therascreen PIK3CA RGQ PCR Kit

P190001 · QIAGEN GmbH · OWD · May 24, 2019 · Medical Genetics

Device Facts

• Record ID: P190001
• Device Name: therascreen PIK3CA RGQ PCR Kit
• Applicant: QIAGEN GmbH
• Product Code: OWD · Medical Genetics
• Decision Date: May 24, 2019
• Decision: APPR
• Device Class: Class 3

Intended Use

The therascreen PIK3CA RGQ PCR Kit is a real-time qualitative PCR test for the detection of 11 mutations in the phosphatidylinositol 3-kinase catalytic subunit alpha (PIK3CA) gene (Exon 7: C420R; Exon 9: E542K, E545A, E545D [1635G>T only], E545G, E545K, Q546E, Q546R; and Exon 20: H1047L, H1047R, H1047Y) using genomic DNA (gDNA) extracted from formalin-fixed, paraffin-embedded (FFPE) breast tumor tissue or circulating tumor DNA (ctDNA) from plasma derived from K2EDTA anticoagulated peripheral whole blood taken from patients with breast cancer. The test is intended to aid clinicians in identifying breast cancer patients who may be eligible for treatment with PIQRAY® (alpelisib) based on a PIK3CA Mutation Detected result. Patients whose FFPE tissue or plasma specimen produces a positive therascreen PIK3CA RGQ PCR Kit test result for the presence of one or more PIK3CA mutations are eligible for treatment with PIQRAY (alpelisib). Patients whose plasma specimen produces a negative result using this test should be reflexed to testing with FFPE tumor tissue for the presence of PIK3CA mutations. FFPE tumor specimens are processed using the QIAamp DSP DNA FFPE Tissue Kit for manual sample preparation. K2EDTA anticoagulated whole peripheral venous blood plasma specimens are processed using the QIAamp DSP Circulating Nucleic Acid Kit for manual sample preparation. For both specimen types, the Rotor-Gene Q (RGQ) MDx (US) instrument is used for automated amplification and detection. The Kit is to be used by trained personnel in a professional laboratory environment.

Device Story

Device detects 11 PIK3CA mutations in FFPE breast tumor tissue or plasma ctDNA. Input: gDNA from FFPE or ctDNA from K2EDTA plasma. Process: Manual extraction via QIAamp kits; automated amplification/detection on Rotor-Gene Q (RGQ) MDx instrument using real-time PCR. Technology: ARMS primers, LNA/TaqMan/MGB probes, PCR clamps. Output: Qualitative mutation status (Detected/Not Detected) via RGAM software. Used in professional laboratories by trained personnel. Output informs clinician eligibility for PIQRAY (alpelisib) therapy. Benefits: Enables targeted therapy selection for patients with PIK3CA-mutated breast cancer.

Clinical Evidence

Clinical evidence from SOLAR-1 Phase III trial (N=572). Retrospective retesting of 395 CTA-enrolled patients showed 97.0% PPA and 95.0% NPA between PIK3CA Kit and CTA. Clinical efficacy in PIK3CA Kit-positive population (HR=0.64, 95% CI: 0.48, 0.85) consistent with SOLAR-1 mutant cohort (HR=0.65).

Technological Characteristics

Real-time PCR assay using ARMS primers, LNA/TaqMan/MGB probes, and PCR clamps. Instrument: Rotor-Gene Q (RGQ) MDx. Connectivity: Standalone. Software: RGAM v2.1, Gamma MDx plug-in v1.0.0, Assay Profile v1.0.0. Sterilization: N/A (reagents).

Indications for Use

Indicated for detection of 11 PIK3CA mutations in FFPE breast tumor tissue or plasma-derived ctDNA in breast cancer patients to identify eligibility for PIQRAY (alpelisib) treatment. Negative plasma results require reflex testing with FFPE tissue.

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.

Reference Devices

• therascreen KRAS RGQ PCR Kit (P110027)

Submission Summary (Full Text)

{0}

PMA P190001: FDA Summary of Safety and Effectiveness Data

# SUMMARY OF SAFETY AND EFFECTIVENESS DATA (SSED)

## I. GENERAL INFORMATION

Device Generic Name: Real-time PCR Test

Device Trade Name: therascreen PIK3CA RGQ PCR Kit

Device Procode: OWD

Applicant's Name and Address: QIAGEN GmbH
QIAGEN Strasse 1
Hilden, 40724
Germany

Date(s) of Panel Recommendation: None

Premarket Approval Application (PMA) Number: P190001

Date of FDA Notice of Approval: May 24, 2019

The current PMA was submitted for an indication for the therascreen PIK3CA RGQ PCR Kit to identify 11 mutations in the phosphatidylinositol 3-kinase catalytic subunit alpha (PIK3CA) gene (Exon 7: C420R; Exon 9: E542K, E545A, E545D [1635G>T only], E545G, E545K, Q546E, Q546R; and Exon 20: H1047L, H1047R, H1047Y) using genomic DNA (gDNA) extracted from formalin-fixed, paraffin-embedded (FFPE) breast tumor tissue from breast cancer patients who may be eligible for treatment for treatment with PIQRAY® (alpelisib).

Another PMA (P190004) for the same device using plasma specimens was also submitted for the qualitative detection of same 11 mutations in PIK3CA gene (Exon 7: C420R; Exon 9: E542K, E545A, E545D [1635G>T only], E545G, E545K, Q546E, Q546R; and Exon 20: H1047L, H1047R, H1047Y) from circulating tumor DNA (ctDNA) isolated from K₂EDTA plasma from breast cancer patients for the treatment with PIQRAY (alpelisib). P190004 was also approved on May 24, 2019 in conjunction with P190001 approval.

The summary of safety and effectiveness data (SSED) to support the indication is available on the Center for Devices and Radiological Health (CDRH) website.

## II. INDICATIONS FOR USE

The therascreen PIK3CA RGQ PCR Kit is a real-time qualitative PCR test for the detection of 11 mutations in the phosphatidylinositol 3-kinase catalytic subunit alpha (PIK3CA) gene (Exon 7: C420R; Exon 9: E542K, E545A, E545D [1635G>T only], E545G, E545K, Q546E, Q546R; and Exon 20: H1047L, H1047R, H1047Y) using

Page 1

{1}

genomic DNA (gDNA) extracted from formalin-fixed, paraffin-embedded (FFPE) breast tumor tissue or circulating tumor DNA (ctDNA) from plasma derived from K₂EDTA anticoagulated peripheral whole blood taken from patients with breast cancer.

The test is intended to aid clinicians in identifying breast cancer patients who may be eligible for treatment with PIQRAY® (alpelisib) based on a PIK3CA Mutation Detected result. Patients whose FFPE tissue or plasma specimen produces a positive therascreen PIK3CA RGQ PCR Kit test result for the presence of one or more PIK3CA mutations are eligible for treatment with PIQRAY (alpelisib). Patients whose plasma specimen produces a negative result using this test should be reflexed to testing with FFPE tumor tissue for the presence of PIK3CA mutations.

FFPE tumor specimens are processed using the QIAamp DSP DNA FFPE Tissue Kit for manual sample preparation. K₂EDTA anticoagulated whole peripheral venous blood plasma specimens are processed using the QIAamp DSP Circulating Nucleic Acid Kit for manual sample preparation. For both specimen types, the Rotor-Gene Q (RGQ) MDx (US) instrument is used for automated amplification and detection. The Kit is to be used by trained personnel in a professional laboratory environment.

## III. CONTRAINDICATIONS

There are no known contraindications.

## IV. WARNINGS AND PRECAUTIONS

The warnings and precautions can be found in the therascreen® PIK3CA RGQ PCR Kit labeling.

## V. DEVICE DESCRIPTION

The following components comprise the overall device for testing FFPE tumor specimens.

- QIAGEN QIAamp DSP DNA FFPE Tissue Kit
- QIAGEN therascreen PIK3CA RGQ PCR Kit
- QIAGEN Rotor-Gene Q (RGQ) MDx Instrument with Automated data analysis and results interpretation using Rotor-Gene AssayManager® (RGAM) software version 2.1, Gamma MDx plug-in version 1.0.0, and therascreen PIK3CA MDx Assay Profile version 1.0.0 (therascreen_PIK3CA_FFPE_MDx)

## Specimen Preparation

FFPE blocks from breast cancer patients are cut into 4-5μm sections and mounted onto glass slides. A hematoxylin and eosin (H&E) stained slide is used to confirm that there is tumor present. For extraction from resected tissue samples (RES), two non-stained tissue sections are scraped from the slide for Deoxyribonucleic Acid (DNA) extraction. For core needle biopsy (CNB) samples, between one and six sections are required for

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 2

{2}

extraction to obtain the total effective tumor area of at least $20\mathrm{mm}^2$ . If, however, $20\mathrm{mm}^2$ is not achievable with 6 CNB sections, then extraction is performed with a maximum of 6 CNB sections. If a resected sample has a tumor content $<10\%$ in the region of interest (ROI), then the sample must be macrodissected.

DNA is manually extracted and purified using the QIAGEN QIAamp DSP DNA FFPE Tissue Kit. The FFPE sample is deparaffinized with xylene, the xylene supernatant is then removed, and any residual xylene is extracted with ethanol. The sample is lysed under denaturing conditions with proteinase K for one hour at $56^{\circ}\mathrm{C}$ . The sample is heated for one hour at $90^{\circ}\mathrm{C}$ to reverse formalin cross-linking of genomic DNA (gDNA). The sample is passed through a silica-based membrane so that the gDNA binds to the membrane and any contaminants are removed. The membrane is washed multiple times with buffers (using a centrifuge to increase flow rates through the membrane). Purified gDNA is eluted from the membrane using elution buffer (ATE) from the QIAamp DSP DNA FFPE Tissue Kit. Elution volumes vary by sample type (RES or CNB) to maintain the desired or expected DNA concentration. For RES, the gDNA samples must be eluted with $120\mu \mathrm{L}$ of ATE, while for CNB, the gDNA samples must be eluted with $70\mu \mathrm{L}$ of ATE.

# PCR Amplification and Detection

The QIAGEN therascreen PIK3CA RGQ PCR Kit contains reagents for the detection of 11 individual mutations in exons 7, 9, and 20 of the phosphoinositide-3-kinase, catalytic, alpha (PIK3CA) gene, and a region in exon 15 of the PIK3CA gene used as a Control Reaction. The detection of the 11 individual mutations and region in exon 15 of the PIK3CA gene is achieved using six reaction mixes listed below:

Table 1: PIK3CA Mutations Detected by the therascreen PIK3CA RGQ PCR Kit

| Reaction Mix | PIK3CA Mutation | Base change | Exon | COSMIC ID1 |
| --- | --- | --- | --- | --- |
| 1 | Control Reaction | N/A | 15 | N/A |
| 2 | E542K | 1624 G>A | 9 | 760 |
| 3 | E545K | 1633 G>A | | 763 |
| | E545D | 1635 G>T | 9 | 765 |
| | E545G | 1634 A>G | | 764 |
| 4 | E545A | 1634 A>C | 9 | 12458 |
| | H1047Y | 3139 C>T | 20 | 774 |
| | Q546R | 1637 A>G | 9 | 12459 |
| 5 | C420R | 1258 T>C | 9 | 757 |
| | Q546E | 1636 C>G | 7 | 6147 |
| 6 | H1047L | 3140 A>T | 20 | 776 |
| | H1047R | 3140 A>G | | 775 |

$^{1}$ COSMIC IDs taken from the Catalogue of Somatic Mutations in Cancer: https://cancer.sanger.ac.uk/cosmic

PMA P190001: FDA Summary of Safety and Effectiveness Data

{3}

The therascreen PIK3CA RGQ PCR Kit uses real-time PCR with the following technologies for PIK3CA sequence related amplification and detection: ARMS® Primers, LNA® Probes, TaqMan® Probes, Minor Groove Binding (MGB™) Probes, and PCR Clamps for Wild Type (WT) and pseudogene sequences.

PCR uses forward and reverse primers to hybridize to a specific DNA sequence to amplify it. The ARMS technique is based on the use of mutation sequence-specific PCR primers that allow amplification of test DNA only when the target allele is contained within the sample.

In addition to the primers, dye-linked oligonucleotides (i.e., probes; e.g., LNA, TaqMan and MGB) are contained in the reaction mixes. The probes, which are labeled with a 5' reporter dye (carboxyfluorescein [FAM™]) and a downstream, 3' dye-free quencher (i.e., BHQ1) which quenches the fluorescence of the reporter dye, also hybridize to the target sequence between the primers. When a probe is intact, the proximity of the reporter dye to the quencher results in suppression of the reporter fluorescence primarily by Förster-type energy transfer.

PCR clamp technology allows selective amplification of the mutant sequence. PCR clamps matched to WT or pseudogene sequence bind to the template and prevent amplification by interference with primer elongation. There are two types of PCR clamps used within the therascreen PIK3CA RGQ PCR Kit; peptide nucleic acids (PNAs) and 3' phosphate groups.

During PCR, forward and reverse primers and a probe bind to the target sequence. DNA polymerase extends the primers and the 5' to 3' exonuclease activity of the enzyme cleaves the probe between the reporter and the quencher leading to an increase in detectable reporter fluorescence. This process occurs in every PCR cycle. The increase in fluorescence is directly proportional to the target amplification during PCR.

The probes used in the mutation specific reaction mixes are labeled with carboxyfluorescein (FAM), hexachloro-fluorescein (HEX) and Cyanine (CY5.5) fluorescent reporter dyes, each with a distinct absorption and emission profile. The probe used in the Internal Control Reaction is labeled with Rhodamine (ROX).

FAM, HEX, ROX and CY5.5 absorb and fluoresce at different wavelengths:

- FAM: 6-carboxyfluorescein: A fluorophore that excites at a wavelength of 495nm and emits at a wavelength of 520nm. This fluoresces in the green RGQ channel.
- HEX: Hexachloro-fluorescein: A fluorophore that excites at a wavelength of 535nm and emits at a wavelength of 556nm. This fluoresces in the yellow RGQ channel.
- ROX: A fluorophore of the Rhodamine family that excites at 578nm and emits at 604nm. This fluoresces in the orange RGQ channel.
- CY5.5: A fluorophore of the Cyanine family that excites at 675nm and emits at 694nm. This fluoresces in the crimson RGQ channel.

The Control Reaction Mix contains a forward and reverse primer and labeled probe (detected in the Green Channel) to amplify a short sequence of exon 15 of the PIK3CA

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 4

{4}

gene. The Control Reaction is used to determine if an appropriate level of amplifiable DNA is present in the sample and is a factor in the analytical calculations that determine mutation status. All samples must be tested with the Control Reaction Mix (Tube 1) to ensure that they give Ct values within a specified range to ensure that there is enough amplifiable DNA to proceed with analysis, but not so much as to overload the assay. The Control Reaction determines whether the quality and quantity of DNA is sufficient for the working range of the assay. The interpretation of the results obtained from the Control Reaction Ct is presented below in Table 2. Any samples that do not give Ct values within this range are invalidated by the RGAM software.

Table 2: Control Reaction Working Range
| Control Ct value | Interpretation | Action |
| --- | --- | --- |
| > 33.38 | Quantity of amplifiable DNA is not sufficient for mutation analysis | Additional samples should be extracted and tested |
| < 23.23 | Quantity of amplifiable DNA is too high for mutation analysis | Dilute with the sample diluent water supplied in the kit |
| Within 23.23 - 33.38 | Quantity of amplifiable DNA is suitable for mutation analysis | No action required, sample is suitable |

The PCR cycling parameters used for assessing the DNA sample with the control reaction mix are the same run parameters for mutation analysis using the mutation assays. The PCR cycling parameters used for assessing the DNA sample are:

- Hold at 95°C for 15 minutes to activate the Taq polymerase;
- PCR for 45 cycles of 95°C for 30 seconds, to denature, and 60°C for 1 minute, to anneal/extend.

If the control assay Ct falls within range, then the sample is analyzed for the presence of the mutation by analyzing the values obtained in the mutation channels and completing the ΔCt calculation. If the control assay Ct is not within range the sample is considered invalid and any results obtained may not be used to make a mutation status evaluation. This assessment is performed automatically by the RGAM software and associated plug-in and assay profile.

## Test Controls

The therascreen PIK3CA RGQ PCR Kit contains three controls: An Internal Control (IC), a Positive Control (PC) and a No Template Control (NTC), which have been designed to detect fault conditions.

Internal Control (IC): Each PIK3CA reaction mix contains reagents (unlabeled primers, probe and oligonucleotide template) for an IC reaction designed to detect failure of the reaction, e.g. due to set up error, and confirms successful PCR reactions in every tube.

PMA P190001: FDA Summary of Safety and Effectiveness Data

{5}

No Template Control (NTC): An NTC test contains nuclease-free water and is required in each RGQ run. The NTC serves as a control to assess potential contamination during assay set up.

Positive Control (PC): A PC test is required in each RGQ run. The PC Tube comprises a mixture of plasmids representing one mutation for each of the mutation assays and the Control Assay. Detection of the targets within acceptable ranges confirms the proper functioning of each of the reaction mixes in the kit.

## Instrument and Software

The therascreen PIK3CA RGQ PCR Kit is designed to be used with the Rotor-Gene Q MDx (hereafter referred to as the RGQ instrument), which is a real-time PCR analyzer designed for rapid thermal cycling and real-time detection of PCR assays. The RGQ incorporates a centrifugal rotary design for thermal cycling where a rotor, containing each tube, spins in a chamber of moving air, keeping all samples at a uniform temperature. Samples are heated and cooled in a low-mass-air oven according to a software-determined cycle that initiates the different phases of the PCR cycle. In the RGQ, fluorophores are excited from the bottom of the sample chamber by a light-emitting diode. Energy is transmitted through the thin wall at the bottom of each PCR tube. Emitted fluorescence passes through the emission filters on the side of the chamber and is detected by a photomultiplier tube. Detection is performed as each tube aligns with the detection optics; tubes spin pass the excitation / emission optics every 150 milliseconds. The fluorescence signals indicate the progress of the PCR reactions. The RGQ instrument has six channels (six excitation sources and six detection filters). Four of these channels; green, yellow, crimson and orange, are used with the therascreen PIK3CA RGQ PCR Kit.

Cycling parameters, data analysis and results interpretation for the therascreen PIK3CA RGQ PCR Kit are performed by the RGAM version 2.1 Software, Gamma MDx plug-in version 1.0.0 and therascreen_PIK3CA_FFPE Assay Profile v1.0.0. Therefore, no manual analysis is required.

The RGAM Software is a core software which provides general functionality including: PCR run set up, cycler control and management of experiment data, results, assay profiles and system configuration.

The Gamma MDx Plug-in extends the functionality of RGAM by providing cycle threshold (Ct) value calculation, data analysis and normalization features.

Assay specific functionality, for example cycling conditions, thresholds and analysis cutoffs, and control ranges, is implemented by the therascreen_PIK3CA_FFPE Assay Profile.

The RGAM software, plug-in and associated assay profile ensure that a user interface with restricted user options is displayed to the user and contains all the information

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 6

{6}

required for automatic real-time PCR analysis including time and temperature profiles, data quality controls, and data analysis algorithms. The software suite also allows printing of test reports and creates result files in the software's file system.

In addition, the RGAM software, plug-in and associated assay profile perform a quality check using Automatic Data Scan (AUDAS) that focuses on parameters of the respective fluorescence curves from which Ct values will be determined.

The AUDAS check is mainly intended to identify problems that occur during the real-time PCR amplification that potentially generate non-typical curve shapes due to saturation, noise, spikes, baseline dips, sloping curves related to the real-time PCR instrument parameters or due to a problem linked to the assay itself. The curves in such situations are automatically invalidated to avoid generating misleading results.

## Interpretation of Results

The first cycle at which the instrument can distinguish the amplification-generated fluorescence as being above the background signal is called the Ct. The RGAM software interpolates fluorescence signals between any two recorded values. Ct values can therefore be any number (not limited to integers) within the range of 0 to 45.

Ct values generated by the Control and Mutation Reactions indicate the quantity of assay specific input DNA. Low Ct values indicate higher input DNA levels and high Ct values indicate lower input DNA levels.

Validity of controls and samples are determined based on the Ct values generated during a run.

## Run Validity Criteria

For therascreen PIK3CA RGQ PCR Kit test runs to be accepted as valid, the RGAM software requires run data for PC and NTC to meet criteria specified in the PIK3CA MDx Assay Profile v1.0.0 in accordance with the analysis rules of the Gamma MDx Plug-in v1.0.0. The PC and NTC validity criteria are shown below. Each test run performed with the therascreen PIK3CA RGQ PCR Kit must meet all the validity criteria listed below.

Table 3: Run, Sample Validity and Call Criteria

| Sample | Reaction Mix | Target | RGQ Channel | Ct Range |
| --- | --- | --- | --- | --- |
| Positive Control | 1 and 5 | Control | FAM | 23.39 – 32.39 |
| | 2 | E542K | FAM | 22.42 – 31.42 |
| | 3 | E545D | HEX | 23.78 – 32.78 |
| | 3 | E545G | CY5.5 | 22.61 – 31.61 |
| | 3 | E545K | FAM | 24.41 – 33.41 |
| | 4 | E545A | FAM | 22.5 – 31.5 |
| | 4 | H1047Y | HEX | 26.57 – 35.57 |
| | 4 and 5 | Q546R | CY5.5 | 24.04 – 33.04 |

PMA P190001: FDA Summary of Safety and Effectiveness Data

{7}

| Sample | Reaction Mix | Target | RGQ Channel | Ct Range |
| --- | --- | --- | --- | --- |
| | 5 | C420R | HEX | 23.31 – 34.31 |
| | 1 and 5 | Q546E | FAM | 24.72 – 35.72 |
| | 6 | H1047L | CY5.5 | 24.02 – 33.02 |
| | 6 | H1047R | HEX | 23.33 – 32.33 |
| NTC | All | All 6 reaction mixes | FAM | Has no value |
| IC | NTC | IC | ROX | 25.52 – 36.51 |
| | Test Sample | IC | ROX | 25.51 – 36.51 |
| | Positive Control | IC | ROX | 25.51 – 36.51 |
| Test Sample | 1 | Control | FAM | 23.39 – 32.39 |

If a test run fails any of the validity criteria, the RGAM software displays the corresponding validity rule related to the failed control but does not provide test results for samples on the RGAM report. If all run validity criteria are met, the RGAM generates a report that confirms the respective controls validity and then displays the sample results.

The individual sample results in each test run are accepted as valid, if the RGAM software obtains Ct values for the PIK3CA assay. If a sample fails to generate a Ct value for any PIK3CA mutant channel, then the RGAM software checks the Ct value obtained in the IC channel to ensure the qPCR reaction validity. If the RGAM fails to detect a signal within the validity criteria range in the IC, the sample is reported as invalid and no PIK3CA mutation status results for that sample are reported. For a therascreen PIK3CA RGQ PCR Kit run to be accepted as valid, the RGAM software, plug-in and associated assay profile require run data for the PC and NTC, to meet specified criteria.

# Sample Validity Criteria and Control Assay Working Range

All samples must be tested with the Control Reaction mix to ensure that they give a Ct value within a specified range. This range is set to ensure that there is sufficient amplifiable DNA to proceed with analysis, but not so much as to overload the assay. The therascreen PIK3CA RGQ PCR Kit has been verified to work within a specific working range (i.e. upper and lower Control Reaction Ct values) and any samples that do not give Ct values within this range are invalidated by the RGAM software, plug-in and associated assay profile.

# Determination of Sample Status

If the Control Assay Ct falls within range, then the sample is analyzed for the presence of the mutation.

The difference in Ct values $(\Delta \mathrm{Ct})$ between the Control Reaction and the mutation-specific reaction (mutation assay) is a qualitative measure of PIK3CA mutation status and is calculated as:

$\Delta \mathrm{Ct} = [\mathrm{Mutation~Reaction~Ct~value}] - [\mathrm{Control~Reaction~Ct~value}]$

PMA P190001: FDA Summary of Safety and Effectiveness Data

{8}

Samples are classed as mutation positive if they give a $\Delta \mathrm{Ct}$ less than or equal to the cut-off $\Delta \mathrm{Ct}$ value for that assay. Above this value, the sample may either contain less than the percentage of mutation that can be detected by the therascreen PIK3CA RGQ PCR Kit (beyond the limit of detection of the assays), or the sample is mutation negative, both of which would be reported as "No Mutation Detected".

The Mutation Assay $\Delta \mathrm{Ct}$ Cut-offs are shown below in Table 4.

Table 4: Cut-off Values for Each Mutation in Each Assay
| Reaction Mix | Mutation name | Ct Cut-off values |
| --- | --- | --- |
| Tube 2 | E542K | 4.8 |
| Tube 3 | E545K | 6.5 |
| | E545D | 7.5 |
| | E545G | 9.5 |
| Tube 4 | E545A | 10.0 |
| | H1047Y | 6.2 |
| | Q546R | 7.0 |
| Tube 5 | Q546E | 10.0 |
| | C420R | 6.0 |
| Tube 6 | H1047R | 7.0 |
| | H1047L | 10.0 |

On the RGAM report each sample is assigned with a status as follows:

Invalid:
- If one of the AUDAS checks failed
- or if one of the run control criteria failed
- or if the sample IC failed
- or if the Control Assay Ct was outside of the acceptance range

PIK3CA Mutation Detected:
- If all AUDAS checks passed
- and if all run control criteria were met
- and if the sample IC was within the defined acceptance range
- and if the Control Assay Ct was within the acceptable range
- and if any PIK3CA mutant signals were equal to or below the predefined $\Delta \mathrm{Ct}$ cut-off

No Mutation Detected:
- If all AUDAS checks passed
- and if all run control criteria were met
- and if the sample IC was within the defined acceptance range
- and if the Control Assay Ct was within the acceptable range
- and if all mutant signals were above the predefined $\Delta \mathrm{Ct}$ cut-off

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 9

{9}

VI. ALTERNATIVE PRACTICES AND PROCEDURES

There are no other FDA cleared or approved alternatives for the testing of FFPE breast tumor tissue for PIK3CA mutation status in the selection of patients who are eligible for treatment with alpelisib.

VII. MARKETING HISTORY

The therascreen PIK3CA RGQ PCR Kit has not been marketed in the United States or any foreign country.

VIII. POTENTIAL ADVERSE EFFECTS OF THE DEVICE ON HEALTH

Failure of the device to perform appropriately, or failure to correctly interpret test results may lead to incorrect PIK3CA mutation results, which could impact patient treatment decisions. A false positive test result may lead to inappropriate treatment and adverse effects associated with treatment with a targeted PIK3CA inhibitor rather than standard of care treatments. A false negative test result may prevent a patient from receiving alpelisib with potential to benefit from a targeted therapy.

For the specific adverse events that occurred in the clinical study, please see Section X below.

IX. SUMMARY OF NONCLINICAL STUDIES

A. Laboratory Studies

The specific performance characteristics of the therascreen PIK3CA RGQ PCR Kit (henceforth referred to as PIK3CA Kit) were determined by studies using formalin-fixed, paraffin-embedded (FFPE) tissue specimens obtained from patients with advanced breast cancer, FFPE human cell lines, and gDNA pre-extracted from human cell lines. FFPE cell lines were sectioned and processed similar to FFPE patient specimens. Cell line DNA was extracted and tested in accordance with the instructions for surgical resections.

Mutation status of specimens was confirmed by Digital Droplet PCR (ddPCR) and/or bi-directional Sanger sequencing. The similarity between FFPE cell lines and FFPE clinical specimens was assessed for high prevalence mutations (E542K, H1047R, and E545K) by comparing the differences in mean $\Delta C_t$ between the two sample types and by comparing the positivity rates and probit models for the two sample types. The results of the evaluation demonstrated that evaluation of FFPE cell lines does not lead to overestimation of assay performance.

1. Correlation with Reference Method

The accuracy of the PIK3CA Kit was demonstrated relative to a validated Next Generation Sequencing (NGS) assay using FFPE clinical specimens from the breast

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 10

{10}

cancer patients randomized in the SOLAR-1 trial from for which there was sufficient quantity and quality of specimen available for testing with the NGS comparator assay. 453 of 572 randomized patient samples had available FFPE slides for testing. Of these 453 clinical specimens, 385 met the NGS comparator sample requirements for tissue volume and tumor content, and 379 yielded a valid result for NGS. The baseline clinical and demographic characteristics of the patients whose specimens were available for this retrospective testing were in general comparable to those of otherwise eligible patients whose specimens were not available for retesting. The overall results are shown in Table 5.

Table 5: PIK3CA Kit Compared to NGS - SOLAR-1 Clinical Trial Specimens

| PIK3CA Kit Call | NGS Mutation Status | | | | | | | | | | | | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | C420R | E542K | E545A | E545D | E545G | E545K | H1047L | H1047R | Q546E | Q546R | C420R & H1047L | C420R & H1047R | E542K & E545K | E542K & H1047R | E545D & H1047R | E545G & H1047R | E545K & H1047R | NMD | Total |
| C420R | 3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 3 |
| E542K | 0 | 26 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 6 |
| E545A | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| E545D | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 2 |
| E545G | 0 | 0 | 0 | 0 | 3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 4 |
| E545K | 0 | 0 | 0 | 0 | 0 | 46 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 48 |
| H1047L | 0 | 0 | 0 | 0 | 0 | 0 | 11 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 13 |
| H1047R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 95 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | 98 |
| Q546E | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| Q546R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 2 |
| C420R & H1047L | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| C420R & H1047R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 1 |
| E542K & E545K | 0 | 0 | 0 | 0 | 0 | 2 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 |
| E542K & H1047R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| E545A & H1047R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 2 |
| E545D & H1047R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| E545G & H1047R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| E545K & H1047R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 1 |
| NMD | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 162 | 164 |
| Total | 3 | 26 | 1 | 0 | 3 | 48 | 11 | 100 | 1 | 2 | 1 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 180 |

PMA P190001: FDA Summary of Safety and Effectiveness Data

{11}

Samples with both NGS and PIK3CA Kit valid results were analyzed to assess overall percent agreement (OPA), positive percent agreement (PPA), and negative percent agreement (NPA) based on the agreements between the two methods for mutation status. The percentages, together with the corresponding two-sided exact 95% CI are summarized below.

The results demonstrate point estimates of PPA, NPA and OPA of 99.0%, 90.0% and 94.7% respectively (Table 6).

Table 6: PIK3CA Kit vs. NGS Concordance in the SOLAR-1 Trial Specimens
| Measure of Agreement | Percent Agreement (N) | 95% CI* |
| --- | --- | --- |
| PPA | 99.0% (197/199) | (96.4%, 99.9%) |
| NPA | 90.0% (162/180) | (84.7%, 94.0%) |
| OPA | 94.7% (359/379) | (92.0%, 96.7%) |

* The 95%CI calculated using the Clopper-Pearson Exact method

For the 20 overall mutation status discordant results, 2 samples with PIK3CA Kit negative results had NGS positive results, while 18 samples with PIK3CA Kit positive results gave NGS negative results. Of the 2 samples with PIK3CA Kit negative results that had NGS positive results, both were detected by the NGS at mutant allele frequencies levels below the PIK3CA Kit limit of detection. Of the 18 samples, determined positive by the PIK3CA Kit and negative by NGS, eleven were low positive (within 1 ΔCt of the cut-off using the PIK3CA Kit and therefore low positive samples). One case was detected as H1047L by the PIK3CA Kit but detected as H1047I by the NGS assay. The underlying cause for the 6 remaining discordant results was not identified.

Additional agreement analyses were also conducted for each specimen type (RES vs. CNB) and for each of the mutation assays in the PIK3CA Kit except for H1047Y as this mutation was not represented in the study. For CNB samples the point estimates of OPA, PPA and NPA were 98.72% (77/78), 97.73% (43/44) and 100% (34/34), respectively. For RES the point estimates of OPA, PPA and NPA were 93.69% (282/301), 99.35% (154/155) and 87.67% (128/146), respectively.

## 2. Analytical Sensitivity

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

The LoB of the PIK3CA Kit was established by testing DNA extracted from 56 individual clinical wild type (WT) FFPE samples (30 WT RES samples and 26 WT CNB samples) with two replicates per sample for each of three PIK3CA Kit lots (generating 336 data points total). The LoB was defined as the highest measurement result that corresponds to the upper 95th percentile in the WT samples. The lowest value of the three LoB estimates (one from each therascreen PIK3CA Kit lot), across both RES and CNB sample types was determined to be the LoB value.

The LoB values for each of the mutation assays (in terms of ΔCt) detected by the PIK3CA Kit were verified to be above the ΔCt cut-off values determined

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 12

{12}

for each of the assays and are summarized below in Table 7 along with the false positive call rates obtained.

Table 7: LoB Values and False Positive Call Rate for all 11 Mutation Assays

| Mutation | LoB (ΔCt) | False Positive Call Rate (Percent) |
| --- | --- | --- |
| E542K | 5.09 | 1.88 |
| E545K | 6.74 | 1.57 |
| E545D | 9.19 | 0.31 |
| E545G | 13.03 | 0.00 |
| E545A | 13.03 | 0.00 |
| H1047Y | 7.61 | 0.63 |
| Q546R | 8.72 | 0.00 |
| Q546E | 13.03 | 0.00 |
| C420R | 7.57 | 0.94 |
| H1047R | 9.80 | 1.25 |
| H1047L | 12.63 | 0.94 |

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

The PIK3CA Kit does not use a specific concentration of DNA as determined by spectrophotometry. DNA input is based on the Control Reaction Ct result which is used to indicate that there is sufficient amplifiable DNA present in the sample. The stated DNA input for the assay is defined by the Control Ct prespecified range of 23.23 to 33.38, as shown in Table 2 above.

For the PIK3CA Kit LoD is defined as the minimum percentage of mutant DNA in a background of WT DNA that can be detected with a $95\%$ probability as determined by a probit analysis. The LoDs for the 11 mutation assays (C420R, E542K, E545A, E545D E545G, E545K, H1047L, H1047R, H1047Y, Q546E and Q546R) in the PIK3CA Kit were assessed using a selection of breast cancer FFPE clinical specimens or FFPE cell lines for the low prevalence mutations. The percent mutant DNA or mutant allele frequency (MAF) of each individual sample was determined previously by using Sanger sequencing for clinical samples and from information available in Cancer Cell line Encyclopedia (CCLE) for the cell lines. Five or six point MAF dilution series were made by serially diluting mutant samples in a FFPE clinical WT background. For each PIK3CA mutation, the percentage of correct calls was assessed across dilution levels using three different PIK3CA Kit lots with 24 replicates tested per kit lot per level. The LoD was determined at low DNA input (Control Ct of $\sim 30.00$ ). The hit rates for each mutant concentration, per kit lot, were obtained and a probit model was fitted to the data. A plot was produced of hit rate against log2 mutation level and the LoD for each mutation was determined as the concentration which gave a $95\%$ estimated probability of a positive call. The final LoD value for each

PMA P190001: FDA Summary of Safety and Effectiveness Data

{13}

mutation was determined as the highest value (in terms of MAF%) across all kit lots.

Table 8: Determined LoD for the 11 Mutations Detected by the PIK3CA Kit

| Mutation Name | Mutation Prevalence | LoD (MAF%) |
| --- | --- | --- |
| C420R | <1% | 2.41* |
| E542K | 11% | 5.47† |
| E545A | 1.50% | 3.54* |
| E545D | <1% | 2.69† |
| E545G | <1% | 4.98† |
| E545K | 20% | 4.13† |
| H1047L | 5% | 2.56† |
| H1047R | 55% | 3.13† |
| H1047Y | <1% | 14.04* |
| Q546E | <1% | 4.50* |
| Q546R | <1% | 6.08† |

* LoD values were established using FFPE cell lines samples
† LoD values were established using breast cancer FFPE clinical specimens

c. Analytical Sensitivity – Control Ct Range, RFI Validation and ΔCt Cut-offs

i. Control Ct Ranges

The objective of this study was to set an appropriate Control Reaction Ct working range for use in assessing DNA sample validity. The Control Reaction Ct working range defines the amount of total amplifiable DNA in a sample and was determined using a total of 20 WT FFPE clinical samples generating 107 data points. The WT status of samples was previously assessed using an independent reference method. To determine the Control Ct working range, only the data gained from the Control Ct produced by the WT samples was used. Different confidence levels were calculated to allow the best range where 95% of the population would fit. Parametric and non-parametric tolerance intervals were calculated for Control Ct values. The Control working range was set based on the non-parametric tolerance intervals since the data was not normally distributed.

The final Control Reaction Ct working range was set at a Ct value of 23.23 to 33.38. This interval provides limits for which 95% of the population falls with a given confidence level of 97.3%. The final Control Reaction Ct working range allows mutation analysis on small amounts of input DNA such as that from CNBs.

ii. Relative Fluorescence Increase (RFI) Validation

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 14

{14}

During development data from the positive control (PC), gDNA from positive cell lines and non-template control (NTC) were used to calculate appropriate fluorescence threshold values for the Green, Yellow, and Crimson RGQ channels used in PIK3CA Kit. Assessment of signal to noise ratio (SNR) was conducted to ensure appropriate fluorescence thresholds were set for each RGQ MDx channel used in the therascreen PIK3CA RGQ PCR System. The fluorescence threshold selected was required to be suitable for all assays detected in any one channel. The fluorescence threshold for the Orange channel, which is used for the detection of the IC, could not be calculated following the same methodology described for the Green, Yellow and Crimson RGQ channels. Since the IC is present in all tubes and is designed to amplify consistently within defined specifications, it was not possible to compare this positive signal to a negative sample. However, since the IC contains a standardized single concentration of synthetic DNA, the fluorescence threshold selected was not required to factor in assay performance across a range of different DNA inputs and as such the difference between signal and noise is not critical for this assay. The Orange channel threshold was selected such that its position was at the start of the exponential phase of the amplification plot. This was determined to be 0.0600 and therefore selected as the threshold for the Orange channel, which is used for the IC. The fluorescence thresholds for the Green, Yellow and Crimson channel were determined and are given as 0.0510, 0.475 and 0.625 respectively.

iii. ΔCt Cut-offs

Assay ΔCt cut-offs were established during development using clinical, WT, and mutation-positive FFPE clinical specimens, FFPE cell line FFPE and cell line gDNA. The cut-offs were determined in terms of ΔCt values and were chosen with respect to the following parameters: false positive fraction, false negative fraction and assay sensitivity. In addition to statistical analysis of ΔCt values, design requirements for false positive rates considering false negative rates were used to define an acceptable target range for cut-off values.

A study was conducted to verify the ΔCt cut-offs for each mutation (refer to Section V for the assays cut-off values) established during development using WT FFPE clinical samples, mutation positive FFPE clinical samples and cell-line FFPE samples. The ΔCt cut-off values were verified by testing three low-input (Ct 30) / low-positive samples (low MAF%). The results of the study demonstrated that each individual mutation assay had a false positive rate below 3%, while the false positive percentage for the overall PIK3CA Kit was 7.21%.

3. Linearity – Effect of DNA input on ΔCt

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 15

{15}

To demonstrate that the performance of the PIK3CA Kit is consistent across the DNA input range a series of nine dilutions with varying DNA input levels with the upper and lower levels being outside of the Control Reaction Ct working range (23.23 - 33.38 Ct), were evaluated with mutation positive samples covering all the 11 variants detected by the PIK3CA Kit. Three different sample types were used in this study: clinical FFPE resection samples, cell line FFPE samples, and gDNA preextracted from cell lines. The mutant allele frequencies were held constant while DNA input was varied; the target Ct values for the lowest and highest concentrations were 23.00 Ct and 33.50 Ct, respectively.

The evaluation was performed using one PIK3CA Kit lot with three replicates tested per DNA level. The data was analyzed using regression analysis to determine the linear range. For the assay to be determined as linear across the DNA input range, there should be no change across the range in $\Delta \mathrm{Ct}$ , i.e., there is no statistically significant linear, quadratic or cubic effect. The E542K, E545D, E545G, E545A, H1047Y, Q546E, C420R and H1047R assays did not show a statistically significant $(p &gt; 0.05)$ first, second or third order terms for all models tested. These assays show no change in $\Delta \mathrm{Ct}$ across the tested DNA input range.

The E545K assay was statistically significant for the first order and second order polynomial in the linear $(p = 0.006)$ and quadratic $(p = 0.002)$ models, respectively. The Q546R assay was statistically significant for the first order and third order polynomial in the linear $(p &lt; 0.001)$ and cubic $(p = 0.001)$ models, respectively. The H1047L assay was statistically significant for the first order and third order polynomial in the linear $(p &lt; 0.001)$ and cubic $(p &lt; 0.001)$ models, respectively. The E545K, Q546R and H1047L assays are not linear for $\Delta Ct$ across the tested DNA input range. E545K showed statistically significant values for only low DNA input samples. Q546R and H1047L showed statistically significant values for both low and high DNA input samples. An investigation determined that the non-linear effects had no effect on the performance of the E545K and H1047L assays. However, an effect on the Q546R assay performance was determined; samples at LoD may be called false negative when the DNA input is high (approximately Control Ct 23), however, the probability of this occurring is extremely low, approximately $0.0052\%$ .

The ranges in which the assays are linear are shown below (Table 9).

Table 9: Linear range for the PIK3CA Kit

| Mutation | Linear Range based on observed Mean Ct |
| --- | --- |
| E542K | 24.42 to 33.77 |
| E545K | 24.08 to 31.02 |
| E545D | 23.02 to 34.99 |
| E545G | 22.83 to 35.71 |
| E545A | 23.18 to 34.31 |
| H1047Y | 23.24 to 34.64 |

PMA P190001: FDA Summary of Safety and Effectiveness Data

{16}

| Q546R | 24.28 to 32.69 |
| --- | --- |
| Q546E | 23.18 to 34.65 |
| C420R | 23.26 to 34.11 |
| H1047R | 23.76 to 33.49 |
| H1047L | 25.74 to 31.61 |

The data was also analyzed to assess the amplification efficiency for each of the assays. Amplification efficiencies ranged from 82.15 to 100.83% for all assays.

## 4. Analytical Specificity

### a. Cross Reactivity/Exclusivity

The PIK3CA Kit is comprised of 6 separate reaction mixes: one single Control Reaction that detects a region in exon 15 of the PIK3CA gene and eleven mutation assays that detect PIK3CA mutations (Table 1). There is no reaction that specifically measures the wild-type PIK3CA sequence at exons 7, 9 or 20. The PIK3CA Kit "No Mutation Detected" result is determined from the absence of any positive mutation results.

The objective of this study was to assess whether cross reactivity has been correctly accounted for in the setting of the analytical $\Delta C_t$ cut-off values. The cross reactivity of the six optimized reaction mixes was assessed by testing each mutation assay against the 11 mutants detected by the PIK3CA Kit. Samples were tested at low DNA input and low MAF% and high DNA input and high MAF%. Mutation positive clinical FFPE resection samples were used for the four most prevalent mutations (E542K, E545K, H1047L and H1047R), while cell line FFPE samples were used for the less prevalent mutations (C420R, E545A, E545D, E545G, H1047Y, Q546E and Q546R). For each sample, two replicates were tested for each of three PIK3CA Kit lots (generating 240 data points total).

Within this study, there was one instance of cross reactivity with E545D and H1047R, and one instance with C420R and H1047R. There were also four instances of mutant non-specific amplification between the high MAF sample E545A and H1047L. Overall 6/240 data points showed mutant non-specific amplification. The six data points showing mutant non-specific amplification were sporadic and inconsistent with other replicates from the same sample. These results were therefore not considered to be a result of cross reactivity.

PCR cross reactivity was observed between H1047L and H1047R. This cross reactivity is uni-directional i.e., if a double H1047R and H1047L sample is seen this will be reported as H1047R Mutation Detected. This rule is incorporated into the automated therascreen_PIK3CA_FFPE Assay Profile algorithm.

### b. Interference – Effect of Necrotic Tissue

PMA P190001: FDA Summary of Safety and Effectiveness Data

{17}

To evaluate the potential interference of necrotic tissue content in breast cancer FFPE specimens on the performance of the PIK3CA Kit, FFPE clinical specimens from SOLAR-1 with both PIK3CA Kit and NGS results were analyzed. A total of 180 specimens determined PIK3CA mutant negative by NGS and 199 specimens determined PIK3CA mutant positive by NGS were evaluated. The specimens included both CNB and RES specimens. Percent necrosis, as identified by a pathologist varied from 0 to 10% for mutant negative and 0 to 20% for mutant positive samples.

For both mutant positive and mutant negative FFPE specimens, all except 20 samples have PIK3CA Kit results that matched the expected NGS results. The 20 mismatched results were from 17 mutant negative and 2 mutant positive samples with less than 5% necrotic content; and 1 mutant negative sample with less than 10% necrotic content; thus, it is unlikely that necrosis was the reason for the discordant results. The results support the use of the PIK3CA Kit with breast cancer FFPE specimen with necrotic tissue content up to 20%.

c. Interference – Exogenous Substances and Hemoglobin

The effect of potential interfering substances introduced from the FFPE Extraction Kit (an exogenous substance) or from the sample itself (hemoglobin) on assay performance were measured by comparison of ΔCt between interferent spiked and control spiked extracts for each mutant and comparison of the correct calls for WT DNA samples. The exogenous substances tested were (1) paraffin wax, (2) ethanol, (3) xylene, (4) extraction buffers (AW1, ATL, AW2 and AL) and (5) proteinase K. Clinical FFPE resection samples were evaluated for mutations E542K, E545K, Q546R, C420R, H1047R and for WT, while FFPE cell line samples were used to evaluate the Q546R and C420R mutations.

Samples that were spiked with exogenous interferents were first normalized to Ct 30.00 and then diluted with WT (also normalized to Ct 30.00) to give the ΔCt expected at a MAF representing 3x LoD. Samples spiked with hemoglobin (endogenous interferent) during the extraction process were not normalized to Ct 30.00 or diluted to 3x LoD prior to mutation assessment but used immediately following extraction. This was to avoid removing any variability which may have been introduced by the interferent.

The study required the preparation of a test sample set and a blank sample set (ATE for exogenous substances and water for hemoglobin). The test sample set included all mutant and WT samples spiked with an interferent. The blank sample set included mutant and WT samples spiked with an appropriate control substance. Samples tested with hemoglobin were spiked during the extraction process to reflect what would be introduced via the FFPE sample. The test concentration of hemoglobin and the estimated tissue volume used in the extraction process were based on CLSI guidelines (CLSI

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 18

{18}

EP7-A2, Appendix D, 2005, Interference Testing in Clinical Chemistry; Approved Guideline). The recommended testing concentration of hemoglobin given in EP07-A, Appendix D, 2005 is 2mg/ml. Samples tested with potential exogenous interferents were spiked following normalization to Ct 30.00 and dilution to 3x LoD at a concentration representing the highest (worst-case) feasible level of the interfering substance carry over into a sample (10x concentration). In total, six replicates of each sample/interferent combination were tested with one PIK3CA Kit lot. All mutation calls in both mutant and WT samples were as expected. Where a significant difference was observed between the spiked and control samples, this was within acceptable intermediate precision of the assay and was, therefore within the inherent variability of the assay. In conclusion, none of the interferents tested had any impact on kit performance.

## 5. Repeatability and Reproducibility

The repeatability and reproducibility of the PIK3CA Kit was investigated by testing DNA extracted from breast cancer FFPE clinical specimens for mutations E542K, E545G, E545K, H1047L, H1047R and Q546R, and Cell Line FFPE samples for mutations C420R, E545A, E545D, H1047Y, Q546E and Q546R. A WT FFPE clinical specimen was also included in the study. To demonstrate repeatability, samples at two mutation levels (LoD and 3x LoD) were tested in duplicate with two (2) runs per day, by three (3) operators across 20 non-consecutive days resulting in 120 data points at one site (located in the United Kingdom) except for samples at LoD with E545A and Q546R PIK3CA mutations. Samples with E545A and Q546R mutations at LoD were evaluated for six (6) days at one site by three (3) operators, with two (2) runs and four (4) replicates for a total of 144 measurements to demonstrate repeatability. For reproducibility two (2) runs a day were performed per operator (3 operators per site) by two additional sites (both located in the USA) over 10 days to give additional 60 data points for each additional site except for samples with at LoD with E545A and Q546R PIK3CA mutations. Samples at LoD with E545A and Q546R PIK3CA mutations were evaluated for six (6) days for two (2) more sites, by three (3) operators, with two (2) runs and four (4) replicates for a total of 144 measurements per site, 432 in total across three (3) sites. At each site, samples were tested using two (2) PIK3CA Kit lots (3 lots across 3 sites). One to two lots of QIAamp® DSP DNA FFPE Tissue Kit were used to extract DNA from FFPE samples. Samples were prepared to have low DNA input levels where a control Ct value of approximately 30 was targeted.

Mutation positive samples were only run with the Control Reaction mix and the relevant reaction mix specific to the tube of the mutation of interest. WT samples were run with all reaction mixes per the PIK3CA Kit Instructions For Use (IFU).

A new repeatability and reproducibility study will be conducted by running samples with all the 11 PIK3CA mutations using 6 reaction tubes in accordance with the PIK3CA Kit IFU; refer to section XIII.

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 19

{19}

For each sample, the proportion of correct mutation calls along with the corresponding two-sided exact 95% confidence limit are reported in Table 10 below, for repeatability.

Table 10: Assay Repeatability – Proportion of Correct Calls for PIK3CA Mutation Tested

| Mutation | Template | Fractional Proportion of Valid Results | Percentage | Two-Sided Lower 95% Confidence Limit |
| --- | --- | --- | --- | --- |
| C420R | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 120 / 120 | 100.00% | 96.97% |
| E542K | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 119 / 119 | 100.00% | 96.95% |
| E545A | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD* | 144 / 144 | 100.00% | 97.47% |
| E545D | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 120 / 120 | 100.00% | 96.97% |
| E545G | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 120 / 120 | 100.00% | 96.97% |
| E545K | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 118 / 120 | 98.33% | 94.11% |
| H1047L | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 117 / 120 | 97.50% | 92.87% |
| H1047R | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 120 / 120 | 100.00% | 96.97% |
| H1047Y | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 117 / 120 | 97.50% | 92.87% |
| Q546E | 3x LoD | 120 / 120 | 100.00% | 96.97% |
| | LoD | 120 / 120 | 100.00% | 96.97% |
| Q546R | 3x LoD | 119 / 119 | 100.00% | 96.95% |
| | LoD* | 139 / 140 | 99.29% | 96.08% |
| WT | Ct 30 | 108 / 120 | 90.00% | 83.18% |

*Samples at LoD with E545A and Q546R PIK3CA mutations were evaluated for 6 days at one site by 3 operators, with 2 runs and 4 replicates for a total of 144 measurements.

For each sample, the proportion of correct mutation calls along with the corresponding two-sided exact 95% confidence limit are reported in the Table 11 below for reproducibility.

Table 11: Assay Reproducibility – Proportion of Correct Calls for PIK3CA Mutation Tested

| Mutation | Template | Fractional Proportion of Valid Results | Percentage | Two-Sided Lower 95% Confidence Limit |
| --- | --- | --- | --- | --- |
| C420R | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD | 240 / 240 | 100.00% | 98.47% |
| E542K | 3x LoD | 240 / 240 | 100.00% | 98.47% |

PMA P190001: FDA Summary of Safety and Effectiveness Data

{20}

| | LoD | 237 / 239 | 99.16% | 97.01% |
| --- | --- | --- | --- | --- |
| E545A | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD* | 431 / 432 | 99.77% | 98.72% |
| E545D | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD | 238 / 240 | 99.17% | 97.02% |
| E545G | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD | 240 / 240 | 100.00% | 98.47% |
| E545K | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD | 238 / 240 | 99.17% | 97.02% |
| H1047L | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD | 230 / 240 | 95.83% | 92.47% |
| H1047R | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD | 240 / 240 | 100.00% | 98.47% |
| H1047Y | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD | 234 / 240 | 97.50% | 94.64% |
| Q546E | 3x LoD | 240 / 240 | 100.00% | 98.47% |
| | LoD | 240 / 240 | 100.00% | 98.47% |
| Q546R | 3x LoD | 239 / 239 | 100.00% | 98.47% |
| | LoD* | 421 / 424 | 99.29% | 97.95% |
| WT | Ct 30 | 222 / 240 | 92.50% | 88.41% |

*Samples at LoD with E545A and Q546R PIK3CA mutations were evaluated for 6 days across 3 sites, by 3 operators, with 2 runs and 4 replicates for a total of 144 measurements per site, 432 total.

The WT sample evaluated in the repeatability and reproducibility study did not meet the acceptance criteria in relation to proportion of correct calls. An investigation indicated that the WT sample had an additional mutation with low allelic frequency as determined by an NGS-based method. A new repeatability and reproducibility study will be conducted postmarket with an appropriately qualified WT sample. Refer to section XIII.

A variance component analysis was used to estimate the standard deviation for between-run, between-day, between-lot, between-operator, between-instrument, and between-day variability for the repeatability study. These estimates were reported along with the number of observations for the mean of $\Delta \mathrm{Ct}$ , Control Ct, and Mutant Ct values. Results by variance components and total variance are presented below for $\Delta \mathrm{Ct}$ . The column "N" included the number of data points that generated a $\Delta \mathrm{Ct}$ value. Across all variance components, the total standard deviation (SD) was $\leq 1.14$ in all PIK3CA mutations tested. Across all mutant panel members, the SD was in general $\leq 0.52$ for between-operators, between lots, between days, and between runs.

Table 12: Repeatability - Overall Mean, Standard Deviation (SD) for $\Delta \mathrm{{Ct}}$ - Between Lot, Between-run, Between-Operator, Between-Day, Between-Instrument and Total Variance

PMA P190001: FDA Summary of Safety and Effectiveness Data

{21}

| Mutation | LoD Level Tested | N | Mean | Between Kit Lot (SD) | Between Run (SD) | Between Operator (SD) | Between Day (SD) | Between Instrument (SD) | Within Run (SD) | Total (SD) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| E542K | 3x LoD | 120 | 1.95 | 0.00 | 0.16 | 0.00 | 0.10 | 0.14 | 0.24 | 0.33 |
| | LoD | 119 | 3.59 | 0.00 | 0.21 | 0.00 | 0.09 | 0.11 | 0.35 | 0.43 |
| E545K | 3x LoD | 120 | 2.68 | 0.02 | 0.14 | 0.08 | 0.10 | 0.08 | 0.30 | 0.36 |
| | LoD | 120 | 4.41 | 0.09 | 0.22 | 0.24 | 0.00 | 0.16 | 0.51 | 0.60 |
| E545A | 3x LoD | 120 | 1.32 | 0.00 | 0.00 | 0.00 | 0.00 | 0.12 | 0.53 | 0.54 |
| | LoD* | 144 | 2.79 | 0.00 | 0.00 | 0.14 | 0.09 | 0.13 | 0.92 | 0.93 |
| Q546E | 3x LoD | 120 | 2.38 | 0.08 | 0.00 | 0.00 | 0.14 | 0.00 | 0.39 | 0.41 |
| | LoD | 120 | 4.71 | 0.00 | 0.00 | 0.07 | 0.00 | 0.21 | 0.70 | 0.72 |
| E545D | 3x LoD | 120 | 2.83 | 0.00 | 0.05 | 0.10 | 0.00 | 0.00 | 0.34 | 0.35 |
| | LoD | 120 | 4.45 | 0.02 | 0.00 | 0.05 | 0.15 | 0.07 | 0.58 | 0.60 |
| H1047Y | 3x LoD | 120 | 3.47 | 0.05 | 0.14 | 0.06 | 0.00 | 0.00 | 0.33 | 0.37 |
| | LoD | 120 | 4.94 | 0.11 | 0.20 | 0.00 | 0.00 | 0.17 | 0.51 | 0.58 |
| C420R | 3x LoD | 120 | 0.96 | 0.00 | 0.29 | 0.00 | 0.08 | 0.06 | 0.30 | 0.43 |
| | LoD | 120 | 2.57 | 0.00 | 0.11 | 0.22 | 0.14 | 0.15 | 0.61 | 0.67 |
| H1047R | 3x LoD | 120 | 2.34 | 0.00 | 0.21 | 0.07 | 0.03 | 0.19 | 0.28 | 0.39 |
| | LoD | 120 | 4.4 | 0.00 | 0.00 | 0.08 | 0.00 | 0.19 | 0.56 | 0.58 |
| E545G | 3x LoD | 120 | 1.92 | 0.08 | 0.15 | 0.00 | 0.00 | 0.00 | 0.26 | 0.31 |
| | LoD | 120 | 5.1 | 0.00 | 0.29 | 0.00 | 0.00 | 0.00 | 0.83 | 0.88 |
| Q546R | 3x LoD | 119 | 1.67 | 0.14 | 0.18 | 0.00 | 0.00 | 0.19 | 0.41 | 0.48 |
| | LoD* | 140 | 3.99 | 0.13 | 0.00 | 0.00 | 0.11 | 0.30 | 0.78 | 0.83 |
| H1047L | 3x LoD | 120 | 4.87 | 0.00 | 0.26 | 0.14 | 0.00 | 0.11 | 0.43 | 0.52 |
| | LoD | 119 | 6.96 | 0.12 | 0.00 | 0.00 | 0.52 | 0.13 | 1.04 | 1.14 |

*Samples at LoD with E545A and Q546R PIK3CA mutations were evaluated for 6 days at one site by 3 operators, with 2 runs and 4 replicates for a total of 144 measurements.

A variance component analysis was used to estimate the standard deviation for between-run, between-day, between-lot, between-operator, between-instrument, and between-day variability for the reproducibility study. These estimates were reported along with the number of observations for the mean of $\Delta \mathrm{Ct}$ , Control Ct, and Mutant Ct values. Results by variance components and total variance are presented below for $\Delta \mathrm{Ct}$ . The column "N" included the number of data points that generated a $\Delta \mathrm{Ct}$ value. Across all mutant panel members, the SD was in general $\leq 0.36$ for within site, between-lots, between instrument, between days, and between operators.

PMA P190001: FDA Summary of Safety and Effectiveness Data

{22}

Table 13: Reproducibility - Overall Mean, Standard Deviation (SD) for $\Delta \mathrm{{Ct}}$ -Between-run, Within - Site, Between Site, Between Lot, Between-Operator, Between-Day, Between-Instrument and Total Variance

| Mutation | LoD Level Tested | N | Mean | Between Site (SD) | Between Run, Within Site (SD) | Between Operator (SD) | Between Instrument (SD) | Between Day (SD) | Between Lot (SD) | Within Run (SD) | Total (SD) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| E542K | 3x LoD | 240 | 1.85 | 0.03 | 0.00 | 0.02 | 0.23 | 0.06 | 0.07 | 0.34 | 0.40 |
| E542K | LoD | 239 | 3.49 | 0.00 | 0.20 | 0.00 | 0.30 | 0.00 | 0.00 | 0.43 | 0.54 |
| E545K | 3x LoD | 240 | 2.73 | 0.00 | 0.14 | 0.05 | 0.08 | 0.09 | 0.03 | 0.29 | 0.34 |
| E545K | LoD | 240 | 4.43 | 0.00 | 0.23 | 0.13 | 0.18 | 0.00 | 0.05 | 0.47 | 0.56 |
| E545A | 3x LoD | 240 | 1.27 | 0.00 | 0.10 | 0.06 | 0.08 | 0.00 | 0.00 | 0.53 | 0.55 |
| E545A | LoD | 432 | 2.85 | 0.00 | 0.11 | 0.00 | 0.17 | 0.08 | 0.08 | 1.06 | 1.09 |
| Q546E | 3x LoD | 240 | 2.41 | 0.00 | 0.12 | 0.00 | 0.00 | 0.06 | 0.11 | 0.41 | 0.43 |
| Q546E | LoD | 240 | 4.78 | 0.00 | 0.00 | 0.04 | 0.15 | 0.10 | 0.00 | 0.74 | 0.77 |
| E545D | 3x LoD | 240 | 2.86 | 0.05 | 0.00 | 0.09 | 0.09 | 0.09 | 0.00 | 0.35 | 0.37 |
| E545D | LoD | 239 | 4.55 | 0.10 | 0.27 | 0.00 | 0.15 | 0.17 | 0.06 | 0.61 | 0.71 |
| H1047Y | 3x LoD | 240 | 3.45 | 0.00 | 0.14 | 0.00 | 0.09 | 0.00 | 0.02 | 0.39 | 0.42 |
| H1047Y | LoD | 240 | 4.93 | 0.00 | 0.12 | 0.00 | 0.18 | 0.00 | 0.05 | 0.54 | 0.58 |
| H1047R | 3x LoD | 240 | 2.38 | 0.00 | 0.14 | 0.04 | 0.11 | 0.04 | 0.00 | 0.32 | 0.37 |
| H1047R | LoD | 240 | 4.53 | 0.11 | 0.09 | 0.00 | 0.15 | 0.00 | 0.00 | 0.60 | 0.63 |
| C420R | 3x LoD | 240 | 1.01 | 0.00 | 0.23 | 0.00 | 0.11 | 0.00 | 0.16 | 0.32 | 0.42 |
| C420R | LoD | 240 | 2.71 | 0.07 | 0.22 | 0.07 | 0.17 | 0.00 | 0.12 | 0.56 | 0.64 |
| E545G | 3x LoD | 240 | 1.89 | 0.00 | 0.16 | 0.02 | 0.05 | 0.00 | 0.07 | 0.28 | 0.34 |
| E545G | LoD | 240 | 5.04 | 0.00 | 0.26 | 0.00 | 0.00 | 0.00 | 0.00 | 0.80 | 0.84 |
| Q546R | 3x LoD | 239 | 1.57 | 0.14 | 0.12 | 0.04 | 0.12 | 0.00 | 0.11 | 0.44 | 0.49 |
| Q546R | LoD | 423 | 3.97 | 0.00 | 0.17 | 0.00 | 0.34 | 0.09 | 0.08 | 0.77 | 0.86 |
| H1047L | 3x LoD | 240 | 5 | 0.22 | 0.36 | 0.17 | 0.07 | 0.02 | 0.09 | 0.46 | 0.63 |
| H1047L | LoD | 239 | 7.15 | 0.17 | 0.22 | 0.26 | 0.22 | 0.15 | 0.17 | 1.24 | 1.32 |

*Samples at LoD with E545A and Q546R PIK3CA mutations were evaluated for 6 days across 3 sites, by 3 operators, with 2 runs and 4 replicates for a total of 144 measurements per site, 432 total.

# 6. Lot-to-Lot Interchangeability

The objective of this study was to demonstrate lot-to-lot interchangeability and to demonstrate consistency of the mutation status across the QIAamp DSP DNA FFPE Tissue Kit and the therascreen PIK3CA RGQ PCR Kit. The study utilized three lots of the FFPE Extraction Kit and three lots of the therascreen PIK3CA RGQ PCR Kit

PMA P190001: FDA Summary of Safety and Effectiveness Data

{23}

to test mutation positive breast cancer FFPE clinical specimens and cell line FFPE samples representing the 11 mutations assays along with WT FFPE clinical samples. A total of 22 samples were used: 13 mutation positive clinical breast FFPE samples, three WT clinical breast FFPE specimens, and six FFPE cell line samples. Samples were used without any modification or dilution. Each sample was extracted with three lots of QIAamp FFPE Extraction Kits. Two replicate extractions were carried out per QIAamp FFPE Extraction Kit, to give 6 extractions per sample. All extracts were tested with three different lots of the therascreen PIK3CA RGQ PCR Kit for a total of 396 data points, including 342 mutation positive and 54 WT data points, reported across all 9 PIK3CA Kit and extraction kit combinations. Further a total of 44 data points, including 38 mutation positive and 6 WT data points, would be reported per kit lot and extraction lot combination. Mutant samples were tested with the Control assay and the mutation assay of interest. WT samples were tested with the full therascreen PIK3CA RGQ PCR Kit.

The percentage of correct overall mutation call and specific mutation calls were reported across all samples and kit lots, along with the corresponding two-sided exact 95% confidence limits. The percentage of correct overall mutation call was also reported for each PCR and extraction kit combination separately, along with the corresponding two-sided exact 95% confidence limits.

Of the nine kit combinations, eight were found to give mutation positive results (E542K and E545K) for WT sample 1. Based on these findings, investigations were performed to determine the sample status of WT Sample 1. It was found that WT Sample 1 had been misclassified as a WT sample. An assessment of the data determined that removal of this sample from the dataset did not affect the capability of the study to determine Lot Interchangeability.

In addition to the removal of the WT sample 1, C420R extract 2 was removed due to sample invalidity. The number of data points was therefore reduced. A total of 375 data points, including 339 mutation positive and 36 WT data points, were reported across all 9 PCR kits and extraction kit combinations. A total of 41 data points per PCR kit lot and extraction kit lot 1 combination was reported including 37 mutation positive samples and 4 WT data points. A total of 42 data points per PCR kit lot and both extraction kit lot 2 and extraction kit lot 3 combinations was reported including 38 mutation positive samples and 4 WT data points.

The purpose of the lot interchangeability study was to determine that extraction and PCR kits are interchangeable. With the removal of WT Sample 1, the data from 2 WT samples was used during the statistical analysis.

The proportion of correct overall mutation calls was 96.80% for each therascreen PIK3CA RGQ PCR Kit lot and Extraction Kit lot combination. For extraction kit lot 3 and all PCR kit combinations, the proportion of correct overall mutation call was 90.48% and for extraction kit lot 1 and 2 and all PCR kit combinations, the proportion of correct overall mutation calls was 100.00%.

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 24

{24}

In conclusion, this study showed that lot-to-lot interchangeability has no impact on assay performance. It also showed that combining different lots of the PIK3CA RGQ PCR Kit and FFPE Extraction Kit does not have an effect on the ability to determine a correct mutation call.

## 7. Specimen Handling – Reproducibility

The objective of this study was to assess sample handling variability, specifically within the DNA extraction as part of the PIK3CA Kit test system process at three different sites. Four mutation positive clinical FFPE breast cancer specimens and six mutation positive cell line FFPE samples representing all 11 mutation assays along with one WT clinical FFPE breast cancer specimen were used in this study. Twelve (12) sections of 11 samples (10 mutant and one WT) were distributed evenly across three different sites. Six extractions per sample (two sections per extraction per sample) were performed at each of the three different sites one located in the United Kingdom, and two in the USA. Testing of the extracts using the PIK3CA Kit was performed at the site in the United Kingdom. Extractions were performed at all three sites using one kit lot of QIAamp DSP DNA FFPE Tissue Kit. There were 48 FFPE sections required for each sample; these FFPE sections were randomized and split into 24 extract sets. These extract sets were then distributed evenly across the three testing sites, six extracts per study site. The remaining six extracts were retained as a contingency set.

Each mutation assessment run tested all six extractions of one sample from each site with the full PIK3CA Kit, that is, with testing with all 6 reaction mixes included in the kit. In total 198 replicates across three sites (that is, 6 replicates at each of the 3 sites for each of the 10 mutation positive samples and 1 WT sample) were planned for evaluation. Sixty-six (66) replicates within sites (that is, 6 replicates at each site for each of the 10 mutation positive samples and 1 WT sample) were planned for testing.

In the final dataset, one run (contained 6 data points) was removed as it was invalid, due to a PC out of specification. This run was not repeated as sufficient data points to conduct the primary analysis were collected from all other associated runs. Due to this, the study has 192 data points instead of 198. There is no impact on the study design by this reduction in sample size. The original power calculations are not impacted, and the power remains &gt;99% over all extractions. When comparing the results of the samples across all three sites, the percentage of correct mutation calls for mutation positive and WT samples was 100.00%. Across specific PIK3CA mutation calls, the proportion of correct calls was 97.92%.

## 8. Guardbanding

The objective of the guard banding studies was to establish the robustness of the PIK3CA Kit. The following studies were conducted to: (1) assess the tolerance of the PIK3CA Kit to temperature variations in the annealing step during PCR that could be introduced by the RGQ instrument, (2) determine the effect of varying

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 25

{25}

reagent volume on the mutation status of samples called by the PIK3CA Kit, (3) determine the effect of varying reagent mixing (4) determine the effect of varying the thaw time of the PIK3CA Kit reagents and samples and (5) determine the effect of varying set-up time on the mutation status called by the therascreen PIK3CA RGQ PCR Kit. For all guardband studies, a WT sample and a representative mutation in each of the reaction mixes within the kit was assessed at 3xLoD and low DNA input.

## a. Cycling Guardband

The study was designed to determine the tolerance of the PIK3CA Kit to temperature variations of the annealing step during PCR that could be introduced by the RGQ instrument.

The standard cycling conditions for the therascreen PIK3CA RGQ PCR Kit are denaturation at 95°C for 30 seconds and annealing at 60°C for 60 seconds. Annealing temperature was tested across a seven-point range 60°C ±2°C. The 59°C to 61°C temperature range was selected to represent temperatures within the RGQ dynamic temperature specification, two conditions, 58°C and 62°C were selected to represent temperatures outside the RGQ specification range. Mutant samples were tested with 4 replicates per run over three runs generating a total of 12 replicates per temperature condition. WT samples were tested with 2 replicates per run. A total of 10 replicates were generated per temperature condition.

When the PCR annealing temperature was varied by ±2°C, correct calls across all tested mutation assays, for each temperature condition, within the dynamic temperature range was 100.00%.

## b. Volumetric Guardbanding

The objective of this study was to determine the effect of varying reagent volume on the mutation status of samples called by the PIK3CA Kit. Six samples, a representation of each of the PIK3CA Kit tubes (five mutants and one WT) were tested.

The standard volumes as stated in the instructions for use are 19.83μL RM, 0.17μL Taq and 5.00μL of sample. The volumetric tolerance was tested by varying the volume of each individual component while keeping the volume of the other components constant. Each component volume was varied by ±6% as shown below (Table 14).

Table 14: Volumetric conditions to be tested

| Condition | Reaction Mix Volume per reaction (μL) | Taq Volume per reaction (μL) | Total Mastermix Volume per reaction (μL) | Sample Volume per reaction (μL) | Total Volume per reaction (μL) |
| --- | --- | --- | --- | --- | --- |
| 1 (Standard) | 19.83 | 0.17 | 20.00 | 5 | 25 |

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 26

{26}

| Condition | Reaction Mix Volume per reaction (μL) | Taq Volume per reaction (μL) | Total Mastermix Volume per reaction (μL) | Sample Volume per reaction (μL) | Total Volume per reaction (μL) |
| --- | --- | --- | --- | --- | --- |
| 2 | 18.64 | 0.17 | 19.81 | 5 | 23.81 |
| 3 | 21.02 | 0.17 | 21.19 | 5 | 26.19 |
| 4 | 19.83 | 0.16 | 19.99 | 5 | 24.99 |
| 5 | 19.83 | 0.18 | 20.01 | 5 | 25.01 |
| 6 | 19.83 | 0.17 | 20.00 | 4.7 | 24.70 |
| 7 | 19.83 | 0.17 | 20.00 | 5.3 | 25.3 |

This represents the total error that can be introduced by pipetting calculated by relative accuracy and precision stated in the pipette specifications.

Mutant samples were tested with a total of 12 replicates were generated per volume condition. WT samples were tested with one replicate per run. A total of 10 replicates were generated per volume condition. The PC and NTC were also be tested as a sample with the seven volume conditions. A total of 4 replicates were generated per volume condition.

Four assays (E542K, Q546E, E545A and H1047L) a 100.00% correct mutation call was observed for these assays, across all conditions tested. For the E545K assay, 11 out of 12 mutation calls were correct. One mutation call for condition 6 gave a double positive mutation status. E545K was detected with a ΔCt value of 2.60 and E545D was reported with a ΔCt 7.05. The impact of volumetric variation on assay performance was low as the change in ΔCt was small (-0.480ΔCt) and no false negative mutation calls were made within the conditions tested.

## c. Guardband Mixing

To determine the effect of varying reagent mixing, conditions that could be potentially introduced by the user, e.g. no mixing and vortexing, were evaluated against the standard, inversion 10 times. Six samples, a representation of each of the PIK3CA Kit tubes (five mutants and one WT) were tested.

The standard mixing conditions as stated in the instructions for use is to mix all reagents by inverting each tube 10 times. The following mixing conditions (as could be introduced by the user) were tested for, once a master mix containing both the reaction mix (Tubes 1 to 6) and the Taq DNA polymerase was made:

- Inversion 10 times (standard condition per the instructions for use (IFU))
- Vortex

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 27

{27}

- No mixing

Intermittent drop outs of mutation positive samples for the no mixing condition were observed for the E542K and E545K assays. The difference in mean $\Delta \mathrm{Ct}$ values for no mixing-inversion for the Q546R assay was 1.578. No difference was observed for all other assays tested.

In conclusion, this study indicated that there is an impact on assay performance if mixing is not performed correctly. The Instructions For Use states that the required mixing method is inversion 10 times. However, this study showed that mixing by vortex is also a suitable method.

d. Thawing guardband

The objective of this study was to determine the effect of varying the thaw time of the therascreen PIK3CA RGQ PCR Kit reagents and samples. Six samples, a representation of each of the PIK3CA Kit tubes (five mutants and one WT) were tested.

The tolerance of reaction mix, Taq and sample was tested against two thawing conditions, thaw time between $\geq 1$ and $\leq 4.5$ hours.

Both thawing conditions were tested on each run. Mutant samples were tested with six replicates per run over two runs and a total of 12 replicates were generated per condition. WT samples were tested with three replicates per run over four runs. A total of 12 replicates were generated per condition. Both PC and NTC were tested with one replicate per run for each condition on mutant runs and one replicate per run for condition 1 on WT runs.

The thawing of kits and templates was staggered, with condition 1 reaction mix removed for thawing 3.5 hours after condition 2 and 1 hour prior to set-up. The time of removal from the freezer and set-up start time was recorded in the worksheets for each condition. During set-up, master mix was prepared in duplicate and in bulk for some runs.

For each mutation sample, the differences in mean $\Delta \mathrm{Ct}$, between the instructions for use stated thaw time of 1 hour, and the guardband tested thaw time of 4.5 hours were within acceptable limits. The objective of the study was met; the impact of thaw time between 1 - 4.5 hours on assay performance was determined to be low as the change in $\Delta \mathrm{Ct}$ was small (i.e. within the pre-specified limits), and no incorrect mutation calls were made for either mutant or WT samples.

e. Set-up guardband

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 28

{28}

The objective of this study was to determine the effect of varying set-up time on the mutation status called by the PIK3CA Kit. Six samples, a representation of each of the PIK3CA Kit tubes (five mutants and one WT) were tested.

Due to complicated run layouts for some studies it can take up to 1.5 hours to set-up. The Instructions For Use state a run should be set-up and ran on the RGQ immediately, the standard set-up condition was Condition 1. The following conditions, which could be introduced by the user, were tested in this study:

1. Set-up and immediate RGQ run (Standard Condition)
2. Set-up (up to 1.5 hours at room temperature then RGQ run)
3. Set-up (up to 4.5 hours at room temperature then RGQ run)
4. Set-up (up to 7.5 hours at room temperature then RGQ run)
5. Set-up (up to 1.5 hours at room temperature), store for 16h at 2-8°C prior to performing RGQ run

This study involved the setting up of 20 runs with 4 runs per condition assessed, to determine the robustness of the assay to environmental stresses during set-up. Twelve replicates were tested per condition. The five conditions that were assessed were a standard run preparation and run immediately on an RGQ, followed by four non-standard conditions where the runs were set-up at room temperature for a prolonged time frame with one condition, involving an overnight incubation step at 2 - 8°C. The set-up of each run consisted of noting down the time the set-up began and the end time being when the run file was created on the RGQ.

To establish the tolerance of the therascreen PIK3CA RGQ PCR Kit to set-up variations, the impact of set-up times on ΔCt and mutation call was determined. The data obtained shows that varying the set-up time has no effect on the mutation call of the kit except for H1047L for Condition 5. For H1047L the data shows that for Condition 5 has an impact on the ΔCt for H1047L. To mitigate the risk of a false mutation call, the instructions for use state that runs must not be left overnight. The data obtained shows that varying the set-up and storage time up to 7.5 hours at room temperature followed by an RGQ run, has no effect on the performance of the therascreen PIK3CA RGQ PCR Kit. The recommended maximum set-up time is 8.5 hours at room temperature; this includes 1-hour thaw time and 7.5 hours for set-up and storage. The recommended maximum set-up time is 8.5 hours at room temperature; this includes 1-hour thaw time and 7.5 hours for set-up and storage

9. Cross-contamination

The objective of this study was to demonstrate the absence of PCR cross contamination of the WT samples by mutant samples within the DNA extraction and

PMA P190001: FDA Summary of Safety and Effectiveness Data
Page 29

{29}

run set-up procedure. The study focused on the DNA extraction of FFPE samples utilizing one kit lot of the FFPE Extraction Kit to identify any cross contamination associated with routine use of the PIK3CA Kit. Contamination could potentially occur at any stage of the testing procedure. This study was designed to investigate the probability of cross contamination during the whole testing procedure (DNA extraction and subsequent PIK3CA Kit).

This study was performed with H1047R (the mutation with highest prevalence) and WT FFPE cell line samples. Two independent sets of samples referred to as "Set A" and "Set B" were extracted following a pre-defined extraction matrix.

Two operators performed the extractions. Eighteen extractions (9 per set) were carried out for the mutation positive (H1047R) samples. 42 extractions (21 per set) were carried out for the WT samples. The extracts were mutation assessed across ten PCR runs; five per sample set were set-up consecutively by the same operator using the same equipment and RGQ instrument, with no other runs set-up using this instrument between these runs. Extracts were tested with the Control assay reaction mix (PIK3CA Kit Tube 1) and mutation of interest (PIK3CA Kit Tube 6).

Per the study design, ten mutation assessment runs were planned and initially carried out. Sufficient data points were collected from the 9 valid runs and they were submitted for statistical analysis. One sample was invalid due to no IC Ct in tube 6 reaction mix, but the sample was not repeat tested as sufficient data points from the associated extraction run had been collected for statistical analysis.

After removal of invalid data points, 189 WT replicates were available for statistical analysis. The observed percentage of correct mutation calls for WT samples was 100%, demonstrating no cross contamination of the WT samples by mutant samples sharing the same DNA extraction and run set up procedure.

## 10. Stability Studies

### a. Stability of FFPE Clinical Specimens and Extracted DNA from FFPE Specimens

The objective of this study was to demonstrate the long-term stability of s…