cobas EGFR MUTATION TEST v2

P150047 · Roche Molecular Systems, Inc. · OWD · Jun 1, 2016 · Pathology

Device Facts

• Record ID: P150047
• Device Name: cobas EGFR MUTATION TEST v2
• Applicant: Roche Molecular Systems, Inc.
• Product Code: OWD · Pathology
• Decision Date: Jun 1, 2016
• Decision: APCB
• Device Class: Class 3
• Attributes: Expedited Review

Intended Use

The cobas® EGFR Mutation Test v2 is a real-time PCR test for the qualitative detection of defined mutations of the epidermal growth factor receptor (EGFR) gene in non-small cell lung cancer (NSCLC) patients. Defined EGFR mutations are detected using DNA isolated from formalin-fixed paraffin-embedded tumor tissue (FFPET) or circulating-free tumor DNA (cfDNA) from plasma derived from EDTA anti-coagulated peripheral whole blood. The test is indicated as a companion diagnostic to aid in selecting NSCLC patients for treatment with the targeted therapies listed in Table 1 below in accordance with the approved therapeutic product labeling: Table 1. Drug FFPET Plasma TARCEVA® (erlotinib) Exon 19 deletions and L858R Exon 19 deletions and L858R TAGRISSO™ (osimertinib) T790M Patients with positive cobas® EGFR Mutation Test v2 test results using plasma specimens for the presence of EGFR exon 19 deletions or L858R mutations are eligible for treatment with TARCEVA® (erlotinib). Patients who are negative for these mutations by this test should be reflected to routine biopsy and testing for EGFR mutations with the FFPET sample type.

Device Story

The cobas® EGFR Mutation Test v2 is an in vitro diagnostic test used to identify specific EGFR gene mutations in NSCLC patients. It processes either FFPET tissue or plasma-derived cfDNA. For plasma, cfDNA is isolated manually using a glass-fiber filter column method. The device uses allele-specific real-time PCR (AS-PCR) on the cobas z 480 analyzer to amplify and detect target mutations. The system uses fluorescently labeled probes and Z05-AS1 DNA polymerase. The software (cobas 4800 system) processes fluorescent signals, assesses control validity, and determines mutation status. Results are provided to clinicians to guide treatment decisions for targeted therapies like erlotinib. By identifying patients with specific EGFR mutations, the test helps select candidates for therapy, potentially improving clinical outcomes for patients with advanced NSCLC.

Clinical Evidence

Clinical evidence based on a retrospective bridging study using plasma samples from the Phase III ENSURE trial (n=180 with 2.0 mL plasma). Primary endpoint was progression-free survival (PFS). Patients with mutation-detected (MD) results by the test treated with erlotinib showed a median PFS of 8.3 months vs. 4.6 months for chemotherapy (HR=0.29; 95% CI: 0.19, 0.45). Analytical accuracy was established by comparing plasma results to a validated NGS method (PPA 87.5%, NPA 96.8%).

Technological Characteristics

Real-time PCR assay using allele-specific primers and fluorescently labeled oligonucleotide probes. Employs Z05-AS1 DNA polymerase and AmpErase (uracil-N-glycosylase) for carryover prevention. Automated amplification/detection on cobas z 480 analyzer. cfDNA isolation via manual glass-fiber filter column. Software-based result interpretation using cycle threshold (Ct) and relative cycle threshold (CtR) values.

Indications for Use

Indicated for NSCLC patients to detect defined EGFR mutations (Exon 19 deletions, L858R, T790M, G719X, S768I, L861Q, Exon 20 insertions) in FFPET or plasma-derived cfDNA to aid in selecting patients for targeted therapy (erlotinib or osimertinib).

Regulatory Classification

Identification

The somatic gene mutation detection system is an in vitro diagnostic nucleic acid amplification and real-time detection system for the identification of genetic mutations in DNA extracted from human tumor tissues.

Predicate Devices

• cobas® EGFR Mutation Test v2 (P120019/S007)

Submission Summary (Full Text)

{0}

# SUMMARY OF SAFETY AND EFFECTIVENESS DATA (SSED)

## I. GENERAL INFORMATION

Device Generic Name: Real-time PCR test

Device Trade Name: cobas® EGFR Mutation Test v2

Device Procode: OWD

Applicant's Name and Address: Roche Molecular Systems, Inc. (RMS)
4300 Hacienda Drive
Pleasanton, CA 94588-2722

Date(s) of Panel Recommendation: None

Premarket Approval Application (PMA) Number: P150047

Date of FDA Notice of Approval: June 1, 2016

Priority Review: Granted priority review status on December 9, 2015 because the device addresses an unmet medical need, as demonstrated by significant clinically meaningful advantage.

The cobas® EGFR Mutation Test v2 (P120019/S007) was approved on November 13, 2015 for the qualitative detection of defined mutations in the epidermal growth factor receptor (EGFR) gene in DNA derived from formalin-fixed paraffin-embedded tumor tissue (FFPET) from non-small cell lung cancer (NSCLC) patients. The indications for use of the device as approved in P120019/S007 is to aid in identifying patients with NSCLC whose tumors have defined EGFR mutations and for whom safety and efficacy of a drug have been established shown in the table below. The SSED to support the FFPET indications are available on the CDRH website.

Table 1. Previously Approved cobas® EGFR Mutation Test v2 Indications for FFPET Specimens.

| Drug | Mutations |
| --- | --- |
| TARCEVA® (erlotinib) | Exon 19 deletions and L858R |
| TAGRISSO™ (osimertinib) | T790M |

The current PMA was submitted to add an indication for the cobas® EGFR Mutation Test v2 to identify NSCLC patients with Exon 19 deletions and L858R substitution mutations from cell-free DNA (cfDNA), also referred to as circulating tumor DNA (ctDNA), isolated from plasma, for treatment with TARCEVA® (erlotinib).

PMA P150047: FDA Summary of Safety and Effectiveness Data

{1}

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 2

# II. INDICATIONS FOR USE

The cobas® EGFR Mutation Test v2 is a real-time PCR test for the qualitative detection of defined mutations of the epidermal growth factor receptor (EGFR) gene in non-small cell lung cancer (NSCLC) patients. Defined EGFR mutations are detected using DNA isolated from formalin-fixed paraffin-embedded tumor tissue (FFPET) or circulating-free tumor DNA (cfDNA) from plasma derived from EDTA anti-coagulated peripheral whole blood.

The test is indicated as a companion diagnostic to aid in selecting NSCLC patients for treatment with the targeted therapies listed in Table 1 below in accordance with the approved therapeutic product labeling:

Table 1.

| Drug | FFPET | Plasma |
| --- | --- | --- |
| TARCEVA®
(erlotinib) | Exon 19 deletions and L858R | Exon 19 deletions and L858R |
| TAGRISSO™
(osimertinib) | T790M | |

Patients with positive cobas® EGFR Mutation Test v2 test results using plasma specimens for the presence of EGFR exon 19 deletions or L858R mutations are eligible for treatment with TARCEVA® (erlotinib). Patients who are negative for these mutations by this test should be reflected to routine biopsy and testing for EGFR mutations with the FFPET sample type.

Drug safety and efficacy have not been established for the following EGFR mutations listed in Table 2 below that are also detected by the cobas® EGFR Mutation Test v2:

Table 2.

| Drug | FFPET | Plasma |
| --- | --- | --- |
| TARCEVA®
(erlotinib) | G719X, exon 20 insertions,
T790M, S768I and L861Q | G719X, exon 20 insertions,
T790M, S768I and L861Q |
| TAGRISSO™
(osimertinib) | G719X, exon 19 deletions,
L858R, exon 20 insertions,
S768I, and L861Q | G719X, exon 19 deletions,
L858R, exon 20 insertions,
T790M, S768I, and L861Q |

For manual sample preparation, FFPET specimens are processed using the cobas® DNA Sample Preparation Kit and plasma specimens are processed using the cobas® cfDNA Sample Preparation Kit. The cobas z 480 analyzer is used for automated amplification and detection.

# III. CONTRAINDICATIONS

There are no known contraindications.

{2}

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 3

# IV. WARNINGS AND PRECAUTIONS

The warnings and precautions can be found in the cobas® EGFR Mutation Test v2 labeling.

# V. DEVICE DESCRIPTION

The cobas® EGFR Mutation Test v2 is based on two major processes: (1) manual sample preparation to obtain DNA from FFPET or plasma; and (2) PCR amplification and detection of target DNA using complementary primer pairs and oligonucleotide probes labeled with fluorescent dyes.

The cobas® EGFR Mutation Test v2 is comprised of the following:

1. The cobas® EGFR Mutation Test v2 kit provides reagents for automated real-time PCR amplification and detection of the EGFR mutations.
2. The cobas® DNA Sample Preparation Kit provides reagents for manual specimen preparation to obtain genomic DNA from formalin-fixed, paraffin-embedded tumor tissue (FFPET).
3. The cobas® cfDNA Sample Preparation Kit provides reagents for manual specimen preparation to obtain cfDNA from plasma.

Two external run controls are provided and the EGFR exon 28 wild-type (WT) allele serves as an internal, full process control.

## A. Specimen Preparation – Plasma

Plasma specimens are processed and cfDNA is isolated using the cobas® cfDNA Sample Preparation Kit, a generic manual specimen preparation based on nucleic acid binding to glass fibers. The components of the cobas® cfDNA Sample Preparation Kit are identical to those of the cobas® DNA Sample Preparation Kit, with the exception of the High Pure Extension Assembly (HPEA) units.

Two milliliters (mL) of plasma are processed with a protease and chaotropic lysis/binding buffer that protects the cfDNA from DNases. Subsequently, isopropanol is added to the binding mixture that is then centrifuged through a column with a glass fiber filter insert. During centrifugation, the cfDNA is bound to the surface of the glass fiber filter. Unbound substances, such as salts, proteins and other impurities, are removed by centrifugation. The adsorbed nucleic acids are washed and then eluted with an aqueous solution.

## B. PCR Amplification and Detection

### Target Selection and Amplification

{3}

The cobas® EGFR Mutation Test v2 kit uses primers that define specific base-pair sequences for each of the targeted mutations. For the exon 19 deletion mutations, sequences ranging from 125 to 141 base pairs are targeted; for the L858R substitution mutation in exon 21, a 138 base pair sequence is targeted; for the T790M substitution mutation in exon 20, a 118 base pair sequence is targeted; for the G719X substitution mutation in exon 18, sequences ranging from 104 to 106 base pairs are targeted; for the S768I substitution mutation in exon 20, a 133 base pair sequence is targeted; for the exon 20 insertion mutations, sequences ranging from 125 to 143 base pairs are targeted; for the L861Q substitution mutation in exon 21, a 129 base pair sequence is targeted; for the internal control in exon 28, an 87 base pair sequence is targeted. Amplification occurs only in the regions of the EGFR gene between the primers; the entire EGFR gene is not amplified.

The cobas® EGFR Mutation Test v2 uses allele-specific PCR (AS-PCR) chemistry for amplification and detection. The selected AS-PCR primers specifically amplify the targeted mutant sequences over the WT sequences and/or other human DNA. The cobas® EGFR Mutation Test v2 is designed to use three master mix (MMx) reagents which are run in three separate wells. The number and types of primers and probes differ based on the particular target(s). The cobas® EGFR Mutation Test v2 detects the following EGFR mutations in exons 18, 19, 20, and 21:

Table 2. EGFR Mutations Detected by the cobas® EGFR Mutation Test v2

| Exon | EGFR Mutation | EGFR Nucleic Acid Sequence | COSMIC ID1 |
| --- | --- | --- | --- |
| Exon 18 | G719X | 2156G>C | 6239 |
| | | 2155G>A | 6252 |
| | | 2155G>T | 6253 |
| Exon 19 | Ex. 19del | 2240_2251del12 | 6210 |
| | | 2239_2247del9 | 6218 |
| | | 2238_2255del18 | 6220 |
| | | 2235_2249del15 | 6223 |
| | | 2236_2250del15 | 6225 |
| | | 2239_2253del15 | 6254 |
| | | 2239_2256del18 | 6255 |
| | | 2237_2254del18 | 12367 |
| | | 2240_2254del15 | 12369 |
| | | 2240_2257del18 | 12370 |
| | | 2239_2248TTAAGAGAAG>C | 12382 |
| | | 2239_2251>C | 12383 |
| | | 2237_2255>T | 12384 |
| | | 2235_2255>AAT | 12385 |
| | | 2237_2252>T | 12386 |
| | | 2239_2258>CA | 12387 |
| | | 2239_2256>CAA | 12403 |
| | | 2237_2253>TTGCT | 12416 |
| | | 2238_2252>GCA | 12419 |
| | | 2238_2248>GC | 12422 |
| | | 2237_2251del15 | 12678 |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{4}

| Exon | EGFR Mutation | EGFR Nucleic Acid Sequence | COSMIC ID1 |
| --- | --- | --- | --- |
| | | 2236_2253del18 | 12728 |
| | | 2235_2248>AATTC | 13550 |
| | | 2235_2252>AAT | 13551 |
| | | 2235_2251>AATTC | 13552 |
| | | 2253_2276del24 | 13556 |
| | | 2237_2257>TCT | 18427 |
| | | 2238_2252del15 | 23571 |
| | | 2233_2247del15 | 26038 |
| Exon 20 | S768I | 2303G>T | 6241 |
| | T790M | 2369C>T | 6240 |
| | Ex. 20ins | 2307_2308ins9GCCAGCGTG | 12376 |
| | | 2319_2320insCAC | 12377 |
| | | 2310_2311insGGT | 12378 |
| | | 2311_2312ins9GCGTGGACA | 13428 |
| | | 2309_2310AC>CCAGCGTGGAT | 13558 |
| Exon 21 | L858R | 2573T>G | 6224 |
| | | 2573_2574TG>GT | 12429 |
| | L861Q | 2582T>A | 6213 |

Catalogue of Somatic Mutations in Cancer (COSMIC), 2011, v.51. http://www.sanger.ac.uk/genetics/CGP/cosmic.

MMx1 (first amplification reaction) contains:

- Fourteen AS-PCR primers, one common primer, and one common probe are used to detect the Exon 19 deletion and complex (defined as the combination of a deletion and an insertion) mutations.
- One AS-PCR primer, one common primer, and one common probe are used to detect the S768I mutation.

MMx2 (second amplification reaction) contains:

- One AS-PCR primer, one common primer, and one common probe are used to detect the L858R mutation.
- One AS-PCR primer, one common primer, and one common probe are used to detect the T790M mutation.

MMx3 v2 (third amplification reaction) contains:

- Three AS-PCR primers, one common primer, and one common probe are used to detect G719X mutations.
- Three AS-PCR primers, one common primer, and one common probe are used to detect Exon 20 insertion mutations.
- One AS-PCR primer, one common primer, and one common probe are used to detect the L861Q mutation.

A derivative of Thermus species Z05-AS1 DNA polymerase is utilized for target amplification. The PCR reaction mixture is heated to denature the DNA and expose the primer target sequences. As the mixture cools, the upstream and downstream

PMA P150047: FDA Summary of Safety and Effectiveness Data

{5}

primers anneal to the target DNA sequences. The Z05-AS1 DNA polymerase, in the presence of divalent metal ion and excess dNTPs, extends each annealed primer, thus synthesizing a second DNA strand. This completes the first cycle of PCR, yielding a double-stranded DNA copy, which includes the targeted base-pair regions of the EGFR gene. This process is repeated for a number of cycles, with each cycle effectively doubling the amount of amplicon DNA.

Selective amplification of target nucleic acid from the specimen is achieved in the cobas® EGFR Mutation Test v2 by the use of AmpErase® (uracil-N-glycosylase) enzyme and deoxyuridine triphosphate (dUTP), which are included in the Master Mix reagents. The AmpErase enzyme recognizes and catalyzes the destruction of DNA strands containing deoxyuridine, but not DNA containing thymidine. Deoxyuridine is always present in the amplicons due to the use of dUTP as one of the nucleotide triphosphates in the Reaction Mix reagent; therefore, only the amplicons contain deoxyuridine. The AmpErase® enzyme is inactive at temperatures above 55°C, i.e., throughout the thermal cycling steps, and therefore does not destroy the target amplicon.

## Automated Real-time Detection

The cobas® EGFR Mutation Test v2 utilizes real-time PCR technology. Each target-specific, oligonucleotide probe in the reaction is labeled with a fluorescent dye that serves as a reporter, and with a quencher molecule that absorbs (quenches) fluorescent emissions from the reporter dye within an intact probe. During each cycle of amplification, a probe complementary to the single-stranded DNA sequence in the amplicon binds and is subsequently cleaved by the 5′ to 3′ nuclease activity of the Z05-AS1 DNA polymerase. Once the reporter dye is separated from the quencher by this nuclease activity, fluorescence of a characteristic wavelength can be measured when the reporter dye is excited by the appropriate spectrum of light. Two different reporter dyes are used to label the mutations targeted by the test. Amplification of the targeted EGFR sequences are detected independently across three reactions by measuring fluorescence at the two characteristic wavelengths in dedicated optical channels.

## Instrument and Software

The cobas® 4800 system is controlled by the cobas® 4800 system software, which provides the core software engines and user interfaces. This core system software was designed to allow multiple assays to be performed on the system using assay specific analysis package software (ASAP). The cobas® z480 analyzer component of the test system also has its own internal instrument control software, which is driven by the core software.

A dedicated Control Unit computer runs the cobas® 4800 system software and provides an interface to the cobas® z 480 and Laboratory Information System (LIS). The computer also processes the fluorescent signals with the assay specific analysis

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 6

{6}

package (ASAP) and stores the test results in a controlled database. The complete system allows a user to create a test work order for each specimen either manually or automatically when connected to a LIS. A software wizard guides the user through the necessary steps to perform a run, which includes cobas z 480 maintenance handling, test selection, specimen ID entry, reagent and microwell plate barcode entry, microwell plate loading and run start.

The cobas® 4800 system tracks each specimen during processing and analysis on the cobas z 480 analyzer. Once the thermal run is complete the ASAP software processes the fluorescence data using data analysis algorithms, assesses the validity of the controls and determines the results using the assay specific result interpretation logic. The software then provides the results to the user in three formats: a printable PDF results report, a GUI based result viewer and a result export file that can be exported to the LIS.

The cobas® 4800 system software includes the assay-specific cobas® 4800 EGFR Analysis Package (AP) software, which contains an algorithm to determine sample results and run validity. The analysis package used to analyze the EGFR results from cfDNA specimens isolated from plasma differs from the one used to analyze EGFR results from DNA specimens isolated from FFPET specimens. The overall cobas® 4800 system components are shown in the diagram below:

Figure 1. cobas® 4800 System

The final version of the cobas® 4800 core system software is v2.1.0 or v2.2.0 and the ASAP software used to analyze all studies in this PMA was EGFR Plasma P2 AP

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 7

{7}

v1.0.1.1555. The final commercial ASAP software is EGFR Plasma P2 Analysis Package v1.0.1.1567.

# Interpretation of Results

If the run is valid, then the cycle threshold (Ct) or CtR (relative cycle threshold) values for each sample will be evaluated against acceptable ranges for each channel. The CtR value is determined by calculating the difference between the mutations observed Ct and the corresponding Internal Control (IC) Ct value from the same Master Mix. Ct values are not available to the user. The tables below summarize how the individual amplification Master Mix results are combined to provide an overall result.

Table 3. Individual Amplification Master Mix Results to Overall Reported Results.

| MMx1 Result | MMx2 Result | MMx3 v2 Result | Reported Result |
| --- | --- | --- | --- |
| Valid, No Mutation Detected | Valid, No Mutation Detected | Valid, No Mutation Detected | Valid, No Mutation Detected |
| Valid, No Mutation Detected | Valid, No Mutation Detected | Valid, Mutation Detected | Valid, Mutation Detected |
| Valid, No Mutation Detected | Valid, No Mutation Detected | Invalid | Invalid |
| Valid, No Mutation Detected | Valid, Mutation Detected | Valid, No Mutation Detected | Valid, Mutation Detected |
| Valid, No Mutation Detected | Valid, Mutation Detected | Valid, Mutation Detected | Valid, Mutation Detected |
| Valid, No Mutation Detected | Valid, Mutation Detected | Invalid | Invalid |
| Valid, No Mutation Detected | Invalid | Valid, No Mutation Detected | Invalid |
| Valid, No Mutation Detected | Invalid | Valid, Mutation Detected | Invalid |
| Valid, No Mutation Detected | Invalid | Invalid | Invalid |
| Valid, Mutation Detected | Valid, No Mutation Detected | Valid, No Mutation Detected | Valid, Mutation Detected |
| Valid, Mutation Detected | Valid, No Mutation Detected | Valid, Mutation Detected | Valid, Mutation Detected |
| Valid, Mutation Detected | Valid, No Mutation Detected | Invalid | Invalid |
| Valid, Mutation Detected | Valid, Mutation Detected | Valid, No Mutation Detected | Valid, Mutation Detected |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{8}

Table 4. Result Interpretation of the cobas® EGFR Mutation Test v2

| Test Result | Mutation Result** | Interpretation |
| --- | --- | --- |
| Mutation Detected (MD) | Ex.19del
L858R
T790M
S768I
G719X
Ex. 20ins
L861Q
(More than one mutation may be present) | Mutation detected in specified targeted EGFR region. |
| No Mutation Detected* (NMD) | N/A | No mutation detected in targeted EGFR regions |
| Invalid | N/A | Specimen result is invalid. |
| Failed | N/A | Failed run due to hardware or software failure. |

* A No Mutation Detected (NMD) result does not preclude the presence of a mutation in the targeted EGFR regions, because results depend on percent mutant sequences, adequate specimen integrity, absence of inhibitors, and sufficient DNA to be detected.
**Italicized mutation results consist of mutations that are intended for analytical detection only for both FFPET and plasma specimens and the T790M mutation is for analytical detection only when identified in plasma specimens.

PMA P150047: FDA Summary of Safety and Effectiveness Data

{9}

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 10

# Test Controls

One EGFR mutant control and one EGFR negative control are provided. The EGFR WT allele located on exon 28 serves as an internal, full process control.

1. **EGFR Mutant Control**: The Mutant Control is a blend of seven DNA plasmids containing specified EGFR mutation sequences and cell line DNA that is WT for EGFR. The Mutant Control is composed of plasmids representing the most frequently observed mutation for each mutation class detected by the test. The Mutant Control will be included in every run and will serve as a process control for amplification and detection. The Mutant Control must yield Cycle threshold (Ct) values for all targeted mutations and the Internal Control (IC) within the respective acceptable ranges for the run to be considered valid.

2. **EGFR Negative Control**: The Negative Control is a full process contamination control for a given test batch of specimens. The Negative Control consists of a blank vial containing no specimen (specimen diluent only) which is processed through specimen preparation and the resulting eluate is subsequently used for amplification and detection. The Negative Control Ct values must be either not detected or greater than the pre-established Ct maximum value for all targeted mutation groups and the IC for the run to be considered valid.

3. **EGFR WT Internal Control (IC)**: The Internal Control in EGFR exon 28 from test specimens serves as a full process control. This control ensures that every step of the process from specimen preparation to amplification and detection has been completed successfully.

# VI. ALTERNATIVE PRACTICES AND PROCEDURES

There are no other FDA-cleared or -approved alternatives for the testing of plasma (or formalin-fixed, paraffin-embedded NSCLC tissue) for EGFR mutation status in the selection of patients who are eligible for treatment with TARCEVA® (erlotinib) or TAGRISSO™ (osimertinib).

# VII. MARKETING HISTORY

The cobas® EGFR Mutation Test (v1) was introduced into the United States starting on May 14, 2013. The cobas® EGFR Mutation Test (v1) is commercially available in the following countries: Argentina, Australia, Austria, Belgium, Brazil, Bulgaria, Canada, Chile, China, Colombia, Costa Rica, Croatia, Cyprus, Czech Republic, Denmark, Ecuador, Estonia, Finland, France, Germany, Greece, Guatemala, Hong Kong, Hungary, Iceland, India, Indonesia, Ireland, Italy, Japan, Korea, Latvia, Liechtenstein, Lithuania, Luxembourg, Malaysia, Malta, Mexico, Netherlands, New Zealand, Nicaragua, Norway, Pakistan, Panama, Peru, Philippines, Poland, Portugal, Romania, Singapore, Slovakia,

{10}

Slovenia, South Africa, Spain, Sweden, Switzerland, Taiwan, Thailand, Turkey, United Arab Emirates, United Kingdom, United States, Uruguay, Venezuela, Vietnam.

The cobas® EGFR Mutation Test (v1) was replaced by the cobas® EGFR Mutation Test v2, which was introduced into the United States starting on November 13, 2015 for use with FFPET specimens. The cobas® EGFR Mutation Test v2 for use with FFPET and plasma specimens is commercially available in Australia, Austria, Belgium, Bolivia, Bulgaria, Colombia, Croatia, Cyprus, Czechia, Denmark, Ecuador, Estonia, Finland, France, Germany, Greece, Guatemala, Hong Kong, Hungary, Iceland, Ireland, Italy, Latvia, Liechtenstein, Lithuania, Luxembourg, Malta, Netherlands, New Zealand, Norway, Panama, Paraguay, Poland, Portugal, Romania, Singapore, Slovakia, Slovenia, Spain, Sweden, Switzerland, Thailand, Turkey, United Arab Emirates, United Kingdom, Uruguay, and Vietnam. The cobas® EGFR Mutation Test v2 for use with FFPET specimens is also commercially available in Japan.

Neither version of the cobas® EGFR Mutation Test for either specimen type, has been withdrawn from the market for reasons related to safety and effectiveness.

## 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 cobas® EGFR Mutation Test v2 results and subsequently improper patient management decisions in NSCLC treatment. No adverse events were reported in connection with the studies used to support this PMA as the studies were performed retrospectively using banked samples.

## IX. SUMMARY OF PRECLINICAL STUDIES

### A. Laboratory Studies

For the non-clinical studies described below, the percentage of mutation present in the prepared plasma specimens was assessed using a validated next generation sequencing (NGS) method. Due to the lack of sufficient volume of clinical specimens, some of the studies used contrived samples, which consisted of sheared cell line DNA diluted into healthy donor (HD) EDTA plasma. A commutability study was conducted to demonstrate comparable performance of sheared cell line DNA diluted in HD plasma as compared to NSCLC EDTA plasma. Clinical specimens or clinical specimen pools were used to assess analytical accuracy, confirm the estimated limit of detection (LoD) and evaluate assay reproducibility.

Contrived samples were prepared using cell line DNA stocks, purchased from a commercial source, sheared to an average size of approximately 220 bp using focused-ultrasonication. DNA fragment sizes were confirmed by an orthogonal method. The sheared cell line DNA stocks were then quantitated against a plasmid-based standard curve in order to dilute each EGFR mutation to pre-specified input cp/mL levels for each of the studies. The cell lines used include the predominant

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 11

{11}

mutation for each mutation group detected by the cobas® EGFR Mutation Test v2 using plasma specimens (cobas® EGFR Plasma Test v2).

The sheared cell line DNA with multiple mutation targets were combined to allow for the evaluation of more than one EGFR mutation target per sample. The combinations included in each sample are shown in the table below. In each non-clinical study, contrived samples at approximately 3x LoD (in cp/mL) per mutation were evaluated unless otherwise specified.

Table 5. EGFR Mutation Combinations for Contrived Samples
| | First EGFR Mutation | Second EGFR Mutation |
| --- | --- | --- |
| Contrived Sample 1 | Ex. 19del | Ex. 20ins |
| Contrived Sample 2 | G719X | S768I |
| Contrived Sample 3 | L861Q | T790M |
| Contrived Sample 4 | L858R | None |

1. Correlation with Reference Method Using Clinical Samples

The analytical accuracy of the cobas® EGFR Plasma Test v2 was assessed by comparing its results to a validated quantitative NGS method using banked clinical samples. Percent mutation present in specimens was determined for all specimens used to demonstrate performance.

In this study, plasma specimens collected at baseline screening from patients from three different studies in stage IIIb/IV NSCLC patients: the ASPIRATION (ML25637/NCT01310036)¹, MetMab (Genentech study GO27821/NCT01496742), and MetLung (GO27761/NCT01456325) studies. The available specimens from these studies (hereafter referred to as the ASPIRATION cohort) were tested by both the cobas® EGFR Plasma Test v2 and a validated NGS method. The specimens from the MetMab and MetLung studies were included to supplement the samples from the ASPIRATION study to ensure a sufficient number of samples for statistical analysis. All patients from the MetMab study were NMD by the cobas® EGFR Tissue Test v1 and were used to supplement the tissue NMD study population.

Specimen Processing, Disposition, and Testing

FFPET specimens were processed according to the standard protocol. In the ASPIRATION study, patients were screened based on an EGFR clinical trial assay and not all samples were re-tested by the cobas® EGFR Tissue Test v1 during the study. Tissue samples from MetLung and MetMab were tested with the cobas® EGFR Tissue Test v1 during the studies and results were not used to enroll the patients. No additional tissue samples were tested under this study protocol.

¹ (Study sponsor’s protocol/ClinicalTrials.gov registration number)

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 12

{12}

Plasma specimens were sent to one of the participating central laboratories after written informed consent was confirmed. cfDNA was isolated manually using cobas® cfDNA Sample Preparation Kit. Testing was performed according to the instructions for use for the cobas® EGFR Plasma Test v2.

Testing was performed using two different lots of the cobas® cfDNA Sample Preparation Kit and one lot of the cobas® EGFR Mutation Test v2 Kit on three separate cobas z 480 Analyzers.

## Accountability of the ASPIRATION Cohort

A total of 411 plasma samples from the ASPIRATION cohort were tested by the cobas® EGFR Plasma Test v2. The specimens included 231 specimens from the ASPIRATION study, 110 specimens from the MetMab study, and 70 specimens from the MetLung study. Due to different study designs, the cobas® EGFR Tissue Test v1 results were only available for 340/411 patients with plasma samples. However, only 128 samples with valid paired cobas® EGFR Tissue Test v1 result had a plasma volume of 2.0 mL. Since the cobas® EGFR Plasma Test v2 is validated for 2.0 mL plasma samples, the primary analysis focused on only those samples with a 2.0 mL volume.

## Agreement to Reference Method:

## Unadjusted Agreement:

Based on the plasma samples tested from the ASPIRATION Cohort (with 2.0 mL plasma volume), the agreements of the cobas® EGFR Plasma Test v2 and the NGS method for detection of EGFR Ex. 19del and L858R mutations are presented in the tables below. Among 128 samples with a 2.0 mL plasma volume, a total of 32 samples had MD and 95 had NMD results by the NGS method. The tables below summarize agreement calculations with associated sensitivity analyses.

Table 6. Unadjusted Agreement between cobas® EGFR Plasma Test v2 and NGS for Detection of EGFR Mutations* in Aggregate with 2.0 mL Plasma Volume

| | NGS (Plasma) | | | | |
| --- | --- | --- | --- | --- | --- |
| cobas® EGFR Plasma Test v2 | MD | NMD | Invalid | Not Tested | Total |
| MD | 28 | 3 | 0 | 0 | 31 |
| NMD | 4 | 92 | 1 | 0 | 97 |
| Invalid | 0 | 0 | 0 | 0 | 0 |
| Total | 32 | 95 | 1 | 0 | 128 |
| With only Valid Result | | | | | |
| PPA (95% CI) | 87.5% (71.9%, 95.0%) | | | | |
| NPA (95% CI) | 96.8% (91.1%, 98.9%) | | | | |
| PPV (95% CI) | 90.3% (75.1%, 96.7%) | | | | |
| NPV (95% CI) | 95.8% (89.8%, 98.4%) | | | | |

Sensitivity Analysis with Invalid Result as Discordant

PMA P150047: FDA Summary of Safety and Effectiveness Data

{13}

| | NGS (Plasma) | | | | |
| --- | --- | --- | --- | --- | --- |
| cobas® EGFR Plasma Test v2 | MD | NMD | Invalid | Not Tested | Total |
| PPA (95% CI) | 84.8% (69.1%, 93.3%) | | | | |
| NPA (95% CI) | 96.8% (91.1%, 98.9%) | | | | |
| PPV (95% CI) | 90.3% (75.1%, 96.7%) | | | | |
| NPV (95% CI) | 94.8% (88.5%, 97.8%) | | | | |

*Detection of either or both Ex. 19del and L858R Mutations
PPA = Positive Percent Agreement; NPA = Negative Percent Agreement;
PPV = Positive Predictive Value; NPV = Negative Predictive Value

Table 7. Unadjusted Agreement between cobas® EGFR Plasma Test v2 and NGS for Detection of Ex. 19del and L858R Mutations (with 2.0 mL Plasma Volume)

| | NGS (Plasma) | | | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | Ex. 19del | | | | | L858R | | | | |
| cobas® EGFR Plasma Test v2 | MD | NMD | Invalid | Not Tested | Total | MD | NMD | Invalid | Not Tested | Total |
| MD | 17 | 2 | 0 | 0 | 19 | 11 | 1 | 0 | 0 | 12 |
| NMD | 2 | 106 | 1 | 0 | 109 | 2 | 113 | 1 | 0 | 116 |
| Invalid | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Total | 19 | 108 | 1 | 0 | 128 | 13 | 114 | 1 | 0 | 128 |
| With only Valid Result | | | | | | | | | | |
| PPA (95% CI) | 89.5% (68.6%, 97.1%) | | | | | 84.6% (57.8%, 95.7%) | | | | |
| NPA (95% CI) | 98.1% (93.5%, 99.5%) | | | | | 99.1% (95.2%, 99.8%) | | | | |
| PPV (95% CI) | 89.5% (68.6%, 97.1%) | | | | | 91.7% (64.6%, 98.5%) | | | | |
| NPV (95% CI) | 98.1% (93.5%, 99.5%) | | | | | 98.3% (93.9%, 99.5%) | | | | |
| Sensitivity Analysis with Invalid Result as Discordant | | |
| --- | --- | --- |
| PPA (95% CI) | 85.0% (64.0%, 94.8%) | 78.6% (52.4%, 92.4%) |
| NPA (95% CI) | 98.1% (93.5%, 99.5%) | 99.1% (95.2%, 99.8%) |
| PPV (95% CI) | 89.5% (68.6%, 97.1%) | 91.7% (64.6%, 98.5%) |
| NPV (95% CI) | 97.2% (92.2%, 99.1%) | 97.4% (92.7%, 99.1%) |

# Adjusted Agreement:

Since the plasma samples tested in this accuracy study were not representative of the intended use population due to enrichment of patients with a tissue MD result as compared to patients with tissue NMD result, the agreement statistics were recalculated based on different tissue MD prevalences (for Ex. 19del and/or L858R mutations) and the results are shown in the table below. PPA ranged from $76.6\%$ to $83.7\%$ and NPA ranged from $97.9\%$ to $98.2\%$ based on $10 - 20\%$ tissue EGFR prevalence in a Caucasian population. Based on $30 - 40\%$ tissue EGFR prevalence in an Asian population, the PPA ranged from $86.3\%$ to $87.7\%$ and NPA ranged from $97.3\%$ to $96.6\%$ , respectively.

PMA P150047: FDA Summary of Safety and Effectiveness Data

{14}

Table 8. Agreement Between cobas® EGFR Plasma Test v2 vs. NGS for Detection of EGFR* in Aggregate Based on Different Tissue Mutation Prevalence (with 2.0 mL Plasma Sample) for the ASPIRATION Cohort

| Tissue Mutation Prevalence | Agreement between cobas® EGFR Plasma Test v2 and NGS | | | |
| --- | --- | --- | --- | --- |
| | PPA (95% CI) | NPA (95% CI) | PPV (95% CI) | NPV (95% CI) |
| 10% | 76.6 (65.2, 89.2) | 98.2 (97.1, 99.3) | 78.3 (67.9, 90.8) | 98.0 (96.9, 99.1) |
| 20% | 83.7 (76.1, 91.6) | 97.9 (96.5, 99.0) | 86.7 (79.2, 93.8) | 97.3 (96.0, 98.6) |
| 30% | 86.3 (79.8, 92.4) | 97.3 (95.7, 98.7) | 89.2 (83.2, 94.4) | 96.5 (94.7, 98.1) |
| 40% | 87.7 (82.6, 92.7) | 96.6 (94.8, 98.2) | 90.5 (85.7, 95.0) | 95.5 (93.7, 97.3) |
| 50% | 88.6 (84.5, 92.8) | 95.8 (93.9, 97.7) | 91.3 (87.4, 95.3) | 94.3 (92.2, 96.5) |

* Detection of Ex. 19del and/or L858R substitution mutations

Comparison of the ENSURE and the ASPIRATION Cohorts

Plasma specimens from patients enrolled in ENSURE were used to demonstrate the clinical utility of the cobas® EGFR Plasma Test v2 as a CDx for TARCEVA® (see Section X). The cobas® EGFR Tissue Test v1 was used to enroll patients in ENSURE. Due to insufficient plasma volumes from patients screened for ENSURE, plasma specimens from the ASPIRATION cohort were used to demonstrate analytical accuracy of the cobas® EGFR Plasma Test v2.

The transportability of analytical accuracy from the ASPIRATION cohort to ENSURE was established by a) demonstrating that the baseline demographics and clinical characteristics information are similar among the enrolled patients who were positive by the cobas® EGFR Tissue Test v1 in the two studies; b) evaluating whether the PPA and NPA between cobas® EGFR Plasma Test v2 vs. cobas® EGFR Tissue Test v1 results (using tissue result as reference) were comparable between samples from the two studies; and c) demonstrating that the propensity score is comparable for enrolled patients in the ENSURE and the ASPIRATION cohorts.

a. Patient Demographics

Demographics and clinical characteristics were compared among the enrolled patients with a plasma sample available between the ENSURE and ASPIRATION cohorts by a propensity score method. The propensity score is the conditional probability that a patient is assigned to the ENSURE trial given the specific characteristic of that patient. The stage of disease and histology result was not included as an independent variable, because only stage III patients were enrolled in ENSURE. Similarly, the propensity was compared between:

- Enrolled patients in ENSURE for whom plasma samples were available and for whom plasma samples were not available;
- Enrolled patients in ASPIRATION cohort for whom plasma samples were available and for whom plasma samples were not available;
- Enrolled patients with available plasma samples in the ENSURE and ASPIRATION cohorts.

PMA P150047: FDA Summary of Safety and Effectiveness Data

{15}

The baseline demographic and clinical covariates of patients with a MD result by the cobas® EGFR Tissue Test v1 and a plasma sample of 2mL from the ENSURE and ASPIRATION cohorts were compared. There were some covariates which were statistically significantly different between the two studies in univariate analysis, including age; smoking history; BMI; tumor stage; histology; and ECOG score. Due to the imbalance of some baseline covariates between the two studies, a propensity score method (logistic regression model) was used by including all seven baseline covariates except the disease stage in the model, and propensity scores were generated for each patient. It was found that none of the covariates was different between the two studies at a significance level of 0.05 after adjusting the propensity score quintiles, indicating that each covariate was balanced within the propensity score strata.

Table 9. Demographics and Baseline Clinical Characteristics of Tissue EGFR Mutation Positive Patients with Plasma Specimen of ${2.0}\mathrm{\;{mL}}$ from ENSURE and ASPIRATION Cohort

| Characteristics | ENSURE (n=180) | ASPIRATION Cohort (n=48)* | p-value | |
| --- | --- | --- | --- | --- |
| | | | Before | After |
| Age (Years) | | | | |
| <65 Years | 143 (79.4%) | 25 (52.1%) | 0.0001 | 0.4783 |
| ≥ 65 Years | 37 (20.6%) | 23 (47.9%) | | |
| Sex (%) | | | | |
| Male | 67 (37.2%) | 22 (45.8%) | 0.2772 | 0.8422 |
| Female | 113 (62.8%) | 26 (54.2%) | | |
| Smoking History (%) | | | | |
| Never Smoked | 129 (71.7%) | 23 (47.9%) | 0.0019 | 0.9885 |
| Past/Current Smoker | 51 (28.3%) | 25 (52.1%) | | |
| BMI (Kg/m2) (Mean ± SD) | 22.9 ± 3.6 | 24.3 ± 3.6 | 0.0261 | 0.5946 |
| Tumor Stage (%) | | | | |
| Stage IIIB | 15 (8.3%) | 0 (0.0 %) | 0.0038 | N/A |
| Stage IV | 165 (91.7%) | 46 (95.8%) | | |
| Missing | 0 (0.0 %) | 2 (4.2%) | | |
| Histology (%) | | | | |
| Adenocarcinoma | 170 (94.4%) | 45 (93.8%) | 0.0328 | 0.9260 |
| Other | 10 (5.6%) | 1 (2.1%) | | |
| Missing | 0 (0.0 %) | 2 (4.2%) | | |
| ECOG Score (%) | | | | |
| 0 | 26 (14.4%) | 20 (41.7%) | 0.0006 | 0.5502 |
| 1 | 142 (78.9%) | 26 (54.2%) | | |
| 2 | 9 (5.0%) | 1 (2.1%) | | |
| Missing | 3 (1.7%) | 1 (2.1%) | | |
| EGFR Mutation (%) by Tissue | | | | |
| Ex. 19del | 98 (54.4%) | 29 (60.4%) | 0.4592 | 0.9678 |
| Ex. 20del | 10 (5.6%) | 1 (2.1%) | | |
| Missing | 0 (0.0 %) | 2 (4.2%) | | |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{16}

| Characteristics | ENSURE (n=180) | ASPIRATION Cohort (n=48)* | p-value | |
| --- | --- | --- | --- | --- |
| | | | Before | After |
| L858R | 82 (45.6%) | 19 (39.6%) | | |

* The 48 patients include 39 patients from MetLung study and 9 from the ASPIRATION study.
Note: Tissue EGFR Mutation positive referred as mutation detected for either Ex. 19del or L858R mutation as determined by cobas® EGFR Tissue Test v1.
Note: N/A = Not Applicable.
Note: Patients with missing covariate information were not used in propensity score analysis.
Note: Covariate tumor stage was not included in the propensity score analysis as stage III patients were only enrolled in the ENSURE study.
Note: p-value was calculated before and after adjusting the propensity score quintiles.

## b. Agreement between plasma and tissue samples

The agreement between plasma and tissue samples by the cobas® EGFR Plasma Test v2 and the cobas® EGFR Tissue Test v1 was performed for detection of exon 19 deletion and L858R mutations in aggregate (for plasma samples with 2.0 mL volume) and a three-way agreement between the two specimen types and NGS using plasma samples. These comparisons are presented in the table below.

Table 10. Agreement between cobas® EGFR Plasma Test v2 and cobas® EGFR Tissue Test v1 for Detection of EGFR Mutations (with 2.0 mL Plasma Sample)

| cobas® Plasma Test v2 | cobas® EGFR Tissue Test v1 | | Total |
| --- | --- | --- | --- |
| | MD | NMD | |
| MD | 30 | 1 | 31 |
| NMD | 18 | 79 | 97 |
| Total | 48 | 80 | 128 |
| | | | |
| PPA (95% CI) | 62.5% (48.4%, 74.8%) | | |
| NPA (95% CI) | 98.8% (93.3%, 99.8%) | | |
| PPV (95% CI) | 96.8% (83.8%, 99.4%) | | |
| NPV (95% CI) | 81.4% (72.6%, 87.9%) | | |

A three-way agreement was provided to compare the EGFR mutation status in aggregate between the cobas® EGFR Tissue Test v1, the cobas® EGFR Plasma Test v2, and NGS. These results are shown in the table below.

Table 11. Three-way Agreement among cobas® EGFR Tissue Test v1, cobas® EGFR Plasma Test v2, and NGS for Detection of EGFR Mutations (with 2.0 mL Plasma Sample) from the ASPIRATION Cohort

| cobas® Tissue Test v1 | cobas® EGFR Plasma Test v2 | NGS (Plasma) | | | |
| --- | --- | --- | --- | --- | --- |
| | | MD | NMD | Invalid | Total |
| MD | MD | 28 | 2 | 0 | 30 |
| | NMD | 3 | 14 | 1 | 18 |
| | Invalid | 0 | 0 | 0 | 0 |
| | Total | 31 | 16 | 1 | 48 |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{17}

| cobas® Tissue Test v1 | cobas® EGFR Plasma Test v2 | NGS (Plasma) | | | |
| --- | --- | --- | --- | --- | --- |
| | | MD | NMD | Invalid | Total |
| NMD | MD | 0 | 1 | 0 | 1 |
| | NMD | 1 | 78 | 0 | 79 |
| | Invalid | 0 | 0 | 0 | 0 |
| | Total | 1 | 79 | 0 | 80 |

c. Comparison of Agreements Between ASPIRATION and ENSURE Cohorts
The agreements, NPA and PPA, were compared between the tissue and plasma results for the patients from each study cohort and p-values calculated. The p-value for PPA was also adjusted based on the calculated propensity scores and are shown below.

Table 12. Comparison of NPA and PPA of cobas® EGFR Plasma Test v2 vs. cobas® EGFR Tissue Test v1 (with 2.0 mL Plasma Sample)

| Sample Source | Total Tissue NMD Patients with Plasma Samples | Results by cobas® EGFR Plasma Test v2 | | | NPA (%) | p-value | |
| --- | --- | --- | --- | --- | --- | --- | --- |
| | | MD | NMD | Invalid | | | |
| ENSURE | 229 | 4 | 217 | 8 | 98.2% (95.4%, 99.3%) | 1.0000 | |
| ASPIRATION Cohort | 80 | 1 | 79 | 0 | 98.8% (93.3%, 99.8%) | | |
| | Total Tissue MD Patients with Plasma Samples | | | | PPA (%) | Before* | After* |
| ENSURE | 212 | 161 | 49 | 2 | 76.7% (70.5%, 81.9%) | 0.043 | 0.206 |
| ASPIRATION Cohort | 48 | 30 | 18 | 0 | 62.5% (48.4%, 74.8%) | | |

* p-value was calculated before and after adjusting the propensity score quintiles.

After adjusting based on propensity scores for those in which a mutation was detected, there is no significant difference in the populations based on the p-values.

d. Rare Mutations:
All five rare mutations [in tyrosine kinase inhibitor (TKI)-naïve patients], targeted by the cobas® EGFR Plasma Test v2 (Ex. 20ins, L861Q, S768I, G719X, and T790M) were detected and confirmed by the validated NGS method using plasma samples from the ASPIRATION Cohort and four rare mutations targeted by the assay (Ex. 20ins, S768I, G719X, and T790M) were detected in the ENSURE clinical study.

- ASPIRATION Cohort
- Three Ex. 20ins mutations were detected by the cobas® EGFR Plasma Test v2. Two were confirmed by NGS using plasma samples, but the

PMA P150047: FDA Summary of Safety and Effectiveness Data

{18}

remaining result could not be confirmed due to the lack of residual sample remaining for NGS testing.

- Two G719X mutations were detected by the cobas® EGFR Plasma Test v2. One sample was confirmed with G719X mutation while the other was not confirmed by NGS.

- G719X was detected in four other samples but not detected by the cobas® EGFR Plasma Test v2; due to the percent mutation of the samples being below the assay cut-off.

- One L861Q and one S768I were mutations detected by the cobas® EGFR Plasma Test v2 were confirmed by NGS.

- Six T790M mutations were detected by cobas® EGFR Plasma Test v2. Four were confirmed by NGS using plasma samples while the other two were not confirmed by NGS due to the percent mutation of the samples being below the NGS cut-off.

- ENSURE Study

- Three Ex. 20ins; three G719X; five S768I; one T790M; and eight L861Q mutations were detected by the cobas® EGFR Plasma Test v2; however, they could not be confirmed due to the lack of sufficient plasma available for NGS testing.

## 2. Contrived Sample Comparison

A study was performed to compare the performance of the cobas® EGFR Plasma Test v2 with intact cell line DNA and mechanically sheared cell line DNA (to an average size of approximately 220 bp) diluted in both HD EDTA plasma and NSCLC EGFR WT EDTA plasma. In two separate experiments, four panels were created with a combination of cell line DNA from Ex. 19del, L858R, and T790M mutations and tested over a total of eight levels each (10x, 5x, 1x (twice), 0.75x, 0.5x, 0.2x, 0.1x, and 0.03x LoD). A total of 20 replicates of each panel member were run and tested with each of two lots of the cobas® cfDNA Sample Preparation Kit in combination with two lots of the cobas® EGFR Mutation Test v2 kit. Ten replicates were split across ten runs to include one replicate of each level and condition per run. A total of 20 runs were performed with five operators and 18 cobas z 480 analyzers. Each experiment was reported separately prior to pooling. The hit rates for each specimen were determined and potential bias estimated using linear regression.

Table 13. Hit Rates for the Contrived Sample Comparison Study (n=20) with Pooled Results

| Target | DNA Type | Plasma Source | 5x LoD (Round 1) | 1x LoD (Round 1) | 1x LoD (Round 2) | 0.75x LoD (Round 2) | 0.5x LoD (Round 1) | 0.2x LoD (Round 2) | 0.1x LoD (Round 2) | 0.03x LoD (Round 2) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Ex. 19del | Intact | NSCLC | 100% (83.2-100) | 100% (83.2-100) | 95% (75.1-99.9) | 100% (83.2-100) | 100% (83.2-100) | 95% (75.1-99.9) | 100% (83.2-100) | 75% (50.9-91.3) |
| | | HD | 100% | 95% | 100% | 100% | 100% | 100% | 90% | 80% |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{19}

Table 14. $\mathbf{R}^2$ , Slope, and Intercept for the Contrived Sample Comparison with Pooled Results

| Target | DNA Type | Donor Type | R2 | Slope | 95% CI | p-value of t-test for Slope* | Intercept | 95% CI | p-value of t-test for Intercept* |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Ex 19del | Intact | NSCLC | 0.93 | -3.51 | -3.65 to -3.37 | 0.74 | 39.40 | 39.14 - 39.65 | 0.71 |
| | | HD | 0.90 | -3.45 | -3.62 to -3.28 | 0.87 | 39.39 | 39.08 - 39.70 | 0.74 |
| | Sheared | NSCLC | 0.92 | -3.47 | -3.62 to -3.31 | N/A | 39.31 | 39.02 - 39.60 | N/A |
| | | HD | 0.91 | -3.27 | -3.44 to -3.11 | 0.23 | 38.91 | 38.61 - 39.22 | 0.20 |
| L858R | Intact | NSCLC | 0.88 | -3.11 | -3.28 to -2.94 | 0.90 | 42.92 | 42.58 - 43.25 | 0.42 |
| | | HD | 0.83 | -2.89 | -3.09 to -2.69 | 0.29 | 42.80 | 42.40 - 43.19 | 0.67 |
| | Sheared | NSCLC | 0.86 | -3.09 | -3.28 to -2.90 | N/A | 42.66 | 42.28 - 43.04 | N/A |

10x LoD values not included as all were $100\%$ (83.2 - 100).

PMA P150047: FDA Summary of Safety and Effectiveness Data

{20}

| Target | DNA Type | Donor Type | R² | Slope | 95% CI | p-value of t-test for Slope* | Intercept | 95% CI | p-value of t-test for Intercept* |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | HD | 0.85 | -3.01 | -3.20 to -2.82 | 0.59 | 42.53 | 42.16 - 42.90 | 0.67 |
| T790M | Intact | NSCLC | 0.92 | -4.34 | -4.54 to -4.14 | 0.46 | 41.77 | 41.38 - 42.15 | 0.17 |
| | | HD | 0.91 | -4.46 | -4.68 to -4.24 | 0.97 | 42.45 | 42.02 - 42.89 | 0.77 |
| | Sheared | NSCLC | 0.92 | -4.47 | -4.67 to -4.27 | N/A | 42.36 | 41.97 - 42.74 | N/A |
| | | HD | 0.88 | -4.35 | -4.59 to -4.11 | 0.52 | 42.06 | 41.58 - 42.53 | 0.44 |

* Comparison to sheared cell line DNA in NSCLC EDTA plasma;
N/A = not applicable.
Note: p-value analysis is based on predicted values determined from the pooled regression analysis

Based on these analyses, the results demonstrate that the performance of the cobas® EGFR Plasma Test v2 is equivalent when using contrived specimens consisting of sheared or intact cell line DNA in a NSCLC or HD EDTA plasma background.

# 3. Analytical Sensitivity

# a. Analytical Sensitivity - Limit of Blank (LoB)

A panel of 33 unique EGFR WT EDTA HD plasma specimens, tested in duplicate, with each of three separately manufactured reagent lots, was used to establish the LoB. The LoB was defined as the 95th percentile of the Ct or CtR (depending on the specific mutation) results obtained with a blank sample.

LoB was determined using the nonparametric option as prescribed in Clinical and Laboratory Standards Institute (CLSI) guideline EP17-A2 for the reported CtR or Ct values of each lot and all lots combined for the EGFR WT samples. Any Ct result that was Not a Number (NaN) was converted to a value of 55 prior to calculating the CtR. The value of 55 corresponds to the total number of cycles in the PCR profile and is considered equivalent to a NaN result. For each lot and for all lots combined, the 5th percentile of the reported values for each channel was determined.

There were no invalid runs out of 146 total runs and 7 replicates gave invalid results due to either the IC not being detected or the Ct value being out of range. Of the invalid replicates, only three had sufficient residual plasma to allow for repeat testing and generated valid results upon repeat. The LoB was determined to be zero for all mutations tested.

# b. Analytical Sensitivity - Limit of Detection (LoD)

The study was performed by testing dilution panels constructed from cell line DNA that was sheared to $\sim 220$ bp, quantified against a plasmid standard curve, and diluted into EGFR WT EDTA plasma from HDs. Specimen blend

PMA P150047: FDA Summary of Safety and Effectiveness Data

{21}

panels were constructed and studied for the predominant mutations for Ex. 19del, L858R, T790M, Ex. 20ins, S768I, G719X, and L861Q mutations.

Six panels were created of two combinations of mutant DNA co-diluted in three separate dilution series. Combination 1 consisted of sheared cell line DNA for Ex. 19del, L858R, and T790M. Combination 2 consisted of sheared cell line DNA for S768I, L861Q, G719X, and Ex. 20ins. Three lots each of the cobas® cfDNA Sample Preparation Kit and the cobas® EGFR Mutation Test v2 kit were used in the study. Twenty-four replicates of each sample at each level were tested per lot for a total of 72 replicates. An extra replicate was tested at one level for one lot with Combination 2 resulting in a total of 73 total replicates.

The lowest level (cp/mL) that yielded at least a 95% MD rate for each mutation was established as the LoD. All mutations were detected at ≤100 cp/mL.

Table 15. Established LoD for each Predominant Mutation using Sheared Cell Line DNA in EDTA HD Plasma
| Mutation | Nucleic Acid Sequence | LoD (cp/mL) |
| --- | --- | --- |
| Ex. 19del | 2235-2249del15 | 75 |
| L858R | 2573 T>G | 100 |
| T790M | 2369 C>T | 100 |
| S768I | 2303 G>T | 25 |
| L861Q | 2582 T>A | 30 |
| G719X | 2156 G>A | 100 |
| Ex. 20ins | 2307_2308insGCCAGCGTG | 25 |

c. Analytical Sensitivity – Clinical Specimen Confirmation

An 11-member panel, consisting of pooled residual clinical plasma specimens along with mutant control and negative control samples were tested in duplicate by each of two operators at three separate sites. Three unique reagent lots were used across the three testing sites with two lots per site (i.e., Site 1, lots A and B; Site 2, lots A and C; and Site 3, lots B and C) and each site performed a total of 88 tests. The panel included five different mutations (three Ex. 19del, one L858R, and one T790M) which were run at the estimated LoD and 2x LoD, and one WT panel member.

Table 16. Limit of Detection Confirmation Panel
| Panel Member | Description of Mutations in Panel | Target Concentration (copies/mL) | Est. LoD by contrived specimens |
| --- | --- | --- | --- |
| 1 | 0 | WT | N/A |
| 2 | Ex 19del (2235_2249DEL15) | 75 | 75 |
| 3 | | 150 | |
| 4 | Ex 19del (2236_2250DEL15) | 75 | N/A |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{22}

| Panel Member | Description of Mutations in Panel | Target Concentration (copies/mL) | Est. LoD by contrived specimens |
| --- | --- | --- | --- |
| 5 | | 150 | |
| 6 | Ex 19del (2240_2257DEL18) | 75 | N/A |
| 7 | | 150 | |
| 8 | T790M | 100 | 100 |
| 9 | | 200 | |
| 10 | L858R | 100 | 100 |
| 11 | | 200 | |

Table 17. Overall Estimates of Agreement by Panel member

| Panel Member | Number of Valid Tests | Agreement N | Agreement % (95% CI)a |
| --- | --- | --- | --- |
| WT – N/A | 23 | 23 | 100 (85.2, 100.0) |
| Ex 19del 1 - 1x LoD | 24 | 24 | 100 (85.8, 100.0) |
| Ex 19del 1 - 2x LoD | 24 | 24 | 100 (85.8, 100.0) |
| Ex 19del 2 - 1x LoD | 23 | 23 | 100 (85.2, 100.0) |
| Ex 19del 2 - 2x LoD | 24 | 24 | 100 (85.8, 100.0) |
| Ex 19del 3 - 1x LoD | 24 | 24 | 100 (85.8, 100.0) |
| Ex 19del 3 - 2x LoD | 24 | 24 | 100 (85.8, 100.0) |
| T790M - 1x LoD | 24 | 23 | 95.8 (78.9, 99.9) |
| T790M - 2x LoD | 24 | 24 | 100 (85.8, 100.0) |
| L858R - 1x LoD | 24 | 24 | 100 (85.8, 100.0) |
| L858R - 2x LoD | 24 | 24 | 100 (85.8, 100.0) |

Note: Results are included as agreement when a valid test of Mutant Type panel member has a result of MD or when a valid test of WT panel member has a result of NMD.
a $95\%$ CI = 95% exact binomial confidence interval.

All 12 runs were valid with 262/264 valid test results. One replicate representing a WT call and one replicate of Ex. 19del 2 at 1x LoD generated invalid results due to no detection of the internal control or generation of a Ct value greater than the $\mathrm{Ct}_{\mathrm{max}}$ cut-off specification. These replicates were not repeated and therefore excluded from the study. Among the valid replicates, one replicate for T790M failed to detect the mutation but all other replicates generated correct results. Based on the valid results using pooled clinical samples, the estimated LoDs from the contrived samples were confirmed. The variability observed in this study is summarized in the table below.

Table 18. Overall Mean and Standard Deviations for CtR from Valid Results of Mutant Panel Members

| Panel Member | N | Mean CtR | Standard Deviation (SD) | | |
| --- | --- | --- | --- | --- | --- |
| | | | Site/ Instrument | Within-Run | Total |
| Ex 19del 1 - 1x LoD | 24 | 8.40 | 0.00 | 0.37 | 0.37 |
| Ex 19del 1 - 2x LoD | 24 | 6.89 | 0.14 | 0.20 | 0.24 |
| Ex 19del 2 - 1x LoD | 23 | 7.65 | 0.12 | 0.34 | 0.36 |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{23}

| Panel Member | N | Mean CtR | Standard Deviation (SD) | | |
| --- | --- | --- | --- | --- | --- |
| | | | Site/ Instrument | Within-Run | Total |
| Ex 19del 2 - 2x LoD | 24 | 7.00 | 0.10 | 0.46 | 0.47 |
| Ex 19del 3 - 1x LoD | 24 | 8.82 | 0.16 | 0.60 | 0.62 |
| Ex 19del 3 - 2x LoD | 24 | 7.63 | 0.10 | 0.26 | 0.28 |
| T790M - 1x LoD | 23 | 8.76 | 0.18 | 0.58 | 0.60 |
| T790M - 2x LoD | 24 | 7.61 | 0.36 | 0.34 | 0.50 |
| L858R - 1x LoD | 24 | 12.48 | 0.00 | 0.84 | 0.84 |
| L858R - 2x LoD | 24 | 11.49 | 0.00 | 0.52 | 0.52 |

## d. Analytical Sensitivity – LoD Verification of Rare Mutations

A study was performed to verify that the cobas® EGFR Plasma Test v2 can detect the non-predominant mutations that are part of the assay's design when present in EDTA plasma and to confirm that each of these mutations had a hit rate that is within the lower bound of the 95% CI of the LoD for the respective predominant mutation. If the non-predominant mutation had a hit rate ≥ 85% at the tested level, corresponding to the predominant mutation, then it was considered equivalent, because the upper bound of the 95% CI would overlap with the 95% hit rate. If the level was not equivalent, then concentrations up to 100 cp/mL were tested.

Plasmids containing the non-predominant mutations were diluted in HD EDTA plasma and tested with 21 replicates per level using one lot of cobas® EGFR Plasma Test v2 reagents. Plasmid DNA for each of the 35 non-predominant mutation targets was diluted in unique units of EDTA HD plasma to the LoD level of the predominant mutation for the respective mutation, or up to 100 cp/mL.

Fifty-four (54) runs out of the 60 total runs were valid. Four of the invalid runs were due to known operator errors and two were due to run controls: one out of range negative control, and one out of range mutation control. There were 11 invalid replicates out of 956 replicates tested, which were due to out of range IC Ct values. All invalids were valid upon repeat testing. A summary of the final LoDs determined for the non-predominant mutations is shown in the table below.

Table 19. Summary of the LoD Hit Rate for EGFR Plasmids in HD Plasma

| EGFR Mutation | EGFR Non-predominant Mutation | LoD for Non-Predominant Mutation [cp/mL] | # Hits/21 Valid Replicates | Hit Rate [%] |
| --- | --- | --- | --- | --- |
| G719X | G719S | 100 | 21 | 100 |
| | G719C | 100 | 21 | 100 |
| Ex 20ins | 2319_2320insCAC | 80 | 19 | 90.5 |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{24}

| EGFR Mutation | EGFR Non-predominant Mutation | LoD for Non-Predominant Mutation [cp/mL] | # Hits/21 Valid Replicates | Hit Rate [%] |
| --- | --- | --- | --- | --- |
| | 2310_2311insGGT | 80 | 20 | 95.2 |
| | 2311_2312ins9 (gcgtggaca) | 30 | 18 | 85.7 |
| | 2309_2310 complex (ac>ccagcgtggat) | 30 | 19 | 90.5 |
| Ex 19del | 2233_2247del15 | 25 | 21 | 100 |
| | 2236_2253 del 18 | 25 | 21 | 100 |
| | 2236-2250 del15 | 25 | 21 | 100 |
| | 2237-2251 del 15 | 25 | 21 | 100 |
| | 2237_2254 del 18 | 25 | 21 | 100 |
| | 2238_2255 del 18 | 25 | 21 | 100 |
| | 2239-2247 del 9 | 25 | 21 | 100 |
| | 2239-2253 del 15 | 25 | 21 | 100 |
| | 2239-2256 del 18 | 25 | 21 | 100 |
| | 2240_2251 del 12 | 25 | 20 | 95.2 |
| | 2240-2257 del 18 | 25 | 21 | 100 |
| | 2235_2252>AAT | 25 | 20 | 95.2 |
| | 2237-2255 >T | 25 | 20 | 95.2 |
| | 2238_2248>GC | 25 | 18 | 85.7 |
| | 2238_2252>GCA | 50 | 21 | 100 |
| | 2239-2248 TTAAGAGAAG>C | 25 | 20 | 95.2 |
| | 2239_2258>CA | 25 | 21 | 100 |
| | 2239-2251 >C | 25 | 21 | 100 |
| | 2253_2276del24 | 25 | 21 | 100 |
| | 2235_2248>AATTC | 50 | 21 | 100 |
| | 2237_2252>T | 25 | 20 | 95.2 |
| | 2238_2252del15 | 25 | 21 | 100 |
| | 2235_2251>AATTC | 50 | 21 | 100 |
| | 2235_2255>AAT | 25 | 18 | 85.7 |
| | 2236_2248>AGAC | 25 | 20 | 95.2 |
| | 2237_2253>TTGCT | 25 | 19 | 90.5 |
| | 2237_2257>TCT | 25 | 21 | 100 |
| | 2239_2256>CAA | 100 | 20 | 95.2 |
| L858R | 2573_2574TG>GT | 20 | 20 | 95.2 |

# 4. Analytical Specificity

Studies to assess both Inclusivity/Cross-Reactivity and Exclusivity were not performed for this PMA as there has been no change to the assay's primers, probes, or targeted mutations. (See Summaries of Safety and Effectiveness Data for P120019/S007.)

To identify potential interference of the cobas® EGFR Plasma Test v2, one lot of the cobas® cfDNA Sample Preparation Kit reagents and one lot of the cobas® EGFR Plasma Test v2 kit were used in testing of potentially interfering endogenous and exogenous substances, and microorganism testing.

PMA P150047: FDA Summary of Safety and Effectiveness Data

{25}

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 26

# a. Interference – Endogenous Interferents

To evaluate the potential interference of triglycerides, hemoglobin, albumin, and bilirubin (conjugated bilirubin and unconjugated bilirubin) on the performance of the cobas® EGFR Plasma Test v2, three HD plasma units (four for albumin) - each spiked with sheared cell line DNA (~220bp) - were used to create a five-member panel: four contrived mutant samples at approximately 3x LoD for each mutation and one EGFR WT sample. The potential endogenous interferents were spiked into the contrived samples prior to processing with the cobas® DNA Sample Preparation Kit, and the results after amplification and detection were compared to a neat (unspiked) sample. The levels of interferents tested were equal to the levels recommended to be tested by the CLSI EP7-A2 and are summarized in the table below.

Table 20. Table Potential Interferent Concentrations

| Potential Interferent | Concentration |
| --- | --- |
| Hemoglobin* | 2.0 g/L |
| | 1.5 g/L |
| Albumin | 60 g/L |
| Unconjugated Bilirubin | 0.2 g/L |
| Conjugated Bilirubin | 0.2 g/L |
| Triglycerides | 37 mM (33 g/L) |

The effect of excess hemoglobin was initially tested by spiking an aliquot with approximately 2 g/L. All results matched the expected results for each sample except for one dropout for Ex. 20ins. This replicate was repeated and was positive for Ex. 20ins. The concentration of hemoglobin was decreased to approximately 1.5 g/L of hemoglobin. All results matched the expected results for each sample after the addition of 1.5 g/L of hemoglobin. Excess albumin at a final concentration of approximately 60 g/L was evaluated and all results matched the expected results for each sample except for one G719X mutation sample which failed to identify the mutation.

No interference with the performance of the cobas® EGFR Plasma Test v2 was observed from either triglycerides, bilirubin (conjugated and unconjugated). Albumin at a concentration of ≥60 g/L may interfere with the cobas® EGFR Test.

# b. Interference – Exogenous Interferents

A study was conducted to evaluate the potential interference of therapeutic drugs (TARCEVA® and Neupogen®) and excess EDTA (to represent a short drawn specimen). As above, the potential interferents were spiked into aliquots from a sample panel prior to processing with the cobas® DNA Sample Preparation Kit. The levels of interferents tested were based on levels

{26}

recommended to be tested by the CLSI EP7-A2 guideline and are summarized in the table below.

Table 21. Potential Interferent Concentrations
| Potential Interferent | Concentration |
| --- | --- |
| EDTA | 9.0 mg/mL |
| Neupogen® | 147 ng/mL |
| TARCEVA® | 90 μg/mL |

All runs and sample replicates generated the expected results and demonstrated that EDTA, Neupogen®, and TARCEVA® do not interfere with the performance of the cobas® EGFR Plasma Test v2 for use with plasma samples.

c. Interference – Microorganisms

To evaluate the potential interference from contamination by Staphylococcus epidermidis, the same sample panel from the endogenous interfering substances study was used to assess 1×10⁶ colony forming units (CFU) of S. epidermidis. There were no invalid runs or samples, and the results demonstrated that 1×10⁶ CFU/mL of S. epidermidis did not interfere with the cobas® EGFR Plasma Test v2.

5. Reproducibility and Precision

The reproducibility of the cobas® EGFR Plasma Test v2 for the detection of mutations in exons 18, 19, 20, and 21 of the EGFR gene was evaluated in two studies. The first study used a nine member sample panel consisting of contrived sheared cell line DNA samples spiked into pooled WT NSCLC plasma. The mutant sheared cell line DNAs were combined into 4 double mutant positive combinations (see table below) at concentrations equivalent to approximately 100 copies/mL and 300 copies/mL and a single WT pooled NSCLC plasma sample.

Table 22. Reproducibility Panel
| Panel Member | Description of Panel member | Target Conc. (copies/mL) |
| --- | --- | --- |
| 1 | WT | 0 |
| 2 | Ex 19del (2235_2249del15) and T790M | 100/100 |
| 3 | Ex 19del (2235_2249del15) and T790M | 300/300 |
| 4 | S768I and G719X | 100/100 |
| 5 | S768I and G719X | 300/300 |
| 6 | L858R and T790M | 100/100 |
| 7 | L858R and T790M | 300/300 |
| 8 | L861Q and Ex 20ins(2307_2308ins9) | 100/100 |
| 9 | L861Q and Ex 20ins (2307_2308ins9) | 300/300 |

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 27

{27}

Each of two operators at each site tested the sample panel in duplicate once a day for three non-consecutive days using each of two lots of reagents for a total of 648 tests per panel member. Three unique lots of reagents were included in the study where each lot was tested at two sites.

One run was invalid due to an operator error and required repeating, resulting in 36/37 valid runs. The invalid run was not included in the analysis. Of 648 tests performed in total from all panel members, 646 generated valid results. There were a total of 1220 valid test results. The final results per mutation member are summarized in the tables below.

Table 23. Overall Estimates of Agreement by Mutation Member

| Mutation Member | copies/mL | Number of Valid Tests | Agreement N | Agreement % (95% CI)a |
| --- | --- | --- | --- | --- |
| WT | N/A | 72 | 72 | 100 (95.0, 100.0) |
| G719X | 100 | 72 | 65 | 90.3 (81.0, 96.0) |
| Ex 19del (2235_2249del15) | 100 | 72 | 72 | 100 (95.0, 100.0) |
| Ex 20ins (2307_2308ins9) | 100 | 72 | 72 | 100 (95.0, 100.0) |
| S768I | 100 | 72 | 72 | 100 (95.0, 100.0) |
| T790M | 100 | 143 | 139 | 97.2 (93.0, 99.2) |
| L858R | 100 | 71 | 70 | 98.6 (92.4, 100.0) |
| L861Q | 100 | 72 | 72 | 100 (95.0, 100.0) |
| G719X | 300 | 71 | 70 | 98.6 (92.4, 100.0) |
| Ex 19del (2235_2249del15) | 300 | 72 | 72 | 100 (95.0, 100.0) |
| Ex 20ins (2307_2308ins9) | 300 | 72 | 72 | 100 (95.0, 100.0) |
| S768I | 300 | 71 | 71 | 100 (94.9, 100.0) |
| T790M | 300 | 144 | 142 | 98.6 (95.1, 99.8) |
| L858R | 300 | 72 | 71 | 98.6 (92.5, 100.0) |
| L861Q | 300 | 72 | 72 | 100 (95.0, 100.0) |

Note: Results are included as agreement when a valid test of Mutation member has a result of MD or when a valid test of WT panel member has a result of NMD.
a 95% CI = 95% exact binomial confidence interval.

All mutation members demonstrated high agreement (e.g., &gt;97%) except G719X at 100 copies/mL (1x LoD). The lower agreement for this sample was due primarily to multiple missed calls (n = 6/24 replicates combined) occurring at one site (site 3) for both operators with one lot of reagents. One other replicate failed at a second site (site 1) with that same lot; however, the overall data does not support the conclusion that there were any manufacturing problems with that lot. Four replicates failed for the T790M mutation at 100 copies/mL, with three of the four failures occurring at one site.

Table 24. Overall Mean and SD for CtR from Valid Results of Mutation Members

| | | Standard Deviation (SD) Observed | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Mutation Member | Copies /mL | N | Mean CtR | Lot | Site/ Instrument | Operator | Day | With-in Run | Total |
| | | | | SD | SD | SD | SD | SD | SD |
| G719X | 100 | 65 | 9.08 | 0.38 | 0.00 | 0.07 | 0.42 | 0.39 | 0.69 |

PMA P150047: FDA Summary of Safety and Effectiveness Data

{28}

| | | Standard Deviation (SD) Observed | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Mutation Member | Copies /mL | N | Mean CtR | Lot | Site/ Instrument | Operator | Day | With-in Run | Total |
| | | | | SD | SD | SD | SD | SD | SD |
| Ex 19del (2235_2249del15) | 100 | 72 | 7.34 | 0.00 | 0.09 | 0.00 | 0.00 | 0.33 | 0.35 |
| Ex 20ins (2307_2308ins9) | 100 | 72 | 4.77 | 0.12 | 0.20 | 0.00 | 0.46 | 0.33 | 0.61 |
| S768I | 100 | 72 | 4.33 | 0.09 | 0.00 | 0.13 | 0.11 | 0.32 | 0.37 |
| T790M | 100 | 139 | 8.55 | 0.50 | 0.11 | 0.00 | 0.00 | 0.67 | 0.84 |
| L858R | 100 | 70 | 11.93 | 0.28 | 0.14 | 0.10 | 0.48 | 0.63 | 0.86 |
| L861Q | 100 | 72 | 5.75 | 0.29 | 0.23 | 0.00 | 0.42 | 0.22 | 0.60 |
| G719X | 300 | 70 | 7.44 | 0.10 | 0.00 | 0.00 | 0.34 | 0.43 | 0.56 |
| Ex 19del (2235_2249del15) | 300 | 72 | 5.69 | 0.05 | 0.11 | 0.10 | 0.00 | 0.24 | 0.29 |
| Ex 20ins (2307_2308ins9) | 300 | 72 | 4.07 | 0.00 | 0.12 | 0.00 | 0.15 | 0.43 | 0.47 |
| S768I | 300 | 71 | 3.24 | 0.00 | 0.00 | 0.00 | 0.11 | 0.23 | 0.26 |
| T790M | 300 | 142 | 6.73 | 0.38 | 0.13 | 0.00 | 0.15 | 0.25 | 0.49 |
| L858R | 300 | 71 | 10.37 | 0.00 | 0.00 | 0.05 | 0.00 | 0.56 | 0.57 |
| L861Q | 300 | 72 | 4.98 | 0.22 | 0.14 | 0.00 | 0.25 | 0.15 | 0.39 |

The second reproducibility study consists of the LoD confirmation study which is described in Section IX.A.3.c., above.

# 6. Lot-to-Lot Reproducibility

To evaluate lot interchangeability of the cobas® EGFR Plasma Test v2 and the cobas® cfDNA Sample Preparation Kit, a study was performed using 12 contrived EGFR mutant samples, covering each of the mutations detected by the assay, consisting of sheared cell line DNA diluted in HD plasma and one EGFR WT sample. One replicate of each sample at each of three concentrations was tested with each of the lots of the cobas® EGFR Plasma Test v2 kit tested in combination with three lots of the cobas® cfDNA Sample Preparation Kit. A $100\%$ correct call rate for each specimen with each combination was observed, supporting that each of the two kits can be used in different lot combinations.

# 7. Specimen Handling

Summary specimen handling information was provided based on the original clinical study protocols. After collection of the EDTA blood samples, plasma was separated by centrifugation within four (4) hours of collection and stored at $\leq -70^{\circ}\mathrm{C}$ until cfDNA isolation.

An additional study is planned based on a submitted protocol to support sample handling (i.e., initial sample processing and storage) and specimen stability studies. See Section XIV, below.

# 8. Guard banding

The objective of the guard banding studies was to establish the robustness of the PCR conditions for the cobas® EGFR Plasma Test v2. Guard banding studies

PMA P150047: FDA Summary of Safety and Effectiveness Data

{29}

were performed on the cobas® EGFR Plasma Test v2 Thermal Cycling Profile and Proteinase K concentration and incubation time (for DNA isolation procedure). Each guardband study used one lot of the cobas® cfDNA Sample Preparation Kit and one lot of the cobas® EGFR Plasma Test v2 kit. Three replicates of each test sample were evaluated by three operators and three cobas® z 480 analyzer.

Two contrived mutant samples, consisting of sheared cell line DNA spiked into pooled HD plasma at approximately 5x LoD for each mutation, and one pooled EGFR wild-type HD plasma specimen were tested. The contrived mutant specimens were pooled to include three and four different mutations each (C1: Ex. 19del, L858R, and T790M mutations and C2: S768I, L861Q, G719X, and Ex. 20ins mutations).

a. Thermal Cycling Profile

Part 1 tested the thermal cycling temperature with ±1°C to the control thermal cycling conditions outlined in the instructions for use. Three replicates of the samples described above were tested in three runs on one instrument. For each condition, all runs and results were valid and the results from each of the 27 replicates were correct.

b. Proteinase K (pK) Volumes

Part 2 tested the volume of the Proteinase K (pK) reagent used in the incubation step of DNA isolation. Three replicates from each mutant and WT sample were extracted according to the instructions for use with 250 μL ±20% of the normal pK reagent volume. The results from each extraction were compared to the third control replicate. For each condition, all runs and replicates were valid. All replicates gave the expected result except for one missed G719X result from a replicate which had 20% additional pK reagent, resulting in a total of 26/27 correct calls (96.3%).

c. pK Incubation Times

Part 3 tested the time of sample incubation with pK reagent. Three replicates from each mutant and WT sample were extracted according to the instructions for use but included incubating the replicates for ±5 minutes. For each condition, all runs and results were valid, and the results from each of the 27 replicates were correct.

Results from the guard-banding study showed that the cobas® EGFR Plasma Test v2 is robust in regard to incubation time of ±5 min, pK volume of ±20% and thermal cycling temperatures of ±1°C.

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 30

{30}

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 31

# 9. Stability Studies

## a. cobas® EGFR Mutation Test v2 (Shelf-Life)

A study was performed to confirm the stability of the existing cobas® EGFR Plasma Test v2 kit amplification and detection reagents. The cobas® EGFR Plasma Test v2 kit recommended storage temperature is 2-8°C. The kit contains six components:

- EGFR MMx1
- EGFR MMx2
- EGFR MMx3 v2
- Magnesium Acetate (MGAC)
- EGFR Mutant Control (EGFR MC)
- DNA Specimen Diluent (DNA SD)

In addition to the EGFR MC and DNA SD, this study included samples composed of eluates from six EDTA plasma samples generated using the cobas® cfDNA Sample Preparation Kit. All eluates were frozen until testing at each stability time point. The six plasma samples consisted of the following: four contrived samples created by spiking combinations of non-sheared cell line DNA for the seven predominant EGFR mutations detected by the assay into HD pooled plasma; one NSCLC EDTA plasma sample positive for the EGFR T790M mutation diluted into HD WT pooled plasma; and one HD EDTA plasma pool (EGFR WT).

Stability was evaluated by performing functional testing at day 0, 4 weeks, 8 weeks, 3 months, 6 months, 9 months, 12 months, 13 months, 18 months, and 19 months. The eluates for the plasma samples were tested at day 0, 6 months, 13 months, and 19 months. For a given storage condition to pass, the EGFR MC and DNA SD results must be within the pre-specified Ct ranges (which are identical to the Ct values for assessing validity of a run). At time points where the plasma samples were tested, the plasma samples had to yield the expected mutation result. To date, real-time stability testing for the cobas® EGFR Plasma Test v2 kit has been completed through 19 months (relative to Day 0 of the stability study).

## b. cfDNA Sample Preparation Kit

A study was conducted to confirm the stability of the cobas® cfDNA Sample Preparation Kit reagents. The cobas® cfDNA Sample Preparation Kit recommended storage temperature is 15-30°C. The real-time temperature used for the stability study was 32°C. In addition to a sample containing EGFR WT cell line DNA, a panel consisting of 3 clinical samples including two EGFR WT HD plasma specimens and one NSCLC WT plasma pool were used.

{31}

All kit components are identical to that approved under P120019/S007, except for the High Pure Extension Assembly Unit, of the cobas® cfDNA Sample Preparation kit. The stability of the cobas® cfDNA Sample Preparation Kit was assessed at various time points after storage at 32°C in upright orientation using three lots of the kit reagents. Stability was evaluated by performing functional testing at day 0, 4 weeks, 8 weeks, 3 months, 6 months, 9 months, 12 months, 13 months, 18 months, and 19 months.. To date, real-time stability testing for the cobas® EGFR Plasma Test v2 kit has been completed through 19 months (relative to Day 0 of the stability study).

## c. Prepared Specimen Stability

A study was performed to assess the stability of prepared specimens. In this study, 4 contrived EGFR mutant samples, consisting of sheared cell line DNA at three different concentrations diluted in pooled plasma, and one WT HD plasma sample were tested. On Day 0, the samples were processed and the eluates were pooled to create 5 pooled samples (4 mutant and one WT) which were then made into aliquots with sufficient volume for one time point. The pooled eluates were then stored at 32°C, 2-8°C, and -20°C (including freeze-thaw cycles) and tested in duplicate at pre-specified time points for each temperature storage condition as listed in the table below.

Table 25. Time Points and Conditions for the Prepared Specimen Stability Study

| Time Point | Temperature |
| --- | --- |
| Day 0 | N/A |
| Day 2 | 32°C |
| Day 8 | 2-8°C |
| | 32°C |
| Day 22 | 2-8°C |
| Day 31 | -20°C (Freeze/thaw cycle 1 and 2) |
| Day 61* | -20°C (Freeze/thaw cycle 1 and 2) |

* One invalid run was successfully repeated (corresponding to the sample panel tested on day 63).

At the pre-specified time points for each temperature storage condition, eluates were tested. One lot each of the cobas® cfDNA Sample Preparation Kit and cobas® EGFR Mutation Test v2 kit were used in the study. One out of 26 runs was invalid due to an out of range negative control. All replicates for each of the valid runs yielded correct calls.

The cfDNA extracted from specimens was stable when stored for up to 21 days at 2-8°C, up to 60 days and two freeze-thaw cycles at -20°C, and up to 7 days at 32°C.

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 32

{32}

d. Open Container Stability

Two stability studies were performed to demonstrate the stability of opened reagents for the cobas® cfDNA Sample Preparation Kit and the cobas® EGFR Mutation Test v2 Kit.

## cobas® cfDNA Sample Preparation Kit Open Container Stability

The first study evaluated the cobas® cfDNA Sample Preparation Kit reagents in two parts. Part A tested four uses (i.e., container opening, removal of reagents, container closure, and storage) of the cobas® cfDNA Sample Preparation Kit reagents over a period of 33 days (with testing at Days 0, 15, and 33, and volume removed at Day 21 [no testing]) and Part B tested four uses of the cobas® cfDNA Sample Preparation Kit reagents over a period of 91 days (with testing on Days 47, 61, and 91, and volume removed on Day 0 [used Day 0 results from Part A were used as Day 0 for Part B]).

The samples used in this study were 4 contrived EGFR mutant samples, covering each of the mutations detected by the assay, consisting of sheared cell line DNA at three different concentrations diluted in HD plasma and one EGFR WT sample. The samples were stored ≤-70 °C until the start of the study. At each designated time point, the samples were thawed, processed in duplicate, and tested. The cobas® cfDNA Sample Preparation Kit was stored at 32°C.

In Part A, all replicates produced the expected results except that one mutant at Day 15 and one at Day 33. One replicate for G719X at 3x LoD, which reported the expected mutation result at Day 15, but had a high Ct value when compared to all other replicates, tested to date. The sample was retested in duplicate and the resulting Ct values were within the range of the other replicates. On Day 33, one L858R replicate at 3x LoD reported the expected mutation result, but also had a Ct value that was considered high compared to all other replicates tested to date. The sample was re-tested in duplicate and both replicates gave correct calls and had Ct values in the expected range. In Part B, all runs and results were valid and correct.

Based on the study results, open reagent vials of the cobas® cfDNA Sample Preparation Kit are stable and are usable up to four times in a period of 90 days when stored at 15-30°C.

## cobas® EGFR Mutation Test v2 Kit Open Container Stability

The second study evaluated the cobas® EGFR Mutation Test v2 kit reagents in two parts. Part A tested four uses (i.e., container opening, removal of reagents, container closure, and storage) of the cobas® EGFR Mutation Test v2 kit reagents over a period of 31 days (with testing at Days 0, 15, 21 and

PMA P150047: FDA Summary of Safety and Effectiveness Data
Page 33

{33}

31). Part B tested four uses of the cobas® EGFR Mutation Test v2 kit reagents over a period of 91 days (with testing on Days 47, 61, and 91).

In this study, 4 contrived EGFR mutant samples, consisting of sheared cell line DNA at three different concentrations diluted in pooled plasma, and one WT HD plasma sample were tested. Prior to Day 0, the samples were processed and the eluates were pooled to create 5 pooled samples (4 mutant and one WT) which were then made into aliquots with sufficient volume for one time point. The pooled eluates were then stored at -20°C. At each designated time point, the eluates were thawed, and tested in duplicate. The cobas® EGFR Mutation Test v2 reagents were stored at 2-8°C.

In Parts A and B, all runs and results were valid and as expected except one replicate for Exon 19 deletion/Exon 20 Insertion at 3xLo…