ONCO/Reveal Dx Lung & Colon Cancer Assay (O/RDx-LCCA)

P200011 · Pillar Biosciences · PQP · Jul 30, 2021 · Medical Genetics

Device Facts

• Record ID: P200011
• Device Name: ONCO/Reveal Dx Lung & Colon Cancer Assay (O/RDx-LCCA)
• Applicant: Pillar Biosciences
• Product Code: PQP · Medical Genetics
• Decision Date: Jul 30, 2021
• Decision: APPR
• Device Class: Class 3
• Attributes: Software as a Medical Device

Intended Use

The ONCO/Reveal™ Dx Lung and Colon Cancer Assay (O/RDx-LCCA) is a qualitative next generation sequencing based in vitro diagnostic test that uses amplicon-based target enrichment technology for detection of single nucleotide variants (SNVs) and deletions in 2 genes from DNA isolated from formalin-fixed paraffin-embedded (FFPE) non-small cell lung cancer (NSCLC) and colorectal cancer (CRC) tumor tissue specimens. The test is intended as a companion diagnostic to identify patients with NSCLC or CRC who may benefit from treatment with the targeted therapies listed in Table 1 in accordance with the approved therapeutic product labeling. The O/RDx-LCCA is intended to be used on the Illumina MiSeqDx® instrument.

Device Story

O/RDx-LCCA is an amplicon-based NGS companion diagnostic for FFPE tumor tissue. Input: DNA extracted from FFPE NSCLC or CRC specimens. Process: DNA deparaffinization; protease digestion; Uracil-DNA Glycosylase treatment; PCR amplification with gene-specific primers; indexing PCR for barcoding; purification; normalization; pooling. Sequencing: Illumina MiSeqDx instrument. Analysis: PiVAT® software (standalone workstation, no network) performs secondary analysis, alignment to hg19, variant calling, and report generation. Output: PDF report identifying specific EGFR or KRAS variants. Used in clinical laboratories by trained personnel. Results inform physician decisions regarding targeted therapy (EGFR TKIs for NSCLC; cetuximab/panitumumab for CRC). Benefits: enables personalized treatment selection for patients with specific actionable mutations.

Clinical Evidence

Clinical concordance studies compared O/RDx-LCCA (FCD) to FDA-approved CDx tests (CCD). NSCLC arm (n=257) compared to cobas® EGFR Mutation Test v2; CRC arm (n=219) compared to therascreen® KRAS RGQ PCR. Non-inferiority statistical approach (Li, 2016) used. Results supported non-inferiority with upper bounds of 95% CIs for agreement margins ≤4% (NSCLC) and ≤5% (CRC). Analytical studies established LoD, accuracy, precision, and stability.

Technological Characteristics

Amplicon-based NGS target enrichment. Materials: 7 reagents (PCR master mixes, oligo pools, indexing primers, purification beads). Energy: Electrical (MiSeqDx, thermal cycler). Connectivity: Standalone workstation (PiVAT®), no network. Software: PiVAT® bioinformatics pipeline. Sterilization: N/A (reagents). Form factor: Kit for 48 reactions.

Indications for Use

Indicated for patients with NSCLC or CRC to identify eligibility for targeted therapies. NSCLC patients: EGFR Exon 19 deletions and Exon 21 L858R substitution mutations. CRC patients: KRAS wild-type (absence of mutations in codons 12 and 13). Not for standalone diagnostic, screening, monitoring, risk assessment, or prognosis.

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

• Praxis™ Extended RAS Panel (Illumina, Inc.)
• cobas® KRAS Mutation Test (P140023) (Roche Molecular Systems, Inc.)
• Therascreen® KRAS RGQ PCR Kit (P110030; P110027) (QIAGEN Manchester Ltd.)
• FoundationOne® CDx (P170019) (Foundation Medicine, Inc.)
• Therascreen® EGFR RGQ PCR Kit (P120022/S018) (QIAGEN Manchester Ltd.)
• cobas® EGFR Mutation Test v2 (P120019/S019; P120019/S031) (Roche Molecular Systems, Inc.)
• Oncomine™ Dx Target Test (P160045; P160045/S019) (Life Technologies Corp.)

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: ONCO/Reveal™ Dx Lung and Colon Cancer Assay (O/RDx-LCCA)

Device Procode: PQP

Applicant's Name and Address: Pillar Biosciences, Inc. 9 Strathmore Road Natick, MA 01760

Date(s) of Panel Recommendation: None

Premarket Approval Application (PMA) Number: P200011

Date of FDA Notice of Approval: July 30, 2021

## II. INDICATIONS FOR USE

The ONCO/Reveal™ Dx Lung and Colon Cancer Assay (O/RDx-LCCA) is a qualitative next generation sequencing based in vitro diagnostic test that uses amplicon-based target enrichment technology for detection of single nucleotide variants (SNVs) and deletions in 2 genes from DNA isolated from formalin-fixed paraffin-embedded (FFPE) non-small cell lung cancer (NSCLC) and colorectal cancer (CRC) tumor tissue specimens. The test is intended as a companion diagnostic to identify patients with NSCLC or CRC who may benefit from treatment with the targeted therapies listed in Table 1 in accordance with the approved therapeutic product labeling. The O/RDx-LCCA is intended to be used on the Illumina MiSeqDx® instrument.

Table 1. List of somatic variants for therapeutic use

| Indication | Gene | Variant | Targeted therapy |
| --- | --- | --- | --- |
| Colorectal Cancer (CRC) | KRAS | KRAS wild-type (absence of mutations in codons 12 and 13) | ERBITUX® (cetuximab), or VECTIBIX® (panitumumab) |
| Non-Small Cell Lung Cancer | EGFR | Exon 19 Deletions and Exon 21 L858R Substitution Mutations | EGFR Tyrosine Kinase Inhibitors approved by FDA* |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{1}

| Indication | Gene | Variant | Targeted therapy |
| --- | --- | --- | --- |
| (NSCLC) | | | |

*For the most current information about the therapeutic products in this group, go to: https://www.fda.gov/medicaldevices/productsandmedicalprocedures/invitrodiagnostics/ucm301431.htm

## III. CONTRAINDICATIONS

The test is not indicated to be used for standalone diagnostic purposes, screening, monitoring, risk assessment, or prognosis.

## IV. WARNINGS AND PRECAUTIONS

The warnings/precautions and limitations can be found in the O/RDx-LCCA labeling.

## V. DEVICE DESCRIPTION

### Test Output

The output of the test includes:

Category 1: Companion Diagnostic (CDx) Claims noted in Table 1 of the Intended Use

### Test Kit Contents

The Assay kit is composed of 7 reagents to allow the processing of 48 reactions (46 patient samples and required controls). The reagents are liquids or suspensions, stored in individual vials, and segregated into four labeled sub-containers. The sub-containers allow kit components to be stored at the recommended temperature, which may be room temperature, 4°C, or -20°C, depending on the components. The kits are labelled according to applicable standards and Instructions for Use. Safety Data Sheets are available from Pillar Biosciences. Outer packaging supports international frozen shipment.

Table 2. Assay kit reagents

| Kit Box 1: GS-PCR Reagent | Quantity | Storage |
| --- | --- | --- |
| Gene Specific PCR Master Mix | 1 tube (red cap) | -15°C to -25°C |
| LC Oligo Pool | 1 tube (yellow cap) | -15°C to -25°C |
| Positive Control (PosCtrl) | 1 tube (clear cap) | -15°C to -25°C |
| Uracil-DNA glycosylase (UDG) | 1 tube (blue cap) | -15°C to -25°C |
| Kit Box 2: Indexing PCR Reagent | Quantity | Storage |
| Indexing PCR Master Mix | 1 tube (green cap) | -15° to -25°C |
| Forward indexing primers (A501-A508) | 8 tubes (white caps) | -15° to -25°C |
| Reverse indexing primers (A701-A706) | 6 tubes (orange caps) | -15° to -25°C |
| Kit Box 3: PCR Product Purification Reagent | Quantity | Storage |
| Purification Beads | 1 bottle | 2°C to 8°C |
| Kit Box 4: Index Tube Caps | Quantity | Storage |
| White caps (for A501-A508 primers) | 24 caps | Ambient |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{2}

| Kit Box 1: GS-PCR Reagent | Quantity | Storage |
| --- | --- | --- |
| Orange caps (for A701-A706 primers) | 18 caps | Ambient |

Other general lab supplies that are needed to execute the protocol include laboratory gloves, ice, ice buckets, tube racks, etc. For reagents, consumables, and equipment required in both pre- and post-PCR processes, dedicated supplies (including gloves, lab coats, etc.) should be available in both areas.

Materials and equipments and software that are required for the test but are not provided with the assay kit are listed in Tables 3 and 4.

Table 3. Materials required but not provided in the assay kit

| Material | Source/Part Number |
| --- | --- |
| Qubit™ dsDNA HS Assay Kit | Thermo Fisher Scientific/Q32851 or Q32854 |
| Qubit™ Assay Tubes | Thermo Fisher Scientific/Q32856 |
| PhiX Control v3, 10 nM | Illumina/FC-110-3001 |
| MiSeqDx® Reagent Kit v3 (600 cycles) | Illumina/20012552 |
| Ethanol, 200 proof for molecular biology | General lab supplier |
| Nuclease-free water | General lab supplier |
| 10 mM Tris-HCl w/ 0.1% Tween-20, pH 8.5 | General lab supplier |
| 10 N NaOH or 1 N NaOH | General lab supplier |
| 1.5 mL microcentrifuge tubes | General lab supplier |
| 96-well PCR plates, 0.2 mL | General lab supplier |
| Microplate sealing film | General lab supplier |
| Conical tubes, 15 mL | General lab supplier |
| Conical tubes, 50 mL | General lab supplier |
| Aerosol filter pipette tips | General lab supplier |
| Solution basin (trough or reservoir) | General lab supplier |

Table 4. Equipment and software required but not provided

| Equipment | Source/Part Number |
| --- | --- |
| MiSeqDx® Instrument† | Illumina/DX-410-1001 |
| Pillar LC-HS module, v1.1.0.316 | Illumina/SW-0001 |
| ONCO/Reveal™ Dx Lung and Colon Cancer Assay™ PiVAT® Software | Pillar Biosciences/SFW-2002 |
| Qubit Fluorometer† | Thermo Fisher Scientific |
| Vortexer | General lab supplier |
| Magnetic stand for 96 wells | Life Technologies/12331D or 12027, Beckman Coulter/A32782 or equivalent |
| Microfuge | General lab supplier |
| Thermal cycler† with heated lid capability | General lab supplier |
| Single- and multi-channel pipettes†, 0.5 to 1000 μl | General lab supplier |
| Centrifuge adapted for PCR plates | General lab supplier |

† Equipment should be maintained and/or calibrated according to the manufacturer's instructions

PMA P200011: FDA Summary of Safety and Effectiveness Data

{3}

The PiVAT® software is for use with O/RDx-LCCA. PiVAT® performs secondary analysis and report generation from sequencing runs that use the O/RDx-LCCA.

PiVAT® is installed on a stand-alone computer system configured with an Ubuntu operating system and a Chromium browser. The system is configured with no network connectivity.

## Test Process

All assay reagents included in the O/RDx-LCCA are qualified by Pillar Biosciences and are compliant with the medical device Quality System Regulation (QSR).

## 1. Library Preparation

The kit is used to prepare sample libraries from FFPE specimens for NGS analysis that may help to inform treatment regimens through the targeted amplification of regions of the genome often mutated during tumor formation. Specimens must be deparaffinized and digested with protease to liberate the DNA target before purification. After purification, the DNA samples are treated with Uracil-DNA Glycosylase to render formalin damaged DNA non-amplifiable.

The samples are then amplified using polymerase chain reaction (PCR) and a gene-specific primer pool to enrich the number of assay target sequences.

The amplification products are purified from remaining primers and each sample is "barcoded" in a second PCR reaction using a unique pair of indexing primers to prepare the samples for pooling and multiplexed analysis.

After purification of the amplification products from residual indexing primers, the indexed libraries are quantified, normalized, and pooled for sequencing.

## 2. Sample Sequencing

Sequence information is extracted from the sample library pool using Illumina's MiSeqDx® NGS analyzer and corresponding reagents according to the manufacturer's protocols. The user selects the appropriate analysis panel type (O/RDx-LCCA) from the Pillar Module on the MiSeqDx® to initiate the sequence analysis utilizing the preset parameters for that panel. The Pillar Module is an interface designed for the MiSeqDx® by Illumina to interface with collaborator assays. Its main function is to automatically configure assay-specific sequencing parameters on the instrument, such as read length, when a particular assay is selected from the Module's drop-down menu. As additional assays are developed, they can be added to extend the capabilities of the Module.

The user must also create a Sample Sheet in Illumina's format that holds information about the samples such as name, whether the sample is a control sample (positive, negative, or no template), and what indices were used to tag that sample. Pillar Biosciences provides a tool that facilitates the aggregation of batches of libraries

PMA P200011: FDA Summary of Safety and Effectiveness Data
4 of 45

{4}

prepared across multiple days onto a single $\mathrm{MiSeqDx}^{\text{®}}$ v3 flow cell, provided that each batch has the required positive and No Template Control (NTC) controls. Up to 48 libraries may be multiplexed onto a single $\mathrm{MiSeqDx}^{\text{®}}$ v3 flow cell. A successful sequencing run will produce sequence data in Illumina's proprietary Binary Base Call format that is converted to a more universal FASTQ sequence format by the Pillar Module.

The O/RDxLCCA requires the user to run positive control (PosCtrl) and NTC for each "Batch" of up to 46 samples (processed on the same plate). Up to 6 batches may be included in a single sequencing run and analyzed through the PiVAT® software. PosCtrl is a cell line DNA containing the CDx variants with expected variant allele frequencies as shown in Table 5 below. The PosCtrl must generate expected mutations to be valid. If the PosCtrl is invalid, the PiVAT® software will fail the entire batch and no results will be reported for all samples within the batch.

Table 5. Positive Control (PosCtrl)

| Gene | Variant | Expected Allelic Frequency, % |
| --- | --- | --- |
| EGFR | ΔE746 - A750 | 2.00% |
| EGFR | L858R | 3.00% |
| KRAS | G13D | 15.00% |
| KRAS | G12D | 6.00% |

NTC is a reaction setup using DNA diluent or nuclease-free water with no template/DNA input. The NTC should not detect any mutations. If the NTC is invalid, the PiVAT® software will fail the entire batch and no results will be reported for all samples within the batch.

# 3. Bioinformatic Analysis

The $\mathrm{PiVAT}^{\text{®}}$ IVD bioinformatics pipeline is used to convert the raw FASTQ output from the $\mathrm{MiSeqDx}^{\text{®}}$ into genetic variation observed for each sample. The $\mathrm{PiVAT}^{\text{®}}$ IVD software is provided on a standalone workstation and raw sequence data are transferred from the $\mathrm{MiSeqDx}^{\text{®}}$ to the $\mathrm{PiVAT}^{\text{®}}$ IVD workstation for analysis using a USB drive. Once the required FASTQ files are uploaded to the $\mathrm{PiVAT}^{\text{®}}$ IVD workstation, the user can select sequence data to be analyzed from the browser-based $\mathrm{PiVAT}^{\text{®}}$ interface and begin the analysis workflow. The software will deconvolute the mixed sample sequence data using the unique sample index and consolidate matching forward and reverse sequence reads. Filters are used to minimize the result of random variation introduced during sample amplification steps and sequencing. The resulting sequences are aligned to the hg19 human genome sequence framework using the BWA-MEM aligner. Local re-alignments are performed to identify longer insertions and deletions (indels). Filters are applied to isolate likely variation from sequencing noise (Table 6). O/RDx-LCCA is designed to detect and report somatic variants for EGFR Exon 19 deletion, EGFR L858R mutation and missense mutations in KRAS G12 and KRAS G13. Non-targeted variants including germline variants are not reported.

PMA P200011: FDA Summary of Safety and Effectiveness Data

{5}

After this step, the remaining variants are annotated using HGVS standards, and a PDF format summary report is generated in two parts.

- The Run Summary section of the PiVAT® Customer PDF Report output file contains various statistics that reflect run quality and an overview of all variation to be reported across all samples analyzed within that run. Samples that fail to meet certain NGS quality criteria are reported as not valid and no genetic variants are reported for these samples.
- The Patient Report section of the PiVAT® Patient Report output file aggregates any variation observed at the patient level along with previously established clinical utility. Samples that fail to meet certain NGS quality thresholds are reported as not valid and no genetic variants are reported for these samples. The user can interact with the PiVAT® IVD pipeline utilizing a browser-based visual interface. At the end of analysis, all intermediate data files and reports may be downloaded to a USB drive and transferred to another location for permanent storage.

Table 6. NGS-QC in PiVAT®: Run, Sample and Variant Calling Passing Criteria

| Category | QC Metrics | Passing Criteria |
| --- | --- | --- |
| Run – FAIL if any QC metric(s) fails | PosCtrl | Expected mutations are detected |
| | PosCtrl | No unexpected mutation(s) detected |
| | NTC | No mutation detected |
| | NTC | Maximum coverage < 50x or < 0.5% of median within-run sample coverage |
| Sample – NOT VALID if any QC metric(s) fail | Sequencing base quality | Bases (with Q Score ≥ Q30) ≥ 75% |
| | Amplification specificity | Effective On-Target Rate2 ≥ 70% |
| | Coverage1 | Minimum of the amplicon depths3 ≥ 1000x |
| Variant reporting threshold | CDx mutations: non-C>T|G>A | Variant coverage >10x and Total coverage ≥ 1000x and Average variant base Q-score ≥ 30 and VAF ≥ 1% |
| | CDx mutations: C>T|G>A | Variant coverage >10x and Total coverage ≥1000x and Average variant base Q-score ≥ 30 and VAF ≥ 1.5% |

1 Coverage: the coverage after paired-end assembly by PiVAT®. All CDx markers in the Pillar assay are bidirectional sequenced with 2x150bp sequencing protocol due to the short amplicon sizes (144-162bp including primers). 1x coverage = 1x forward + 1x reverse of sequencing reads. Only uniquely mapped reads are analyzed.
2 Effective On-Target Rate = Mapping rate * On-target rate
3 Include the 7-key amplicons: three amplicons covering the CDx markers and other four amplicons covering EGFR-exon18, KRAS-exon3, KRAS-exon4 and BRAF-V600E.

Abbreviations: VAF = Variant allele frequency

# VI. ALTERNATIVE PRACTICES AND PROCEDURES

There are several FDA-approved CDx alternatives for the detection of genetic alterations using FFPE tumor specimens, as listed in the O/RDx-LCCA intended use statement. The

PMA P200011: FDA Summary of Safety and Effectiveness Data

{6}

approved CDx tests are listed in Table 7 below; for additional details see FDA List of Cleared or Approved CDx Devices at https://www.fda.gov/medical-devices/vitro-diagnostics/list-cleared-or-approved-companion-diagnostic-devices-vitro-and-imaging-tools. Each alternative has its own advantages and disadvantages. A patient should fully discuss these alternatives with his/her physician to select the method that best meets expectations and lifestyle.

Table 7. FDA-approved companion diagnostic (CDx) alternatives to O/RDx-LCCA

| Indication | Gene | Device | Company | Technology | Therapy |
| --- | --- | --- | --- | --- | --- |
| CRC | KRAS | Praxis™ Extended RAS Panel | Illumina, Inc. | NGS | VECTIBIX® (panitumumab) |
| CRC | KRAS | cobas® KRAS Mutation Test (P140023) | Roche Molecular Systems, Inc. | PCR | ERBITUX® (cetuximab) VECTIBIX (panitumumab) |
| CRC | KRAS | Therascreen® KRAS RGQ PCR Kit (P110030; P110027) | QIAGEN Manchester Ltd. | PCR | ERBITUX® (cetuximab) VECTIBIX (panitumumab) |
| CRC | KRAS | FoundationOne® CDx (P170019) | Foundation Medicine, Inc. | NGS | ERBITUX® (cetuximab) VECTIBIX® (panitumumab) |
| NSCLC | EGFR | Therascreen® EGFR RGQ PCR Kit (P120022/S018) | QIAGEN Manchester Ltd. | PCR | IRESSA® (gefitinib) GILOTRIF® (afatinib) VIZIMPRO® (dacomitinib) |
| NSCLC | EGFR | cobas® EGFR Mutation Test v2 (P120019/S019; P120019/S031) | Roche Molecular Systems, Inc. | PCR | EGFR Tyrosine Kinase Inhibitors approved by FDA |
| NSCLC | EGFR | FoundationOne® CDx (P170019) | Foundation Medicine, Inc. | NGS | GILOTRIF® (afatinib) IRESSA® (gefitinib) TARCEVA® (erlotinib) |

PMA P200011: FDA Summary of Safety and Effectiveness Data
7 of 45

{7}

| | | | | | TAGRISSO®
(osimertinib) |
| --- | --- | --- | --- | --- | --- |
| NSCLC | EGFR | Oncomine™ Dx
Target Test (P160045;
P160045/S019) | Life
Technologies
Corp. | NGS | IRESSA®
(gefitinib) |

Abbreviations: NSCLC = Non-Small Cell Lung Cancer, CRC=Colorectal Cancer

## VII. MARKETING HISTORY

The O/RDx-LCCA has not been marketed previously in the United States or any foreign country. It has so far been used only in Research Use Only format for research applications.

## 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 test results, and subsequently, inappropriate patient management decisions. Patients with false positive results may undergo treatment with one of the therapies listed in the above intended use statement without clinical benefit and may experience adverse reactions associated with the therapy. Patients with false negative results may not be considered for treatment with the indicated therapy. There is also a risk of delayed results, which may lead to delay of treatment with the indicated therapy.

For the specific adverse events related to the approved therapeutics, please see approved drug product labels.

## IX. SUMMARY OF NONCLINICAL STUDIES

### A. Laboratory Studies

Performance characteristics of the O/RDx-LCCA were established using DNA derived from FFPE NSCLC and CRC tumor tissue specimens. Studies included reportable CDx SNV and deletion variant representatives of EGFR and KRAS genes indicated in Table 1 of the intended use statement. The contrived samples (cell line, reference cell line standards in genomic DNA, FFPE and formalin-compromised DNA formats) were used only in cross-contamination study and reagent kit stability studies.

### 1. Analytical Accuracy

Analytical accuracy was performed to demonstrate the concordance between the O/RDx-LCCA and an externally validated comparator method for the ability of O/RDx-LCCA to detect reportable SNVs and short and medium deletions for CDx genes EGFR and KRAS.

A total of 263 samples (177 CRC and 86 NSCLC) were tested. Of these samples, 6 yielded invalid results with the externally validated NGS (evNGS) comparator method and 6 yielded invalid results or did not meet workflow quality control (QC) with O/RDx-LCCA. Among the 251 valid samples, 87 positive and 160

PMA P200011: FDA Summary of Safety and Effectiveness Data

{8}

negative samples were concordant between the two assays. There were 4 discordant samples between the O/RDx-LCCA and the comparator assay. The samples included simple SNVs, complex SNV and deletions that are targeted by the O/RDx-LCCA.

The aggregated results at the variant, sample and bin levels are shown in Table 8 below.

Table 8. Overall agreement results by variant, sample, and bin type

| Binned by Gene | Test + Comp + TP | Test + Comp - FP | Test - Comp + FN | Test - Comp - TN | Total N | PPA (95%CI) | NPA (95%CI) | PPV (95%CI) | NPV (95%CI) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| EGFR variant in NSCLC | 18 | 2 | 0 | 7675 | 7695 | 100.0% (82.4%, 100.0%) | 100.0% (99.9%, 100.0%) | 90.0% (69.9%, 97.2%) | 100.0% (100%, 100.0%) |
| KRAS variant in CRC | 69 | 2 | 0 | 5199 | 5270 | 100.0% (94.7%, 100.0%) | 100.0% (99.9%, 100.0%) | 97.2% (90.3%, 99.2%) | 100.0% (99.9%, 100.0%) |
| Binned by Sample | Test + Comp + TP | Test + Comp - FP | Test - Comp + FN | Test - Comp - TN | Total n | PPA (95%CI) | NPA (95%CI) | PPV (95%CI) | NPV (95%CI) |
| Sample (EGFR) | 18 | 2 | 0 | 61 | 81 | 100.0% (82.4%, 100.0%) | 96.8% (89.1%, 99.1%) | 90.0% (69.9%, 97.2%) | 100.0% (94.1%, 100.0%) |
| Sample (KRAS) | 69 | 2 | 0 | 99 | 170 | 100.0% (94.7%, 100.0%) | 98.0% (93.1%, 99.5%) | 97.2% (90.3%, 99.2%) | 100.0% (96.3%, 100.0%) |

Note: O/RDx-LCCA does not have "No Call" in "Valid" samples. Invalid data are excluded from this analysis.
$\mathrm{PPA} = \mathrm{TP} / (\mathrm{TP} + \mathrm{FN}) \times 100\%$
$\mathrm{NPA} = \mathrm{TN} / (\mathrm{TN} + \mathrm{FP}) \times 100\%$
$\mathrm{PPV} = \mathrm{TP} / (\mathrm{TP} + \mathrm{FP}) \times 100\%$
$\mathrm{NPV} = \mathrm{TN} / (\mathrm{TN} + \mathrm{FN}) \times 100\%$
Comp+ = Sample positive for at least one targeted variant when tested with comparator method
Test+ = Sample positive for at least one target variant when tested with O/RDx-LCCA

Agreement of $EGFR$ variants in NSCLC and $KRAS$ variants in CRC are summarized Tables 9 and 10 below, respectively.

Table 9. EGFR variants (in NSCLC) agreement by variant type and class

| Binned by Variant Type | Test + Comp + TP | Test + Comp - FP | Test - Comp + FN | Test - Comp - TN | Total n | PPA (95%CI) | NPA (95%CI) | PPV (95%CI) | NPV (95%CI) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| SNV | 8 | 0 | 0 | 73 | 81 | 100.0% (67.6%, 100.0%) | 100.0% (95.0%, 100.0%) | 100.0% (67.6%, 100.0%) | 100.0% (95.0%, 100.0%) |
| Complex SNV | 0 | 0 | 0 | 243 | 243 | N/A | 100.0% (98.4%, 100.0%) | N/A | 100.0% (98.4%, 100.0%) |
| Deletion (15-18bp) | 10 | 2 | 0 | 7359 | 7371 | 100.0% (72.2%, 100.0%) | 100.0% (99.9%, 100.0%) | 83.3% (55.2%, 95.3%) | 100.0% (99.9%, 100.0%) |
| Binned by Variant Class | Test + Comp + TP | Test + Comp - FP | Test - Comp + FN | Test - Comp - TN | Total n | PPA (95%CI) | NPA (95%CI) | PPV (95%CI) | NPV (95%CI) |
| C>T|G>A | 0 | 0 | 0 | 0 | 0 | N/A | N/A | N/A | N/A |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{9}

Table 10. KRAS variants (in CRC) agreement by variant type and class

| Binned by Variant Type | Test + Comp + TP | Test + Comp - FP | Test - Comp + FN | Test - Comp - TN | Total n | PPA (95%CI) | NPA (95%CI) | PPV (95%CI) | NPV (95%CI) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| SNV | 69 | 2 | 0 | 1969 | 2040 | 100.0% (94.7%, 100.0%) | 99.9% (99.6%, 100.0%) | 97.2% (90.3%, 99.2%) | 100.0% (99.8%, 100.0%) |
| Complex SNV | 0 | 0 | 0 | 3230 | 3230 | N/A | 100.0% (99.9%, 100.0%) | N/A | 100.0% (99.9%, 100.0%) |
| Deletion (15-18bp) | 0 | 0 | 0 | 0 | 0 | N/A | N/A | N/A | N/A |
| Binned by Variant Class | Test + Comp + TP | Test + Comp - FP | Test - Comp + FN | Test - Comp - TN | Total n | PPA (95%CI) | NPA (95%CI) | PPV (95%CI) | NPV (95%CI) |
| C>T|G>A | 43 | 1 | 0 | 636 | 680 | 100.0% (91.8%, 100.0%) | 99.8% (99.1%, 100.0%) | 97.7% (88.2%, 99.6%) | 100.0% (99.4%, 100.0%) |
| Non C>T|G>A | 26 | 1 | 0 | 4563 | 4590 | 100.0% (87.1%, 100.0%) | 100.0% (99.9%, 100.0%) | 96.3% (81.7%, 99.3%) | 100.0% (99.9%, 100.0%) |

As the accuracy study samples were enrolled by the O/RDx-LCCA, PPA and NPA values were adjusted using a prevalence of $6.9\%$ for EGFR variants and $36.1\%$ for KRAS variants in the intended use population. The summary of the agreement statistics is shown in the Table 11 below.

Table 11. Summary of agreement statistics

| Binned by Sample | PPV (95%CI) | NPV (95%CI) | Unadjusted | | Adjusted | |
| --- | --- | --- | --- | --- | --- | --- |
| | | | PPA (95%CI) | NPA (95%CI) | PPA (95%CI) | NPA (95%CI) |
| Sample (EGFR) | 90.0% (69.9%, 97.2%) | 100.0% (94.1%, 100.0%) | 100.0% (82.4%, 100.0%) | 96.8% (89.1%, 99.1%) | 100.0% (46.6%, 100.0%) | 99.3% (97.7%, 99.8%) |
| Sample (KRAS) | 97.2% (90.3%, 99.2%) | 100.0% (96.3%, 100.0%) | 100.0% (94.7%, 100.0%) | 98.0% (93.1%, 99.5%) | 100.0% (93.2%, 100.00%) | 98.4% (94.6%, 99.6%) |

In total, there were 4 samples that were discordant. Among the 4 discordant results, 3 mutations detected by the O/RDx-LCCA (one for $EGFR$ and 2 for $KRAS$ ) had low VAFs, which were below the assay cut-off for the evNGS assay. The remaining discordant variant was a complex $EGFR$ Exon 19 deletion, which was detected by O/RDx-LCCA but not the evNGS assay.

To evaluate additional samples with positive calls at the low VAF, DNA from 11 representative positive clinical samples with adequate leftover material from the accuracy study were diluted with normal FFPE DNAs to create samples with low VAFs. In total 16 low VAF samples were generated and tested with O/RDx-LCCA. Of these, 5 NSCLC samples that were positive for 4 different EGFR Exon

PMA P200011: FDA Summary of Safety and Effectiveness Data

{10}

19 deletions were diluted to 0.6x-2.1x LoD levels, 2 NSCLC clinical samples positive for EGFR L858R mutation were diluted to 0.5x-1.7x LoD levels, and 4 CRC samples positive for KRAS mutations were diluted to 0.6x-2.2x LoD levels. All these sample runs met the sequencing quality metrics criteria. The results of the diluted samples were positive based on the original calls.

## 2. Analytical Sensitivity

### a. Limit of Blank

A Limit of Blank (LoB) of zero was determined across 70 independent sample libraries prepared from four FFPE specimens each of normal (non-tumor) colon and normal (non-tumor) lung tissue. Each sample was tested with 9 replicates spanning low and high DNA input, two reagent lots, and three sequencing analyses. No false positives were reported for the CDx variants.

### b. Limit of Detection

The limit of detection (LoD) for each positive variant detected by the O/RDx-LCCA was estimated using the hit rate approach where LoD is defined as the lowest VAF with 100% hit rate. A total of 4 clinical NSCLC and CRC specimens were evaluated, which included CDx SNVs and a deletion/insertion (DelIns) variant which is a complex mutation with a deletion followed by an insertion. Five titration levels and 2 reagent lots were tested. Each level was tested with 10 replicates per sample at the minimum DNA input (30 ng) for each of the two reagent lots (20 replicates per level).

The claimed LoD for each variant summarized in the table below is based on the conservative hit rate approach where the assay produced 100% positive calls. Adequate dilutions were not tested for all samples in the LoD study to determine the lowest VAF at which at least 95% of the test replicates produce correct calls using the probit approach or the lowest level with 100% hit rate (i.e., worst case scenario, if the minimum requirements for the probit approach are not met).

Table 12. Summary of O/RDx-LCCA variant limit of detection

| Gene | Variant | Variant Category | Estimated VAF% |
| --- | --- | --- | --- |
| KRAS | G13D | SNV | 3.3 |
| KRAS | G12D | SNV | 3.4 |
| EGFR | L858R | SNV | 3.0 |
| EGFR | Exon 19 Del | DelIns (a complex mutation with 19 bp deletion and 1bp insertion) | 3.7 |

### Additional Study

Since adequate dilutions were not tested in the original study, a second study was conducted with additional dilutions at lower VAFs with one dilution level

PMA P200011: FDA Summary of Safety and Effectiveness Data
11 of 45

{11}

below $100\%$ hit rate for all samples were tested to determine the lowest VAF with $100\%$ hit rate. Samples were tested at a minimum of five titration levels with 20 replicates per dilution level using 2 reagent lots at the minimum DNA input of $30~\mathrm{ng}$ . The newly defined LoD for each variant summarized in the table below is based on the hit rate approach where the assay produced $100\%$ positive calls.

Table 13. Summary of O/RDx-LCCA variant limit of detection (second LoD study)

| Gene | Variant | Variant Category | VAF% |
| --- | --- | --- | --- |
| KRAS | G13D | SNV | 2.6 |
| KRAS | G12D | SNV | 1.8 |
| EGFR | L858R | SNV | 1.5 |
| EGFR | Exon 19 Del | DelIns (a complex mutation with 19 bp deletion and 1bp insertion) | 1.7 |

# c. Tumor Content

The minimum tumor fraction required to support the robustness of the O/RDx-LCCA was evaluated. Four clinical samples with different percentages of initial tumor cell content (30% to 80%) as estimated by an external pathology lab before the study, were diluted with DNA extracted from tissue-matched normal FFPE samples resulting in five levels of final tumor content. These samples were analyzed with 20 replicates per level using the O/RDx-LCCA.

The resulting data show robustness of O/RDx-LCCA in samples with tumor content above $10\%$ at $30\mathrm{ng}$ DNA input (Table 14). The data supports O/RDx-LCCA requirement of $30\%$ tumor content.

Table 14. Detection rate of diluted tumor content by variant

| Gene Exon | Nucleotide Change | Amino Acid Change | Test Level | Detection Rate | VAF Range | VAF Mean | VAF SD | Diluted Tumor Content (%) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| KRAS Exon 2 | c.35G>A | p.Gly12Asp | L1 | 20/20 | 8.7 - 10.7 | 9.78 | 0.47 | 28.8 |
| | | | L2 | 20/20 | 4.3 - 5.8 | 4.87 | 0.39 | 14.3 |
| | | | L3 | 20/20 | 2.7 - 4 | 3.41 | 0.37 | 10.0 |
| | | | L4 | 20/20 | 2.1 - 3 | 2.52 | 0.24 | 7.4 |
| | | | L5 | 20/20 | 1.5 - 2.2 | 1.81 | 0.19 | 5.3 |
| KRAS Exon 2 | c.38G>A | p.Gly13Asp | L1 | 20/20 | 5.4 - 7.3 | 6.27 | 0.53 | 25.0 |
| | | | L2 | 20/20 | 3.8 - 4.7 | 4.25 | 0.26 | 16.9 |
| | | | L3 | 20/20 | 2.7 - 3.8 | 3.30 | 0.32 | 13.1 |
| | | | L4 | 20/20 | 2 - 3.1 | 2.60 | 0.34 | 10.3 |
| | | | L5 | 15/20 | 1.6 - 2 | 1.71 | 0.12 | 6.8 |
| | | | L1 | 20/20 | 6.2 - 9.7 | 7.85 | 1.05 | 39.9 |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{12}

| Gene Exon | Nucleotide Change | Amino Acid Change | Test Level | Detection Rate | VAF Range | VAF Mean | VAF SD | Diluted Tumor Content (%) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| EGFR Exon 19 | c.2237_225 5delinsT | p.Glu746_Ser 752delinsVal | L2 | 20/20 | 3.1 - 6.4 | 4.92 | 0.86 | 25.0 |
| | | | L3 | 20/20 | 2.4 - 5.1 | 3.72 | 0.65 | 18.9 |
| | | | L4 | 20/20 | 1.6 - 3.4 | 2.42 | 0.42 | 12.3 |
| | | | L5 | 20/20 | 1 - 2.6 | 1.66 | 0.41 | 8.4 |
| EGFR Exon 21 | c.2573T>G | p.Leu858Arg | L1 | 20/20 | 7.2 - 9.6 | 8.32 | 0.70 | 18.0 |
| | | | L2 | 20/20 | 4.9 - 7.1 | 6.05 | 0.62 | 13.1 |
| | | | L3 | 20/20 | 1.9 - 4.3 | 3.02 | 0.56 | 6.6 |
| | | | L4 | 20/20 | 2 - 3.7 | 2.59 | 0.49 | 5.6 |
| | | | L5 | 20/20 | 1.2 - 1.9 | 1.53 | 0.18 | 3.3 |

## d. DNA Input

The recommended DNA input range of the O/RDx-LCCA is between 30 ng to 80 ng. The DNA input range was evaluated at 5, 10, 20, 40, 80, and 160 ng in duplicate using DNA extracted from 10 FFPE samples containing reportable CDx SNV and deletion variant representatives of EGFR and KRAS genes indicated in Table 1 of the intended use statement. The expected variants (KRAS G12X, KRAS G13X EGFR Exon 19 deletion and EGFR L858R) present in the 10 samples were called correctly at DNA inputs of 5-160 ng. At 5 ng of DNA input, 5 out of 20 samples failed to generate sequencing libraries that met the library yield of ≥ 3.5 nM, which is the minimum library yield requirement recommended for O/RDx-LCCA. At 10 ng of DNA input, 2 out of 20 samples failed the library yield requirement. Seven samples that failed library yield QC requirement were processed to completion to assess results below minimum DNA input of 30 ng/test. The data showed that 10-80 ng of DNA input for the O/RDx-LCCA produced accurate results (at the variant level: PPA=100.0% [95% CI: 95.4%, 100%] (80/80), NPA=100.0% [95% CI: 99.9%, 100%] (9999/10000)); and therefore, supports a DNA input range of 30 ng to 80 ng for the O/RDx-LCCA.

## 3. Analytical Specificity

### a. Interfering Substances

To evaluate the potential impact of interfering substances on the performance of the O/RDx-LCCA, four CRC and four NSCLC FFPE specimens including reportable CDx SNV and deletion variant representatives of EGFR and KRAS genes indicated in Table 1 of the intended use statement were evaluated in the presence of exogenous and endogenous substances. Each specimen was assessed with two replicates for each of the eight interferents tested at low and high concentrations for a total of 16 test libraries. Testing was performed at 1-1.5x LoD for the CDx variants for exogenous interfering substances and near the minimum assay requirement of 30 ng DNA input. The following interferents and interferent concentrations were used: paraffin in xylene (0.000002% and 0.000267%), Proteinase K (0.004 and 0.043 ug/mL), lysis buffers Buffer ATL

PMA P200011: FDA Summary of Safety and Effectiveness Data
13 of 45

{13}

(0.0002% and 0.0019%), and Buffer AL (0.0002% and 0.0021%), extraction wash buffers AW1 (0.06% and 0.33%), and AW2 (5.7% and 16.7%), ethanol (4.0% and 11.9%) and hemoglobin (1 mg/mL and 2 mg/mL) (Table 15). The concentrations for exogenous interferents is given relative to the eluted DNA sample, and for hemoglobin, relative to the lysis solution postdeparaffinization. No impact on the performance of the O/RDx-LCCA was observed for each substance and at each level tested (Table 16).

Table 15. List of interfering substances tested

| Exogenous/ Endogenous | Interfering Substance | Min Value | Max Value | # tested |
| --- | --- | --- | --- | --- |
| Endogenous | Control | N/A | N/A | 8 |
| | Hemoglobin | 1 mg/mL | 2 mg/mL | 16 |
| Exogenous | Control | N/A | N/A | 8 |
| | Buffer AL | 0.0002% | 0.0021% | 16 |
| | Buffer ATL | 0.0002% | 0.0019% | 16 |
| | Buffer AW1 | 0.06% | 0.33% | 16 |
| | Buffer AW2 | 5.7% | 16.7% | 16 |
| | Ethanol | 4.0% | 11.9% | 16 |
| | Xylene | 0.000002% | 0.000267% | 16 |
| | Proteinase K | 0.000004 | 0.000043 | 16 |
| | Total | | | 144 |

Table 16. Summary of interference study results

| Study | Interfering Substance | Test Value | TP | FP | FN | TN | PPA (95%CI) | NPA (95%CI) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Endogenous | Hemoglobin | Min | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | | Max | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| Endogenous | Buffer AL | Min | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | | Max | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | Buffer ATL | Min | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | | Max | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | Buffer AW1 | Min | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | | Max | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{14}

| Study | Interfering Substance | Test Value | TP | FP | FN | TN | PPA (95%CI) | NPA (95%CI) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | Buffer AW2 | Min | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | | Max | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | Ethanol | Min | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | | Max | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | Xylene | Min | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | | Max | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | Proteinase K | Min | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |
| | | Max | 8 | 0 | 0 | 1000 | 100% (67.6%, 100%) | 100% (99.6%, 100%) |

# b. Necrotic Fraction

Retrospective analyses of impact of necrotic tissue content in FFPE samples from clinical validation and analytical accuracy studies are shown below. Samples with $&lt; 20\%$ necrotic tissue content in analytical accuracy study excluded from retrospective analysis. 274 CRC and 276 NSCLC FFPE specimens with varying quantities of necrosis (1% to 70%) were assessed. All samples $&gt;20\%$ necrotic content passed library yield QC, PiVAT® results were valid and concordant with comparator assays. Five discordant results were observed in CRC that were not correlated with high necrotic content. Three discordant calls were observed in NSCLC that were not correlated with high necrotic content. For details on the discordant results, see the footnotes in Table 17. No clear trend in decreasing performance with increasing necrotic fraction in the sample was observed supporting the conclusion that the performance of the O/RDx-LCCA is robust within the recommended range of necrotic content less than $50\%$ (Table 17).

PMA P200011: FDA Summary of Safety and Effectiveness Data

{15}

Table 17. Summary of necrotic content results

| Indication | % Necrotic Bin | # Enrolled/Tetse | # Library yield QC fail | # Library yield QC pass | # PiVAT® invalid | # PiVAT® valid | # Included in analysis | # of CDx- | # of CDx+ | # CDx- Concordant result 1 | # CDx+ Concordant result 1 |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| CRC | ≤10 | 116 | 6 | 110 | 0 | 110 | 96 | 62 | 34 | 61² | 32³,⁴a |
| | 11-20 | 11 | 1 | 10 | 0 | 10 | 9 | 6 | 3 | 6 | 3 |
| | 21-30 | 4 | 0 | 4 | 0 | 4 | 4 | 1 | 3 | 1 | 3 |
| | 31-40 | 1 | 0 | 1 | 0 | 1 | 1 | 1 | 0 | 1 | 0 |
| | 41-50 | 1 | 0 | 1 | 0 | 1 | 1 | 1 | 0 | 1 | 0 |
| | >50 | 1 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 0 | 1 |
| | Not available | 140 | 9 | 131 | 0 | 131 | 114 | 58 | 56 | 58 | 54³,⁴b |
| | CRC Total | 274 | 16 | 258 | 0 | 258 | 226 | 129 | 97 | 128 | 93 |
| NSCLC | ≤10 | 187 | 8 | 179 | 0 | 179 | 176 | 120 | 56 | 120 | 56 |
| | 11-20 | 15 | 0 | 15 | 0 | 15 | 15 | 12 | 3 | 12 | 3 |
| | 21-30 | 7 | 0 | 7 | 0 | 7 | 7 | 4 | 3 | 4 | 3 |
| | 31-40 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| | 41-50 | 1 | 0 | 1 | 0 | 1 | 1 | 1 | 0 | 1 | 0 |
| | >50 | 1 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 0 | 1 |
| | Not available | 65 | 6 | 59 | 0 | 59 | 59 | 37 | 22 | 34⁵ | 22 |
| | NSCLC Total | 276 | 14 | 262 | 0 | 262 | 259 | 174 | 85 | 171 | 85 |
| | CRC + NSCLC | 550 | 550 | | 520 | | 485 | 485 | | 477 | |

1 FDA-approved comparator companion diagnostic (CCD) assay used for concordance analysis of CRC samples in clinical validation studies: therascreen® KRAS assay. The CCD assay used for concordant analysis of NSCLC samples in clinical validation studies: cobas® EGFR Mutation Test v2. The concordance results above is provided based on the CCD1 and FCD results only. For details on the discordant results, see Table 37 and Table 40.
2 The replicates of the comparator (CCD1/CCD2 = KRAS negative) were discordant with FCD (KRAS 13VAL; c.38_39delinsTT). It is inferred that therascreen® KRAS is not designed to detect complex SNVs, this result may indicate an error by therascreen® KRAS Assay.
3 The replicates of the comparator (CCD1/CCD2 = KRAS 12VAL) were discordant with FCD (KRAS 12PHE; c.34_35delinsTT). It is inferred that therascreen® KRAS is not designed to detect complex SNVs, this result may indicate an error by therascreen® KRAS Assay.
4 The replicates of the comparator were discordant.
$^{\mathrm{a}}$ CCD1 = KRAS 12ALA; CCD2 = KRAS 12VAL; FCD = KRAS 12VAL

PMA P200011: FDA Summary of Safety and Effectiveness Data

{16}

${}^{\mathrm{b}}\mathrm{{CCD}}1 = {KRAS12}\mathrm{{ARG}};\mathrm{{CCD}}2 = {KRAS12}\mathrm{{CYS}};\mathrm{{FCD}} = {KRAS12}\mathrm{{CYS}}$

5 A total of three unique clinical specimens with $EGFR$ L858R mutation showed discordant results. For all three samples, their CCD1/CCD2 results using cobas were both negative and their O/RDx-LCCA results were positive with VAF range $1.9\%$ to $4.9\%$. These results suggest the discordant cases are likely due to difference in detection sensitivity (Limit of Detection: cobas=5%) and the O/RDx-LCCA results are likely correct.

Abbreviations: Lib=sample library

## c. Cross-Contamination

To assess intra-run cross-contamination, 24 replicates of a positive cell line sample containing $EGFR$ L858R at $\sim 50\%$ VAF and 24 replicates of NTC were processed on the same plate in a checkerboard format. No false positive calls (0/24, 0%) were detected in all NTC samples. Therefore, no cross-contamination was observed.

To assess inter-run cross-contamination, a retrospective study utilizing sequencing runs generated as part of validation testing were analyzed. Indices that were used in Sequencing Run 1 and theoretically absent from Sequencing Run 2 (unexpected indices) were identified and enumerated in the output of Sequencing Run 2. Reads from index combinations used in Sequencing Run 1 could arise from run-to-run carryover, or they could arise from within run events, such as PCR errors and index hopping. The fraction of reads associated with unexpected indices across all five Run 2 data sets analyzed was less than $1\%$ ($\leq 0.4\%$) of the minimum number of reads for any sample within that run, well below the level where the unexpected reads could generate false positive results.

## d. Cross-Reactivity

An in-silico cross-reactivity analysis was performed to evaluate the specificity of the primers used in the OR/Dx-LCCA. The primers were checked for specificity to the human genome (hg19) and the genomes of representative protozoal, viral, fungal, and bacterial human pathogens. A total of 177 human and 259 pathogen non-target sequences with some similarity to the human genome were identified using in silico PCR and BLAT analysis. These sequences were converted to FASTQ format and processed through the PiVAT® software. The test samples produced no on-target reads and no variant calls for any of the non-target sequences while producing the expected variant calls for positive controls included in the analysis. These results demonstrated that the primers are specific for the intended targeted sequences.

## 4. Precision and Reproducibility

The reproducibility of the O/RDx-LCCA was evaluated using 10 clinical samples with target variants adjusted to a variant allele frequency percent (VAF%) in the range of 1-3x of the currently established LoD using DNA extracted from clinically normal tissue. The sample panel included two (2) FFPE CRC specimens with KRAS mutations (Gly12Asp and Gly13Asp), two (2) FFPE NSCLC specimens with EGFR mutations (Glu746_Ser752delinsVal and Leu858Arg) and

PMA P200011: FDA Summary of Safety and Effectiveness Data
17 of 45

{17}

one (1) FFPE CRC specimen negative for CDx variants. Each variant was present at both high and low VAF% levels.

The study was conducted at three sites with 2 operators at each site performing 3 runs on non-consecutive days. One sequencing instrument and 2 reagent lots were used at each site. Each sample was tested in 4 replicates in each run for a total of 36 possible results (3 sites by 3 runs by 4 replicates). The study produced a total of 360 test results.

PPA and NPA values with two-sided 95% confidence intervals were calculated across all tests performed. The observed PPA value for target variants was 100% (98.7%, 100%), and NPA was 100% (100%, 100%). A variance component analysis was performed for each of the sample/variant level to estimate variability of the assay including site, operator, day (site, operator), replicate and reagent lot. The total standard deviations of VAF% ranged from 0.33% to 0.70%.

## Agreement at the Variant and Variant Frequency Level

Observed mean VAF% and positive call rates with 95% confidence intervals across sample variants at both high and low VAF% concentration for the 36 replicates is presented in Table 18 below.

Table 18. Agreement at the variant and variant frequency level

| VAF Level | Gene Exon | Nucleotide Change | Amino Acid Change | N | Mean VAF (%) | Fold LoD* | Positive call rate (%) | 95% CI (LB, UB) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| High | KRAS Exon 2 | c.35G>A | p.Gly12Asp | 36 | 6.80 | 2.0 | 36/36 (100%) | 90.4%, 100% |
| | KRAS Exon 2 | c.38G>A | p.Gly13Asp | 36 | 6.91 | 2.1 | 36/36 (100%) | 90.4%, 100% |
| | EGFR Exon 19 | c.2237_2255delinsT | p.Glu746_Ser 752delinsVal | 36 | 5.14 | 1.4 | 36/36 (100%) | 90.4%, 100% |
| | EGFR Exon 21 | c.2573T>G | p.Leu858Arg | 36 | 8.91 | 3.0 | 36/36 (100%) | 90.4%, 100% |
| Low | KRAS Exon 2 | c.35G>A | p.Gly12Asp | 36 | 4.69 | 1.4 | 36/36 (100%) | 90.4%, 100% |
| | KRAS Exon 2 | c.38G>A | p.Gly13Asp | 36 | 3.66 | 1.1 | 36/36 (100%) | 90.4%, 100% |
| | EGFR Exon 19 | c.2237_2255delinsT | p.Glu746_Ser 752delinsVal | 36 | 3.02 | 0.8 | 36/36 (100%) | 90.4%, 100% |
| | EGFR Exon 21 | c.2573T>G | p.Leu858Arg | 36 | 4.96 | 1.7 | 36/36 (100%) | 90.4%, 100% |

* The LoD is based on Table 12.

## a. Site-to-Site Reproducibility

Each of the study sites performed a total of 120 tests. An analysis of test performance across study sites, measured as PPA and NPA with 95%

PMA P200011: FDA Summary of Safety and Effectiveness Data
18 of 45

{18}

confidence intervals, is provided in Table 19 below. The PPA and NPA observed at each site was 100%. All 4 replicates tested for each sample at both high and low VAF% levels were concordant in each of the 9 runs tested across 3 sites with no false negatives, i.e. 100% concordant (within run precision).

Table 19. Overall agreement by site and 3-sites combined
| Bin | N | TP | FP | FN | TN | PPA (2-sided 95% CI) | NPA (2-sided 95% CI) |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Overall 3 Sites | 360 | 288 | 0 | 0 | 45072 | 100% (98.7%,100%) | 100% (100%,100%) |
| Site 1 | 120 | 96 | 0 | 0 | 15024 | 100% (96.2%, 100%) | 100% (100.0%, 100%) |
| Site 2 | 120 | 96 | 0 | 0 | 15024 | 100% (96.2%, 100%) | 100% 100.0%, 100%) |
| Site 3 | 120 | 96 | 0 | 0 | 15024 | 100% (96.2%, 100%) | 100% (100.0%, 100%) |

b. Lot-to-Lot Precision

A total of 3 manufactured reagent lots were used in the study with 2 reagent lots tested at each site. The calculated PPA and NPA values were identical across reagent lots with mean and two-sided 95% confidence intervals of 100% (96.2%, 100%) for PPA and 100% (100%, 100%) for NPA. The reagent lot component of the total standard deviation of VAF% ranged from 0.08% to 0.33%.

c. Thermocycler Variability

A total of 3 make/model of thermocyclers were used in the study. The calculated PPA and NPA values were identical across thermocyclers with a mean of 100% for PPA and 100% for NPA and with the two-sided 95% confidence intervals as shown in the table below.

Table 20. Precision by Thermocycler
| Bin | # tests | TP | FP | FN | TN | PPA (2-sided 95% CI) | NPA (2-sided 95% CI) |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Eppendorf MasterCycler | 120 | 96 | 0 | 0 | 15024 | 100% (96.2%, 100%) | 100% (100%, 100%) |
| ABI GeneAmp 9700 | 40 | 32 | 0 | 0 | 5008 | 100% (89.3%, 100%) | 100% (99.9%, 100%) |
| Bio-Rad C1000 | 200 | 160 | 0 | 0 | 25040 | 100% (97.7%, 100%) | 100% (100.0%, 100%) |

d. Additional Study

An additional single-site precision study was conducted using 11 clinical samples with target CDx variants adjusted to a VAF% in the range of 1-1.5x

PMA P200011: FDA Summary of Safety and Effectiveness Data
19 of 45

{19}

of the LoD defined based on the second LoD study (see Section IX.A.2.b, Table 13) using DNA extracted from tissue-matched clinically normal FFPE tissue. The sample panel included three (3) FFPE NSCLC specimens with unique EGFR Exon 19 deletion variants, three (3) FFPE NSCLC specimens with EGFR Exon 21 L858R mutations, three (3) FFPE CRC specimens with unique KRAS G12 variants and two (2) FFPE CRC specimens with KRAS G13 variants.

The study was conducted at a single site with 2 operators performing 3 runs on non-consecutive days. Each sample was tested in 4 replicates in each run for a total of 12 possible results (3 runs by 4 replicates).

PPA and NPA values with two-sided $95\%$ confidence intervals were calculated across all tests performed. The observed PPA value for target variants for the site was $99.2\%$ (95.8%, 99.9%) and was $98.6\%$ (92.4%, 99.8%) for EGFR and $100\%$ (94.0%, 100%) for KRAS on a gene-level. The NPA on a site-level and on a gene-level was $100\%$ for all comparisons.

# 1. Agreement by Sample at the Variant Level

Observed mean VAF% and positive call rates with 95% confidence intervals across 11 clinical samples with target variants for the 12 replicates is presented in Table 21 below.

Table 21. Agreement by specimen at the variant level

| Gene Exon | Nucleotide Change | Amino Acid Change | Total Calls | Mean VAF (%) | Fold LoD** | Positive call rate (%) | 95% CI (LB, UB) |
| --- | --- | --- | --- | --- | --- | --- | --- |
| EGFR Exon 19 | c.2240_2254del | p.Leu747_Thr751del | 12 | 2.39 | 1.4 | 12/12 (100%) | 75.8%, 100% |
| KRAS Exon 2 | c.35G>A | p.Gly12Asp | 12 | 2.80 | 1.6 | 12/12 (100%) | 75.8%, 100% |
| KRAS Exon 2 | c.34G>T | p.Gly12Cys | 12 | 1.94 | 1.1 | 12/12 (100%) | 75.8%, 100% |
| EGFR Exon 21 | c.2573T>G | p.Leu858Arg | 12 | 2.30 | 1.5 | 12/12 (100%) | 75.8%, 100% |
| EGFR Exon 19 | c.2236_2250del | p.Glu746_Ala750del | 11* | 2.23 | 1.3 | 11/11 (100%) | 74.1%, 100% |
| KRAS Exon 2 | c.38G>A | p.Gly13Asp | 12 | 3.42 | 1.3 | 12/12 (100%) | 75.8%, 100% |
| EGFR Exon 21 | c.2573T>G | p.Leu858Arg | 12 | 1.73 | 1.2 | 12/12 (100%) | 75.8%, 100% |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{20}

| Gene Exon | Nucleotide Change | Amino Acid Change | Total Calls | Mean VAF (%) | Fold LoD** | Positive call rate (%) | 95% CI (LB, UB) |
| --- | --- | --- | --- | --- | --- | --- | --- |
| KRAS Exon 2 | c.35G>T | p.Gly12Val | 12 | 1.54 | 0.9 | 12/12 (100%) | 75.8%, 100% |
| EGFR Exon 21 | c.2573T>G | p.Leu858Arg | 12 | 1.55 | 1.0 | 11/12 (91.7%) | 64.6%, 98.5% |
| EGFR Exon 19 | c.2235_2249del | p.Glu746_Ala7 50del | 12 | 1.89 | 1.1 | 12/12 (100%) | 75.8%, 100% |
| KRAS Exon 2 | c.38G>A | p.Gly13Asp | 12 | 3.16 | 1.2 | 12/12 (100%) | 75.8%, 100% |

* One replicate produced a 'Not Valid' result. Investigation suggests the library was inadvertently excluded during library pooling and was not sequenced.
** The LoD is based on Table 13.

2. Agreement on Site and Gene Level

The study site performed a total of 132 tests. An analysis of test performance on a site and gene level, measured as PPA and NPA with 95% confidence intervals, is provided in Tables 22 and 23 below, respectively. The PPA and NPA observed on a site level was 99.2% (95.8%, 99.9%) and 100% (100%, 100%), respectively. The observed PPA value for target variants was 98.6% (92.4%, 99.8%) for EGFR and 100% (94.0%, 100%) for KRAS on a gene-level and the NPA on a gene-level was 100% for both genes.

Table 22. Overall agreement by site

| Site | Positive /Total Calls | PPA (2-sided 95% CI) | Negative/ Total Calls | NPA (2-sided 95% CI) |
| --- | --- | --- | --- | --- |
| Site 1 | 130/131 | 99.2% (95.8%,99.9%) | 16375/16375 | 100% (100%,100%) |

Table 23. Overall agreement by gene

| Gene | Positive /Total Calls | PPA (2-sided 95% CI) | Negative/ Total Calls | NPA (2-sided 95% CI) |
| --- | --- | --- | --- | --- |
| EGFR | 70/71 | 98.6% (92.4%, 99.8%) | 8875 / 8875 | 100% (100%,100%) |
| KRAS | 60/60 | 100% (94.0%, 100%) | 7500 / 7500 | 100% (99.9%, 100%) |

Adequate number of samples harboring CDx biomarkers/variants at the LoD levels based on the second LoD study (Table 13) were not evaluated in a 3-site reproducibility study. This was due to the onset of the COVID-19 emergency and associated precautionary measures taken by Pillar, specifically an unanticipated reduction in lab supplies and personnel. Therefore, a post-market 3-site

PMA P200011: FDA Summary of Safety and Effectiveness Data

{21}

reproducibility study is planned with samples carrying CDx variants and covering different EGFR Exon 19 deletions, EGFR Exon 21 L858R mutations and KRAS codon 12/13 variants at the newly established LoD levels (Table 13) to supplement the existing studies such that the assay precision is demonstrated to be robust near the true LoD levels of variants that are detected by O/RDx-LCCA.

## 5. DNA Extraction Method Equivalence

A study evaluating performance of three commercially available FFPE tissue extraction kits was conducted because extraction kits are not included in the O/RDx-LCCA kit. Four FFPE CRC (including one FFPE CRC negative for CDx variants), one normal colon tissue, four FFPE NSCLC (including one FFPE negative for CDx variants), and one normal lung tissue samples were used in the study. The six tumor specimens that were selected to be CDx variant positive included reportable CDx SNV and deletion variant representatives of EGFR and KRAS genes indicated in Table 1 of the intended use statement. Genomic DNA was extracted using 3 commercially available FFPE extraction kits. Each extracted DNA sample was run in duplicate using the O/RDx-LCCA. PPA was 100% (95% CI: 75.8%, 100%) (12/12) and NPA was 100% (95% CI: 99.8%, 100%) (2508/2508) at the variant level for each of the two commercially available FFPE DNA extraction kits compared to the validated reference kit. The results demonstrate that the 3 methods yield DNA with comparable quality and quantity to generate reliable results when used with O/RDx-LCCA.

## 6. Guard Band Studies

The tolerances encompassing the library preparation and sequencing workflow steps were assessed, which correspond to the test's most critical steps that could lead to assay failure. Each workflow step tested included a minimum of 3 test conditions: low; nominal, as defined by the assay instructions for use; and high. The guard-banding range for each study was designed such that the maximum and minimum test points challenged the system, while still being within operational error range.

Ten FFPE-extracted DNA samples were analyzed over 4 sequencing runs to assess library preparation workflow steps such as PCR input and thermal cycling temperature offset. The seven tumor specimens were selected to be CDx variant positive and included reportable CDx SNV and deletion variant representatives of EGFR and KRAS genes indicated in Table 1 of the intended use statement. The six CDx variant positive tumor specimens were tested for each assay specification. One reference standard DNA (HD799: Quantitative Multiplex Formalin Compromised (Moderate) formalin compromised DNA) containing reportable CDx SNV and deletion variant representatives of EGFR and KRAS genes indicated in Table 1 of the intended use statement was prepared and analyzed over 5 sequencing runs to assess library sequencing workflow steps such as library concentration and number of libraries per run. The conditions of testing of the assay's most critical steps are shown in Table 24. All studies resulted in zero failures and 100% agreement across conditions as shown in Table 25.

PMA P200011: FDA Summary of Safety and Effectiveness Data
22 of 45

{22}

Table 24. DNA input test levels and sample concentration

| Process | Variable | Nominal Value or Range | Test values |
| --- | --- | --- | --- |
| Gene-specific PCR | DNA input/test | 10 to 80 ng | 5 ng
30 ng
160 ng |
| Indexing PCR | Purified GS-PCR product input volume/test | 6 μL | 3 μL
6 μL
9 μL |
| Library pooling | Number of libraries† per sequencing run | 12 to 48 libraries per run | 6 libraries/ run
12 libraries/run
48 libraries/run
54 libraries/run |
| Library normalization | Library input per sequencing run | 3.5 to 5.0 nM | 1 nM
2 nM
3 nM
5 nM
6 nM |
| Gene-specific and Indexing PCR | Cycling temperatures | Cycling profile in instructions for use | -1°C offset
Nominal
+1°C offset |

† Sample libraries including PosCtrl and NTC

Table 25. Summary of results for guard banding studies

| Process | Variable | Nominal value /range | Test values | TP | FP | FN | TN | PPA (2-sided 95% CI) | NPA (2-sided 95% CI) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| GS-PCR | DNA input/test | 30 - 80 ng | 5 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| | | | 160 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| I-PCR | Purified GS-PCR product input volume/test | 6 μL | 3 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| | | | 9 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| GS-LP | Cycling temperatures | Standard profile in | Standard - 1°C | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{23}

| Process | Variable | Nominal value /range | Test values | TP | FP | FN | TN | PPA (2-sided 95% CI) | NPA (2-sided 95% CI) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | User Manual | Standard + 1°C | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| Library pooling | Number of libraries† per sequencing run | 12 - 48 libraries/ run | 6 | 4 | 0 | 0 | 500 | 100% (51.0%, 100%) | 100% (99.2%, 100%) |
| | | | 12 | 10 | 0 | 0 | 1250 | 100% (72.2%, 100%) | 100% (99.7%, 100%) |
| | | | 54 | 52 | 0 | 0 | 6500 | 100% (93.1%, 100%) | 100% (99.9%, 100%) |
| Library normalization | Library input†† per sequencing run | 3.5 – 4.5 nM | 1, 2, 3, 4, 5 and 6 nM | 46 | 0 | 0 | 5750 | 100% (92.3%, 100%) | 100% (99.9%, 100%) |

† Sample libraries including PosCtrl and NTC
†† 1 library each tested at 1, 2, 3, 5 and 6nM; 41 libraries tested at 4 nM (reference)

## 7. Stability Studies

### a. Reagent Kit Shelf-life Stability:

Three separately manufactured kit lots including all components of the O/RDx-LCCA were stored according to the storage conditions specified in product labeling. The stability of the reagents was evaluated by testing at least three (3) reference standard DNA including reportable CDx SNV and deletion variant representatives of EGFR and KRAS genes indicated in Table 1 of the intended use statement at specified time points from baseline (Table 26).

- HD701 - Quantitative Multiplex gDNA Multiplex
- HD803 - Quantitative Multiplex Formalin Compromised (Severe) formalin compromised DNA
- HD799 - Quantitative Multiplex Formalin Compromised (Moderate) formalin compromised DNA

Each of the assay QC metrics were evaluated in addition to final calls. Calls and metrics were confirmed against the calls for the kit at the baseline time (i.e., month 0). The shelf-life stability will continue to be evaluated to extend the shelf-life stability claim (see additional studies below [Table 28] for extended stability duration).

Table 26. Summary of reagent kit shelf-life stability studies

PMA P200011: FDA Summary of Safety and Effectiveness Data

{24}

| Kit Lot | Timepoint (months) | # tests | PPA (2-sided 95% CI)* | NPA (2-sided 95% CI)* |
| --- | --- | --- | --- | --- |
| 1 | 2 | 3 | 100% (67.6%, 100%) | 100% (99.0%, 100%) |
| | 4 | 3 | 100% (67.6%, 100%) | 100% (99.0%, 100%) |
| | 7 | 3 | 100% (67.6%, 100%) | 100% (99.0%, 100%) |
| | 9 | 3 | 100% (67.6%, 100%) | 100% (99.0%, 100%) |
| | 14 | 6 | 100% (80.6%, 100%) | 100% (99.5%, 100%) |
| | 16 | 6 | 100% (80.6%, 100%) | 100% (99.5%, 100%) |
| 2 | 4 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |
| | 7 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |
| | 10 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |
| 3 | 3.5 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |
| | 6.5 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |
| | 10 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |

## Additional Studies

## Reagent Kit Shelf-Life Stability – Clinical Samples

The stability of the reagents was further evaluated in an additional study by testing three (3) clinical samples with target CDx variants adjusted to a VAF% in the range of 1-1.5x of the LoD (Table 13) using DNA extracted from tissue-matched clinically normal FFPE tissue. The sample panel included one (1) FFPE NSCLC specimens with EGFR Exon 19 deletion variant, one (1) FFPE CRC specimen with KRAS G12 variant and one (1) FFPE CRC specimen with KRAS G13 variant. Three reagent kit lots aged 17, 18 and 24 months were used as representative assay reagent lots to test the samples in replicates of five with each of the 3 reagent lots for a total of 15 replicates per sample at the minimum DNA input of 30 ng.

Each of the assay QC metrics were evaluated in addition to final calls. The detection rate of each sample across all three lots tested was 100% (15/15) as shown in Table 27 below.

PMA P200011: FDA Summary of Safety and Effectiveness Data

{25}

Table 27. Performance of each reagent kit lot across clinical samples

| Gene Exon | Nucleotide Change | Mean VAF (%) | Fold LoD* | Total Calls | Lot | Detection rate (%) |
| --- | --- | --- | --- | --- | --- | --- |
| KRAS Exon 2 | c.35G>A | 2.49 | 1.4 | 5 | 1 | 5/5 (100%) |
| | | | | 5 | 2 | 5/5 (100%) |
| | | | | 5 | 3 | 5/5 (100%) |
| KRAS Exon 2 | c.38G>A | 3.53 | 1.4 | 5 | 1 | 5/5 (100%) |
| | | | | 5 | 2 | 5/5 (100%) |
| | | | | 5 | 3 | 5/5 (100%) |
| EGFR Exon 19 | c.2235_2249del | 1.94 | 1.1 | 5 | 1 | 5/5 (100%) |
| | | | | 5 | 2 | 5/5 (100%) |
| | | | | 5 | 3 | 5/5 (100%) |

* The LoD is based on Table 13.

Reagent Kit Shelf-Life Stability – Extended Stability Duration
The shelf-life stability of reagent kits was evaluated for additional time points to extend the shelf-life stability claim. The data currently available support at least 13 months of stability for O/RDx-LCCA kit components for all 3 lots evaluated as shown in Table 28 below.

Table 28. Summary of reagent kit shelf-life stability studies

| Kit Lot | Timepoint (months) | # tests | PPA (2-sided 95% CI)* | NPA (2-sided 95% CI)* |
| --- | --- | --- | --- | --- |
| 1 | 2 | 3 | 100% (67.6%, 100%) | 100% (99.0%, 100%) |
| | 4 | 3 | 100% (67.6%, 100%) | 100% (99.0%, 100%) |
| | 7 | 3 | 100% (67.6%, 100%) | 100% (99.0%, 100%) |
| | 9 | 3 | 100% (67.6%, 100%) | 100% (99.0%, 100%) |
| | 14 | 6 | 100% (80.6%, 100%) | 100% (99.5%, 100%) |
| | 16 | 6 | 100% (80.6%, 100%) | 100% (99.5%, 100%) |
| 2 | 4 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |
| | 7 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |
| | 10 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |
| | 13 | 6 | 100% (86.2%, 100%) | 100% (99.5%, 100%) |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{26}

| Kit Lot | Timepoint (months) | # tests | PPA (2-sided 95% CI)* | NPA (2-sided 95% CI)* |
| --- | --- | --- | --- | --- |
| | 17 | 6 | 100%
(86.2%, 100%) | 100%
(99.5%, 100%) |
| 3 | 3.5 | 6 | 100%
(86.2%, 100%) | 100%
(99.5%, 100%) |
| | 6.5 | 6 | 100%
(86.2%, 100%) | 100%
(99.5%, 100%) |
| | 10 | 6 | 100%
(86.2%, 100%) | 100%
(99.5%, 100%) |
| | 13 | 6 | 100%
(86.2%, 100%) | 100%
(99.5%, 100%) |
| | 16 | 6 | 100%
(86.2%, 100%) | 100%
(99.5%, 100%) |

b. Reagent Kit Transport Stability:

The reagent kit stability studies were performed as one large study that included data points for in-use freeze-thaw stability and transport stability testing under recognized summer and winter profiles for international shipments (Table 29). The transport stability study was performed to demonstrate that the shipping configurations for all kit components provide adequate thermal and physical protection as packages are transported from the manufacturing site to customers. Three (3) separately manufactured kits and component reagent lots were exposed to simulated transport challenges intended to simulate the longest estimated international shipping times of 72 hours and 144 hours. The simulated transport conditions included both physical and temperature challenges, which include 2 packaging configurations (one (1) kit per shipping box and four (4) kits per shipping box) and 4 temperature profiles (72-hour summer, 72-hour winter, 144-hour summer and 144-hour winter). The 144-hour profiles correspond to two runs of the 72-hour profiles. The 72-hour international profile is considered to be a worse case than the 48-hour domestic profile, so a domestic profile was not performed.

Temperature challenge was performed at 3, 4, and 8 months to simulate shipping of aged components. After each temperature challenge, kits produced QC metrics and variant calls equivalent to baseline (month 0) QC metrics and variant calls of the control kit in the lot. No individual kit boxes experienced temperature higher than $-15^{\circ}\mathrm{C}$. No sign of deterioration or degradation was observed for all labels.

Physical challenge was performed only at 0 months, since the acceptance criteria was visual integrity rather than function (i.e., physical challenge is extremely unlikely to affect the functional integrity of the reagents, and so it was not tested). No packages having undergone simulated transport shipping

PMA P200011: FDA Summary of Safety and Effectiveness Data

{27}

displayed signs of physical damage which may impede the function of the assay or workstation and monitor.

Each of the kits and components undergoing temperature challenge was functionally tested using at least three (3) reference standard DNA containing reportable CDx SNV and deletion variant representatives of $EGFR$ and $KRAS$ genes indicated in Table 1 of the intended use statement to establish transport stability. The data demonstrate that all kit components show acceptable transport stability at the simulated time points (see Table 30).

Table 29. Test configuration for reagent lots 1, 2, and 3

| Line ID | Shipping Box | Number of Kits/Box | Thermal Profile | Purpose of kits used for testing at months... | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | 0-1 | 2-3 | 4-5 | 6-8 | 9-11 | 12-14 | 15-17 | 18 |
| LOT 1 | | | | | | | | | | | |
| 1 | None | Min 1 | None | SL | SL | SL | SL | SL | SL | SL | SL |
| 2 | 1 | 1 | Summer | | | | TC | | | | |
| 3 | 2 | 4 * | Summer | | | | TC | IU | | | |
| LOT 2 | | | | | | | | | | | |
| 4 | None | Min 1 | None | SL | SL | SL | SL | SL | SL | SL | SL |
| 5 | 3 | 1 | Winter | | TC | IU | | | | | |
| 6 | 4 | 4 * | Winter | | TC | | | | | | |
| LOT 3 | | | | | | | | | | | |
| 7 | None | Min 1 | None | SL | SL | SL | SL | SL | SL | SL | SL |
| 8 | 5 | 4 * | Summer | TC; IU | | | | | | | |
| 9 | 6 | 1 | Summer | TC; IU | | | | | | | |
| 10 | 7 | 4 * | Winter | TC; IU | | | | | | | |
| 11 | 8 | 1 | Winter | TC; IU | | | | | | | |
| 12 | 9 | 4 | None | 0 only, this shipping box and kits was used for simulated transport for physical challenge | | | | | | | |
| 13 | 10 | 1 | None | 0 only, this shipping box and kits wasused for simulated transport for physical challenge | | | | | | | |

Abbreviations: SL = Shelf Life; TC = Thermal Challenge; IU = In-Use.
* One (1) representative kit was tested in-use after completion of profile and storage.

## c. In-Use Stability:

The in-use stability study evaluated both open vial stability and freeze-thaw stability. For each of Lots 1, 2, and 3, representative kits were subjected to at least five (5) freeze-thaw cycles and four (4) uses. Additional testing was performed for Lot 3 to assess at least five (5) freeze-thaw cycles and two (2)

PMA P200011: FDA Summary of Safety and Effectiveness Data
28 of 45

{28}

uses. The reagents were evaluated by testing at least three (3) reference standard DNA containing reportable CDx SNV and deletion variant representatives of $EGFR$ and $KRAS$ genes indicated in Table 1 of the intended use statement. In-use stability was tested using Lots 1-3 at the baseline (i.e., month 0), 4 and 9 months in a combined study with transport stability (see above).

To perform reagent freeze-thaw cycle, the reagents were removed from freezer storage and placed in a $2^{\circ}\mathrm{C}$ to $8^{\circ}\mathrm{C}$ environment overnight (minimum of 12 hours) to simulate use, then returned to freezer storage for a minimum of 12 hours.

Each of the assay QC metrics were evaluated in addition to final calls. The data demonstrate in-use stability for at least 5 freeze-thaw cycles (Table 30).

Table 30. Summary of reagent kit transport stability and in-use stability studies

| KIT LOT | Ship Configuration (# kits/shipping box) | Stimulated Thermal Profile* | Freeze-thaw cycles** | PPA (2-sided 95% CI)*** | NPA (2-sided 95% CI)*** |
| --- | --- | --- | --- | --- | --- |
| 1 | 4 | Summer | 3 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 4 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 5 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 6 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| 3 | 4 | Summer | 3 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 4 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 5 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 6 | 92% (64.6%, 99.0%) | 100% (99.0%, 100%) |
| 3 | 1 | Summer | 5 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 6 | 92% (75.8%, 100%) | 100% (99.0%, 100%) |
| 3 | 4 | Winter | 3 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 4 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 5 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{29}

| KIT LOT | Ship Configuration (# kits/shipping box) | Stimulated Thermal Profile* | Freeze-thaw cycles** | PPA (2-sided 95% CI)*** | NPA (2-sided 95% CI)*** |
| --- | --- | --- | --- | --- | --- |
| | | | 6 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| 2 | 1 | Winter | 3 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 4 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 5 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| | | | 6 | 100% (75.8%, 100%) | 100% (99.0%, 100%) |
| 3 | 1 | Winter | 5 | 92% (64.6%, 99.0%) | 100% (99.0%, 100%) |
| | | | 6 | 92% (75.8%, 100%) | 100% (99.0%, 100%) |

* Summer and winter international profiles per ISTA 7D
** Test kit lots that had undergone simulated thermal challenges were removed from storage and subjected to the indicated freeze-thaw cycles prior to testing.
*** Agreement between test kit lots (temperature challenged and/or freeze-thawed) and control kit lot.

d. FFPE (Section and Block) and DNA Sample Stability

The stability of FFPE clinical samples (section and block) was assessed using both retrospective and a real-time analysis.

Retrospective analysis leveraged data from the clinical validation study and paired observed performance parameters with FFPE block age as determined by the difference between the reported year of specimen collection and the year when specimens were tested. Analysis focused on the QC metrics used throughout the O/RDx-LCCA workflow: sample library concentration, PiVAT® valid/invalid status, and concordance with results determined by the appropriate comparator assay. Results are summarized in Table 31 below. No significant trend in poorer overall performance with increasing FFPE block age was observed, and robust assay performance was observed for samples over 10 years old. A post-market study will be conducted to prospectively test FFPE block stability for CDx variants.

PMA P200011: FDA Summary of Safety and Effectiveness Data

{30}

Table 31. Retrospective analysis of the performance of sample libraries prepared from clinical FFPE specimens of various ages and with known comparator test results

| Tumor | FFPE Block Age (years) | # Tested | # Lib Yield QC Fail | # Lib Yield QC Pass | # PiVAT® Invalid | # PiVAT® Valid | # Included in Analysis | # of CCD- | # of CCD+ | # CCD- Concordant | # CCD+ Concordant | % FCD Discordant |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| CRC | 0-2 | 29 | 0 | 29 | 0 | 29 | 29 | 16 | 13 | 15 | 13 | 3.4% |
| | 3-5 | 38 | 0 | 38 | 0 | 38 | 38 | 28 | 10 | 28 | 10 | 0.0% |
| | 6-10 | 20 | 0 | 20 | 0 | 20 | 20 | 14 | 6 | 14 | 6 | 0.0% |
| | 11-20 | 1 | 0 | 1 | 0 | 1 | 1 | 1 | 0 | 1 | 0 | 0.0% |
| | >20 | 2 | 0 | 2 | 0 | 2 | 2 | 1 | 1 | 1 | 1 | 0.0% |
| | Total | 90 | 0 | 90 | 0 | 90 | 90 | 60 | 30 | 59 | 30 | 1.1% |
| NSCLC | 0-2 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | n/a |
| | 3-5 | 35 | 0 | 35 | 0 | 35 | 35 | 1 | 34 | 1 | 34 | 0.0% |
| | 6-10 | 1 | 0 | 1 | 0 | 1 | 1 | 1 | 0 | 1 | 0 | 0.0% |
| | 11-20 | 2 | 0 | 2 | 0 | 2 | 2 | 1 | 1 | 1 | 1 | 0.0% |
| | >20 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | n/a |
| | Total | 38 | 0 | 38 | 0 | 38 | 38 | 3 | 35 | 3 | 35 | 0.0% |
| All | 0-2 | 29 | 0 | 29 | 0 | 29 | 29 | 16 | 13 | 15 | 13 | 3.4% |
| | 3-5 | 73 | 0 | 73 | 0 | 73 | 73 | 29 | 44 | 29 | 44 | 0.0% |
| | 6-10 | 21 | 0 | 21 | 0 | 21 | 21 | 15 | 6 | 15 | 6 | 0.0% |
| | 11-20 | 3 | 0 | 3 | 0 | 3 | 3 | 2 | 1 | 2 | 1 | 0.0% |
| | >20 | 2 | 0 | 2 | 0 | 2 | 2 | 1 | 1 | 1 | 1 | 0.0% |
| | Total | 128 | 128 | | 128 | | 128 | 128 | | 127 | | 0.5% |

Abbreviations: Lib=sample library, CCD=FDA-approved comparator companion diagnostic (+/- indicate whether test was positive or negative), FCD=follow-on companion diagnostic (O/RDx-LCCA)

Stability of FFPE curls was assessed at baseline, 30 days and 60 days to support stability at 30 days. At each timepoint, eight FFPE specimens stored at ambient conditions were extracted using QIAGEN QIAamp FFPE extraction kit and processed for sequencing within 3 days. The samples contained EGFR (L858R and Ex19 del) and KRAS (G12D and G13D) variants. FFPE curls were stable, as measured by PPA/NPA analysis, at both the 30-day and 60-day time points, supporting a claim of a 30-day stability. PPA was $100\%$ (95% CI: $80.6\%$ , $100\%$ ) (16/16) and NPA was $100\%$ (95% CI: $99.8\%$ , $100\%$ ) (2000/2000) compared to baseline testing.

Stability of DNA extracted from FFPE clinical samples using QIAGEN QIAamp FFPE extraction kit was assessed after storage at $4^{\circ}\mathrm{C}$ or $-20^{\circ}\mathrm{C}$ , and after 5 cycles of freeze-thaw. Stability at $4^{\circ}\mathrm{C}$ was assessed after 60 days, 8 months, and 8.25 months and stability at $-20^{\circ}\mathrm{C}$ was assessed after 6 months and 6.5 months. FFPE derived DNA was stable, as measured by PPA/NPA analysis, at all the time points tested following storage at $4^{\circ}\mathrm{C}$ . PPA was $100\%$ (95% CI: $80.6\%$ , $100\%$ ) (16/16) and NPA was $100\%$ (95% CI: $99.8\%$ , $100\%$ )

PMA P200011: FDA Summary of Safety and Effectiveness Data

{31}

(2000/2000) for the timepoints 60 days ans 8.25 months compared to baseline testing. For the timepoint 8 month, PPA was $100\%$ (95% CI: 67.6%, 100%) (8/8) and NPA was $100\%$ (95% CI: 99.6%, 100%) (1000/1000). FFPE derived DNA was stable, as measured by PPA/NPA analysis, at all the time points tested following storage at $-20^{\circ}\mathrm{C}$ . PPA was $100\%$ (95% CI: 80.6%, 100%) (16/16) and NPA was $100\%$ (95% CI: 99.8%, 100%) (2000/2000) for the timepoints 6 months and 6.5 months compared to baseline testing.

Data presented here supports a claim of DNA storage stability at 8 months at $4^{\circ}\mathrm{C}$ and 6 months at $-20^{\circ}\mathrm{C}$ . The DNA stability at $-20^{\circ}\mathrm{C}$ will continue to be evaluated to extend the stability claim. FFPE derived DNA that was subjected to 5 cycles of freeze-thaw was stable, as determined by PPA/NPA analysis using DNA that undergone 1 cycle of freeze-thaw as a control. PPA was $100\%$ (95% CI: $80.6\%$ , $100\%$ ) (16/16) and NPA was $100\%$ (95% CI: $99.8\%$ , $100\%$ ) (2000/2000). This data supports a DNA freeze-thaw stability claim of 5 cycles.

# e. Stability of Assay Intermediates

The workflow for the O/RDx-LCCA incorporates several optional stopping points to hold assay intermediates. The stability of the intermediate products was evaluated by incorporating two optional stopping points specified in the assay instructions for use. Ten FFPE-extracted DNA samples were included in this study which contained reportable CDx SNV and deletion variant representatives of $EGFR$ and $KRAS$ genes indicated in Table 1 of the intended use statement. Each sample was processed to completion for sequencing at baseline and the resulting intermediates stored. The intermediates were then removed from storage at different time points and processed to completion to assess the impact of storage on assay performance.

The study results support the conclusion that the 60-day hold of Gene-Specific PCR (GS-PCR) products and 90-day hold on indexed libraries at recommended storage condition did not result in a decrease in O/RDx-LCCA performance as shown in Tables 32 and 33.

Table 32. Summary of results for stability of assay intermediates study

| Assay intermediate | Timepoint (days) | TP | FP | FN | TN | PPA (2-sided 95% CI)* | NPA (2-sided 95% CI)* |
| --- | --- | --- | --- | --- | --- | --- | --- |
| GS-PCR | 0 | N/A | N/A | N/A | N/A | N/A | N/A |
| GS-PCR | 35 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| GS-PCR | 62 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |

PMA P200011: FDA Summary of Safety and Effectiveness Data

{32}

| Assay intermediate | Timepoint (days) | TP | FP | FN | TN | PPA (2-sided 95% CI)* | NPA (2-sided 95% CI)* |
| --- | --- | --- | --- | --- | --- | --- | --- |
| GS-PCR | 90 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| I-PCR | 35 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| I-PCR | 90 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| I-PCR | 233 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |
| Combo* | GS-PCR 35 + I-PCR 27 | 6 | 0 | 0 | 1254 | 100% (61.0%, 100%) | 100% (99.7%, 100%) |

* Combination intermediate stability was tested by performing indexing on GS-PCR products that were 35 days old. The resulting indexed libraries was stored for 27 days at -25 to -10°C before being sequenced.

Table 33. Stability of assay intermediates

| Assay intermediate(s) | Storage condition | Intermediate stability |
| --- | --- | --- |
| GS-PCR products | -15°C to -25°C | 60 days |
| Indexed libraries | -15°C to -25°C | 90 days |

B. Animal Studies
No animal studies were conducted using the O/RDx-LCCA.

C. Additional Studies
No additional studies were conducted using the O/RDx-LCCA.

Software verification and validation activities, including unit testing, integration testing, and system testing were performed for the PiVAT® Software.

X. SUMMARY OF PRIMARY CLINICAL STUDIES

The reasonable assurance of safety and effectiveness for the O/RDx-LCCA CDx claims were established through clinical concordance studies using a non-inferiority statistical testing approach. Two clinical concordance studies were conducted to support the CDx claims indicated in Table 1 of the Indications for Use statement for EGFR Exon 19del/L858R in NSCLC and KRAS wild-type (absence of mutation in codons 12 and 13) in CRC. A non-inferiority statistical testing approach was used according to Li (2016)¹. O/RDx-LCCA test, considered the follow-on companion diagnostic (FCD), was compared to an FDA-approved CDx test, considered the comparator companion diagnostic (CCD) test

PMA P200011: FDA Summary of Safety and Effectiveness Data

{33}

for each of the clinical concordance study, using samples representati…