | 4 | 0 | 0 | 410 | 100.0 (51.0, 100.0) | 100.0 (99.1, 100.0) |
| NRAS | Gly12Ala | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{24}

| Gene | Variant | Concordant Positive Calls | Called only in evNGS (Discordant Negative) | Called only in TSO Comp (Discordant Positive) | Concordant Negative Calls | PPA (%) (95% CI) | NPA (%) (95% CI) |
| --- | --- | --- | --- | --- | --- | --- | --- |
| PALB2 | Gln472Ter | 0 | 1 | 0 | 413 | 0.0 (0.0, 79.3) | 100.0 (99.1, 100.0) |
| PTEN | Leu320Ter | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| RET | Ala883Phe | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| | Cys609Tyr | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| | Cys630Arg | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| | Cys634Arg | 2 | 0 | 0 | 412 | 100.0 (34.2,100.0) | 100.0 (99.1, 100.0) |
| | Cys634Phe | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |
| | Cys634Ser | 3 | 0 | 0 | 411 | 100.0 (43.9,100.0) | 100.0 (99.1, 100.0) |
| | Cys634Tyr | 2 | 0 | 0 | 412 | 100.0 (34.2, 100.0) | 100.0 (99.1, 100.0) |
| | Met918Thr | 3 | 0 | 0 | 411 | 100.0 (43.9, 100.0) | 100.0 (99.1, 100.0) |
| | Val804Met | 1 | 0 | 0 | 413 | 100.0 (20.7, 100.0) | 100.0 (99.1, 100.0) |

PMA P230011: FDA Summary of Safety and Effectiveness Data
25 of 176

{25}

# Wild Type Accuracy

For assessing accuracy of wild-type calls, evaluation of clinically relevant positions was performed by observation of the NPA of Level 2 small DNA variants. Results in Table 14 (by variant type) show an NPA of $>99.9\%$ for all relevant Level 2 variant types.

Table 14. Evaluation of Wild-Type Accuracy using NPA of Variants of Clinical Significance (Level 2)

| Category | NPA (n/N) | 95% CI (Wilson Score) | 95% CI (Exact Method) |
| --- | --- | --- | --- |
| SNVs | >99.9% (9832/9833) | >99.9, >99.9 | 99.9, >99.9 |
| MNVs | 100.0% (9931/9931) | >99.9, 100.0 | >99.9, 100.0 |
| Insertions | 100.0% (9935/9935) | >99.9, 100.0 | >99.9, 100.0 |
| Deletions | >99.9% (9932/9933) | >99.9, >99.9 | 99.9, >99.9 |

# Tumor Mutational Burden (TMB) Score

Concordance of TMB score was determined by comparing the TMB scores (somatic mutations/megabase) between a validated Whole Exome Sequencing (WES) comparator method and TSO Comprehensive assay. 246 tumor-normal paired samples were sent for reference testing by the WES, and 29 were excluded due to failure of QC metrics. For the remaining 217 pairs, 33 were determined to be mismatched, and 60 pairs were excluded due to contamination or insufficient coverage. The remaining 124 FFPE samples from 14 different tissue types across a range of TMB scores (0.8-115.7 Mut/Mb) with valid results by both TSO Comprehensive and WES comparator method were analyzed. Linear regression analysis with WES as the predictor and TSO Comprehensive as the outcome had a y-intercept of 2.53, a slope of 0.89, and Pearson's correlation coefficient of 0.94 (Figure 1).

PMA P230011: FDA Summary of Safety and Effectiveness Data

{26}

Figure 1. TMB Score Correlation between WES and TSO Comprehensive

# RNA Splice Variants

Accuracy for splice variant detection was calculated by comparing TSO Comprehensive assay results to a validated qPCR comparator method for EGFRvIII. Concordance analysis was performed on a total of 16 unique FFPE RNA samples from brain tissue. Table 15 summarizes the concordance study results for EGFRvIII. In summary, EGFR had a PPA and NPA of $100\%$ .

Table 15. Concordance Analysis Between TSO Comprehensive assay and qPCR Assay for EGFRvIII

| TSO Comprehensive Results | qPCR Positive | qPCR Negative |
| --- | --- | --- |
| Positive | 3 | 0 |
| Negative | 0 | 13 |
| PPA (n/N) (95% CI) | 100% (3/3) (44%, 100%) | |
| NPA (n/N) (95% CI) | 100% (13/13) (77%, 100%) | |

# RNA Fusions - Comparison to a Composite Method

TSO Comprehensive assay fusions were compared to a validated composite method consisting of an RNA whole exome NGS panel (weNGS), a targeted NGS fusion panel

PMA P230011: FDA Summary of Safety and Effectiveness Data

{27}

(tNGS), and droplet digital PCR (ddPCR). A breakdown of the composite methodology is shown in Table 16. In short, the NGS exome panel overlapped with all the genes for which TSO Comprehensive assay can detect fusions. However, because the limit of detection of the weNGS assay was $4 - 8\mathrm{x}$ that of TSO Comprehensive assay (based on the number of supporting reads observed in the overlapping fusion calls), a composite method using two additional methods with greater sensitivity, but less breadth, was used for fusions (NGS Targeted Panel and ddPCR).

Table 16. Composite Reference Method Testing for RNA Fusions

| | weNGS | | |
| --- | --- | --- | --- |
| | | Detected | Not Detected |
| TSO Comprehensive | Detected | Tested with tNGS if covered by assay; if not, tested with ddPCR | Tested with tNGS if covered by assay; if not, tested with ddPCR |
| | Not Detected | Tested with tNGS if covered by assay; if not, tested with ddPCR | Randomly selected samples tested with tNGS or ddPCR if fusions are not covered by tNGS |

A total of 255 unique RNA samples representing 14 tissue types were tested. All samples passed TSO Comprehensive metrics. Of the 255 samples, 220 samples were not selected based on prior screening (hereafter "uncharacterized"), while 35 samples were selected for the study because they were fusion positive with TSO Comprehensive assay or a predecessor assay (hereafter known as "characterized" samples). Performance for these 35 characterized samples was adjusted using a mean fusion prevalence. Composite concordance results for the characterized and uncharacterized samples (and their respective performance characteristics) are shown in Tables 17 and 18, respectively. The relatively low PPA for the characterized samples is reflective of the prevalence of the included fusions, as PPA and NPA were adjusted for fusion prevalence; therefore, the combined performance across characterized and uncharacterized samples was calculated using an inverse-variance weighted average to account for the adjustment for prevalence in the characterized samples. These concordance results are shown in Table 19. In summary, the PPA of the combined fusions was 87.29.

PMA P230011: FDA Summary of Safety and Effectiveness Data

{28}

Table 17. Summary of TSO Comp Versus Composite Method Results for RNA Fusions in 35 Characterized Samples)

| | Composite Method | | | |
| --- | --- | --- | --- | --- |
| | | Detected | Not Detected | Total |
| TSO Comprehensive Assay | Detected | 25 | 4 | 29 |
| | Not Detected | 2 | 809¹ | 811 |
| | Total | 27 | 813 | 840 |
| | PPA (95% CI)²,³ | 60.34% (29.73%, 100%) | | |
| | NPA (95% CI)²,³ | 99.94% (99.86%, 99.98%) | | |

¹809 calculated as 35 samples x 24 fusion outcomes/sample - 25 TP - 4 FP - 2 FN
²Confidence intervals calculated by bootstrap.
³Performance characteristics have been adjusted for prevalence

Table 18. Summary of TSO Comp Versus Composite Method Outcomes for RNA Fusions in 220 Uncharacterized Samples

| | Composite Method | | | |
| --- | --- | --- | --- | --- |
| | | Detected | Not Detected | Total |
| TSO Comprehensive Assay | Detected | 12 | 0 | 12 |
| | Not Detected | 1 | 5267¹ | 5268 |
| | Total | 13 | 5267 | 5280 |
| | PPA (95% CI)² | 92.31% 66.69% - 99.61%) | | |
| | NPA (95% CI)² | 100% (99.93% - 100%) | | |

¹5,267 calculated as 220 samples x 24 fusion outcomes/sample – 12 TP – 0 FP – 1 FN
²Confidence intervals calculated by the Wilson Method.

Table 19. Summary of Performance for RNA Fusions (Combined for Characterized and Uncharacterized)

| PPA (95% CI¹) | NPA (95% CI¹) |
| --- | --- |
| 87.29% (67.83%, 96.35%) | 99.99% (99.98%, >99.99%) |

¹Confidence intervals calculated by bootstrap.

A summary of the concordance data by characterization and by gene is also presented in Table 20 below, which shows PPA/NPA for all 24 fusions genes in the TSO Comprehensive panel. Please note that, for the PPA/NPA values, characterized samples were adjusted by prevalence, while uncharacterized samples were not.

PMA P230011: FDA Summary of Safety and Effectiveness Data
29 of 176

{29}

Table 20. Concordance for RNA Fusions by Gene and Sample Characterization

| Gene | # TP | # FP | # FN | # TN | Characterized | PPA (n/N) (95% CI*) | NPA (n/N) (95% CI*) | PPV (n/N) (95% CI*) | NPV (n/N) (95% CI*) | Adjusted PPA (%) | Adjusted NPA (%) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| AXL | 2 | 0 | 0 | 33 | YES | 100.0 (2/2) (34.2-100.0) | 100.0 (33/33) (89.6-100.0) | 100.0 (2/2) (34.2-100.0) | 100.0 (33/33) (89.6-100.0) | 100.0 | 100.0 |
| AXL | 1 | 0 | 0 | 219 | NO | 100.0 (1/1) (20.7-100.0) | 100.0 (219/219) (98.3-100.0) | 100.0 (1/1) (20.7-100.0) | 100.0 (219/219) (98.3-100.0) | - | - |
| BCL2 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1) (20.7-100.0) | 100.0 (34/34) (89.8->99.9) | 100.0 (1/1) (20.7-100.0) | 100.0 (34/34) (89.8->99.9) | 100.0 | 100.0 |
| BCL2 | 0 | 0 | 0 | 220 | NO | - | 100.0 (220/220) (98.3->99.9) | - | 100.0 (220/220) (98.3->99.9) | - | - |
| BRAF | 2 | 0 | 0 | 33 | YES | 100.0 (2/2) (34.2-100.0) | 100.0 (33/33) (89.6-100.0) | 100.0 (2/2) (34.2-100.0) | 100.0 (33/33) (89.6-100.0) | 100.0 | 100.0 |
| BRAF | 0 | 0 | 0 | 220 | NO | - | 100.0 (220/220) (98.3->99.9) | - | 100.0 (220/220) (98.3->99.9) | - | - |
| CDK4 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1) (20.7-100.0) | 100.0 (34/34) (89.8->99.9) | 100.0 (1/1) (20.7-100.0) | 100.0 (34/34) (89.8->99.9) | 100.0 | 100.0 |
| CDK4 | 0 | 0 | 0 | 220 | NO | - | 100.0 (220/220) (98.3->99.9) | - | 100.0 (220/220) (98.3->99.9) | - | - |
| EGFR | 1 | 3 | 0 | 31 | YES | 100.0 (1/1) (20.7-100.0) | 91.2 (31/34) (77.0-97.0) | 25.0 (1/4) (4.6-69.9) | 100.0 (31/31) (89.0-100.0) | 100.0 | 100.0 |
| EGFR | 0 | 0 | 0 | 220 | NO | - | 100.0 (220/220) (98.3->99.9) | - | 100.0 (220/220) (98.3->99.9) | - | - |
| EML4 | 1 | 0 | 1 | 33 | YES | 50.0 (1/2) (9.5-90.5) | 100.0 (33/33) (89.6-100.0) | 100.0 (1/1) (20.7-100.0) | 97.1 (33/34) (85.1-99.5) | 1.3 | 100.0 |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{30}

| Gene | #TP | #FP | #FN | #TN | Characterized | PPA(n/N)(95% CI*) | NPA(n/N)(95% CI*) | PPV(n/N)(95% CI*) | NPV(n/N)(95% CI*) | Adjusted PPA (%) | Adjusted NPA (%) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| EML4 | 1 | 0 | 0 | 219 | NO | 100.0 (1/1)(20.7-100.0) | 100.0(219/219)(98.3-100.0) | 100.0 (1/1)(20.7-100.0) | 100.0(219/219)(98.3-100.0) | - | - |
| ERG | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| ERG | 6 | 0 | 1 | 213 | NO | 85.7 (6/7)(48.7-97.4) | 100.0(213/213)(98.2->99.9) | 100.0 (6/6)(61.0-100.0) | 99.5(213/214)(97.4-99.9) | - | - |
| ESR1 | 2 | 0 | 0 | 33 | YES | 100.0 (2/2)(34.2-100.0) | 100.0(33/33)(89.6-100.0) | 100.0 (2/2)(34.2-100.0) | 100.0(33/33)(89.6-100.0) | 100.0 | 100.0 |
| ESR1 | 2 | 0 | 0 | 218 | NO | 100.0 (2/2)(34.2-100.0) | 100.0(218/218)(98.3-100.0) | 100.0 (2/2)(34.2-100.0) | 100.0(218/218)(98.3-100.0) | - | - |
| ETV1 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| ETV1 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| ETV4 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| ETV4 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| EWSR1 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| EWSR1 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{31}

| Gene | #TP | #FP | #FN | #TN | Characterized | PPA(n/N)(95% CI*) | NPA(n/N)(95% CI*) | PPV(n/N)(95% CI*) | NPV(n/N)(95% CI*) | Adjusted PPA (%) | Adjusted NPA (%) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| FGFR1 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| FGFR1 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| FGFR2 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| FGFR2 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| FGFR3 | 1 | 1 | 1 | 32 | YES | 50.0 (1/2)(9.5-90.5) | 97.0(32/33)(84.7-99.5) | 50.0 (1/2)(9.5-90.5) | 97.0 (32/33)(84.7-99.5) | 0.2 | 100.0 |
| FGFR3 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| FLI1 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| FLI1 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| KIF5B | 2 | 0 | 0 | 33 | YES | 100.0 (2/2)(34.2-100.0) | 100.0(33/33)(89.6-100.0) | 100.0 (2/2)(34.2-100.0) | 100.0(33/33)(89.6-100.0) | 100.0 | 100.0 |
| KIF5B | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| NRG1 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{32}

| Gene | #TP | #FP | #FN | #TN | Characterized | PPA(n/N)(95% CI*) | NPA(n/N)(95% CI*) | PPV(n/N)(95% CI*) | NPV(n/N)(95% CI*) | Adjusted PPA (%) | Adjusted NPA (%) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| NRG1 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| NTRK1 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| NTRK1 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| NTRK2 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| NTRK2 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| NTRK3 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| NTRK3 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| PAX3 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| PAX3 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |
| RAF1 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 (1/1)(20.7-100.0) | 100.0(34/34)(89.8->99.9) | 100.0 | 100.0 |
| RAF1 | 0 | 0 | 0 | 220 | NO | - | 100.0(220/220)(98.3->99.9) | - | 100.0(220/220)(98.3->99.9) | - | - |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{33}

| Gene | # TP | # FP | # FN | # TN | Characterized | PPA (n/N) (95% CI*) | NPA (n/N) (95% CI*) | PPV (n/N) (95% CI*) | NPV (n/N) (95% CI*) | Adjusted PPA (%) | Adjusted NPA (%) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| RET | 2 | 0 | 0 | 33 | YES | 100.0 (2/2) (34.2-100.0) | 100.0 (33/33) (89.6-100.0) | 100.0 (2/2) (34.2-100.0) | 100.0 (33/33) (89.6-100.0) | 100.0 | 100.0 |
| RET | 2 | 0 | 0 | 218 | NO | 100.0 (2/2) (34.2-100.0) | 100.0 (218/218) (98.3-100.0) | 100.0 (2/2) (34.2-100.0) | 100.0 (218/218) (98.3-100.0) | - | - |
| TMPR SS2 | 1 | 0 | 0 | 34 | YES | 100.0 (1/1) (20.7-100.0) | 100.0 (34/34) (89.8->99.9) | 100.0 (1/1) (20.7-100.0) | 100.0 (34/34) (89.8->99.9) | 100.0 | 100.0 |
| TMPR SS2 | 6 | 0 | 1 | 213 | NO | 85.7 (6/7) (48.7-97.4) | 100.0 (213/213) (98.2->99.9) | 100.0 (6/6) (61.0-100.0) | 99.5 (213/214) (97.4-99.9) | - | - |

¹Performance characteristics were adjusted for prevalence for characterized samples, and non-adjusted for uncharacterized samples.

## 2. Precision and Reproducibility

### a. Precision

Two studies were conducted to evaluate within-laboratory precision for the TSO Comprehensive assay. Study 1 evaluated CDx variants (NTRK and RET fusions). Study 2 evaluated TMB.

### CDx Precision

Within-laboratory precision was evaluated for NTRK1, NTRK2, and NTRK3 fusions in 6 FFPE samples from 5 tumor types (lower grade glioma, glioblastoma multiforme, colorectal cancer, myofibroblastic sarcoma, secretory breast cancer) and RET fusions in 4 samples from 3 tumor types (non-small cell lung cancer, thyroid cancer, and atypical Spitz tumor from skin tissue specimen). Each sample was tested at two variant levels: ~1x LoD (low variant level) and ~2–3x LoD (high variant level) with the exception of the sample harboring CCDC6-RET, which was only tested at the low variant level. Each of the samples at each test level was run in duplicates in each library preparation event across three (3) operators. Each operator started library preparation on three (3) non-consecutive start days and sequenced on three (3) designated NextSeq 550Dx instruments. Three (3) reagents lots were tested, generating 54 observations per sample per level.

Variant calling was evaluated separately for the two variant levels for a given variant from pooled observations across all variables (operators, reagent lots, instruments, days, and replicates). The percent positive calls (PPC) and percent negative calls (PNC) for NTRK fusions and associated two-sided 95% confidence interval (Wilson score) are summarized

PMA P230011: FDA Summary of Safety and Effectiveness Data

{34}

in Table 21 (overall), and the PPC confidence interval (Wilson score) by operator and by lot are shown in Table 22 and Table 23, respectively. The percent positive calls (PPC) and percent negative calls (PNC) for RET fusions and associated two-sided $95\%$ confidence interval (Wilson score) are summarized in Table 24 (overall), and the PPC confidence interval (Wilson score) by operator and by lot are shown in Table 25 and Table 26, respectively. It should be noted that some levels had fewer than 54 observations; these were due to invalid libraries that did not meet the necessary QC metrics (4 out of 1096 libraries failed, $0.4\%$ invalid rate).

For NTRK fusions, at the higher variant level ( $\sim 2 - 3\mathrm{x}$ LoD), the TSO Comprehensive assay demonstrated $100\%$ PPC for all fusions except one (LMNA-NTRK1 had $99.0\%$ PPC). It should be noted that LMNA-NTRK1 at the high level had more than 54 observations; this was due to the preparation of two dilution mixtures targeting to the same variant level. Therefore, the observations were combined. At the low variant level ( $\sim 1\mathrm{x}$ LoD), the PPC for RNA fusions ranged from $94.44\%$ to $100\%$ . For variants with PPC $< 95\%$ (BCAN-NTRK1), the supporting reads were below the respective Limits of Detection (53.2 supporting reads for BCAN-NTRK1). $100\%$ PNC was achieved for all variants at both levels.

Table 21. Overall Qualitative Results for NTRK Fusions

| Variant Level | Fusion | Mean Supporting Reads | PPC (95% CI) | PNC (95% CI) |
| --- | --- | --- | --- | --- |
| ~1x LoD | TPM3-NTRK1 | 20.2 | 100.0 (54/54) (93.4, 100.0) | 100.0 (537/537) (99.3, 100.0) |
| | BCAN-NTRK1 | 22.1 | 94.4 (51/54) (84.9, 98.1) | 100.0 (591/591) (99.4, 100.0) |
| | LMNA-NTRK1 | 12.2 | 98.1 (51/52) (89.9, 99.7) | 100.0 (539/539) (99.3, 100.0) |
| | ETV6-NTRK2 | 20.3 | 100.0 (54/54) (93.4, 100.0) | 100.0 (591/591) (99.4, 100.0) |
| | ETV6-NTRK3 | 16.2 | 100.0 (54/54) (93.4, 100.0) | 100.0 (537/537) (99.3, 100.0) |
| | ETV6-NTRK3 (FFPE cell line) | 23.1 | 98.1 (53/54) (90.2, 99.7) | |
| | KANK1-NTRK3 | 13.5 | 100.0 (54/54) (93.4, 100.0) | 100.0 (591/591) (99.4,100.0) |
| ~2-3x LoD | TPM3-NTRK1 | 57.1 | 100.0 (54/54) (93.4, 100.0) | 100.0 (481/481) (99.2, 100.0) |
| | BCAN-NTRK1 | 53.2 | 100.0 (54/54) (93.4, 100.0) | 100.0 (535/535) (99.3, 100.0) |
| | LMNA-NTRK1 | 35.1 | 99.0 (103/104) (94.8, 99.8) | 100.0 (431/431) (99.1, 100.0) |

PMA P230011: FDA Summary of Safety and Effectiveness Data

{35}

PMA…