THERASCREEN EGFR RGQ PCR KIT

{0} SUMMARY OF SAFETY AND EFFECTIVENESS DATA (SSED) I. GENERAL INFORMATION Device Generic Name: Real-time PCR test Device Trade Name: therascreen® EGFR RGQ PCR Kit Device Procode: OWD Applicant’s Name and Address: QIAGEN Manchester Ltd Skelton House, Lloyd Street North Manchester, M15 6SH United Kingdom Date(s) of Panel Recommendation: None Premarket Approval Application (PMA) Number: P120022 Date of FDA Notice of Approval: July 12, 2013 Expedited: Granted priority review status on January 15, 2012 because the device addresses an unmet medical need, as demonstrated by significant clinically meaningful advantage. II. INDICATIONS FOR USE The therascreen® EGFR RGQ PCR Kit is a real-time PCR test for the qualitative detection of exon 19 deletions and exon 21 (L858R) substitution mutations of the epidermal growth factor receptor (EGFR) gene in DNA derived from formalin-fixed paraffin-embedded (FFPE) non-small cell lung cancer (NSCLC) tumor tissue. The test is intended to be used to select patients with NSCLC for whom GILOTRIF™ (afatinib), an EGFR tyrosine kinase inhibitor (TKI), is indicated. Safety and efficacy of GILOTRIF™ (afatinib) have not been established in patients whose tumors have L861Q, G719X, S768I, exon 20 insertions, and T790M mutations, which are also detected by the therascreen® EGFR RGQ PCR Kit. Specimens are processed using the QIAamp® DSP DNA FFPE Tissue Kit for manual sample preparation and the Rotor-Gene® Q MDx instrument for automated amplification and detection. III. CONTRAINDICATIONS None. PMA P120022: FDA Summary of Safety and Effectiveness Data {1} PMA P120022: FDA Summary of Safety and Effectiveness Data Page 2 ## IV. WARNINGS AND PRECAUTIONS The warnings and precautions can be found in the QIAGEN therascreen® EGFR RGQ PCR Kit labeling. ## V. DEVICE DESCRIPTION The following components comprise the overall device: - QIAGEN QIAamp® DSP DNA FFPE Tissue Kit - QIAGEN therascreen® EGFR RGQ PCR Kit - QIAGEN Rotor-Gene Q MDx Software version 2.1.0 build 9, and EGFR Assay Package version 1.1.2 ### Specimen Preparation Formalin-fixed, paraffin-embedded (FFPE) blocks are sectioned onto glass slides. A stained slide is used to confirm that there is tumor present. Two non-stained tissue sections are scraped from the slide for DNA extraction. DNA is manually extracted and purified from two 5-10 µm glass-mounted sections of FFPE tissue taken from NSCLC patients using the QIAGEN QIAamp® DSP DNA FFPE Tissue Kit. The tumor tissue is deparaffinized with xylene and the xylene is extracted with ethanol. The sample is lysed under denaturing conditions with proteinase K for one hour. The sample is heated at 90°C to reverse formalin cross-linking of genomic DNA. The sample is passed through a silica-based membrane so that genomic DNA binds to the membrane and contaminants are removed. Purified genomic DNA is eluted from the membrane into 120-µl of elution buffer. Extracted DNA is stored at -20°C. ### PCR Amplification and Detection The QIAGEN therascreen® EGFR RGQ PCR Kit contains reagents for eight separate reactions; seven mutation-specific reaction mixes to amplify and detect mutations in codons 18, 19, 20 and 21 of the EGFR oncogene, and one Control Reaction mix that amplifies and detects a region of exon 2 in the EGFR oncogene. The seven mutation-specific reaction mixes are the exon 19 Deletions, L858R, T790M, L861Q, G719X, S768I and exon 20 Insertions reaction mixes. Each reaction in the therascreen® EGFR RGQ PCR Kit makes use of an amplification refractory mutation system (ARMS) allele specific polymerase chain reactions (PCR) to selectively amplify mutated genomic DNA templates (mutation-positive) in a background of non-mutated genomic DNA (mutation-negative; wild-type) combined with a fluorophore-labeled Scorpion primer to detect any resultant amplification product. ARMS technology exploits the ability of Taq polymerase to distinguish between a match and a mismatch at the 3' end of a PCR primer. Scorpions are bi-functional molecules containing a PCR primer covalently linked to a probe. The probes incorporate both a fluorophore (carboxyfluorescein [FAM™]) and a quencher which quenches the fluorescence of the fluorophore. During PCR, when the probe binds to the ARMS {2} amplicon, the fluorophore and quencher become separated leading to a detectable increase in fluorescence. Before testing with the mutation-specific reaction mixes, each DNA sample must be tested with the Control Reaction mix to determine whether the quality and quantity of DNA is sufficient and appropriate for the working range of the assay. The Control Reaction Ct value is used to assess the total amplifiable DNA in a sample and must fall within pre-specified ranges for each sample. The interpretation of the results obtained from the Control Reaction is as follows: | Control Ct value | Interpretation | Action | | --- | --- | --- | | > 31.10 | Quantity of amplifiable DNA is not sufficient for mutation analysis. | Additional samples should be extracted and tested | | < 23.70 | Quantity of amplifiable DNA is too high for mutation analysis. | Dilute with the sample diluent water supplied in the kit | | 23.70 ≤ Control Ct ≥ 31.10 | Quantity of amplifiable DNA is suitable for mutation analysis | --- | The run 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 run parameters are: - Hold at 95°C for 15 minutes to activate the Taq polymerase; - PCR for 40 cycles of 95°C for 30 seconds, to denature, and 60°C for 1 minute, to anneal/extend. The PCR cycle at which the fluorescence from a particular reaction crosses the predefined threshold value is defined as the Ct value. The therascreen® EGFR RGQ PCR Kit detects the following mutations in exon 19 of the EGFR gene: 2235_2249del15, 2235_2252&gt;AAT, 2237_2255&gt;T, 2236_2250del15, 2238_2255del18, 2238_2248&gt;GC, 2238_2252&gt;GCA, 2239_2253del15, 2239_2256del18, 2239_2248TTAAGAGAAG&gt;C, 2239_2258&gt;CA, 2240_2257del18, 2240_2254del15, 2239_2251&gt;C, as well as the L858R substitution mutation 2573 T&gt;G in exon 21 of the EGFR gene. The therascreen® EGFR RGQ PCR Kit also detects the following mutations where safety and efficacy of GILOTRIF™ (afatinib) has not been established: T790M (2369C&gt;T), L861Q (2582T&gt;A), G719A (2156G&gt;C), S768I (2303G&gt;T), and exon 20 insertions (2319_2320insCAC and 2310_2311insGGT). The list of mutation is also included in "SUMMARY OF PRECLINICAL STUDIES" section below. ## Test Controls Each test run must contain an Internal Control (IC, already included in the PCR reaction mixes), a Positive Control (PC), and a No Template Control (NTC, Negative Control) tested with each reaction mix. A test run is considered invalid if any NTC indicates that the test run has been contaminated (one or more Ct values below a set value for the green FAM channel or outside a set range for the yellow Hex channel), or if a PC Ct value lies outside a set range (green FAM channel). For EGFR Kit test runs to be accepted as valid, PMA P120022: FDA Summary of Safety and Effectiveness Data {3} the RGQ software requires Ct values for PC, and NTC, meet validity criteria specified in the EGFR Locked Templates as shown in table below. Run Validity Criteria | Reaction Control | Reaction Mix | Channel | Ct Range* | | --- | --- | --- | --- | | PC | Control | FAM (green) | 28.13 – 34.59 | | | T790M | FAM (green) | 30.22 – 34.98 | | | Deletions | FAM (green) | 28.90 – 34.90 | | | L858R | FAM (green) | 29.97 – 34.81 | | | L861Q | FAM (green) | 28.49 – 34.02 | | | G719X | FAM (green) | 29.42 – 34.19 | | | S768I | FAM (green) | 28.98 – 35.19 | | | Insertions | FAM (green) | 27.92 – 34.09 | | NTC | All 8 reaction mixes | FAM (green) | No Amp (i.e. not <40.00) | | | | HEX (yellow) | 29.85 – 35.84 | *Ranges are inclusive. # Internal Control: All eight reaction mixes contain an additional ARMS primer and a HEX-labeled Scorpions primer for the amplification and detection of a synthetic non EGFR related oligonucleotide template that is used as an Internal Control. The Scorpions primer is labeled with HEX to distinguish from the FAM-labeled Scorpions in the Control and mutation assays. In each reaction mix, the Internal Control Reaction is designed to be the weaker of the two reactions. This is achieved through the use of a very low concentration of Internal Control template. The Internal Control Reaction is designed to work independently of mutation-specific amplification, but can fail in the presence of strong amplification if it is "outcompeted" by the FAM reaction. A no mutation detected result with a failed Internal Control Reaction in any one of the seven mutation assays will be reported as an invalid result. The Internal Control is used to detect inhibitors or gross reaction failures. # Positive Control (PC): The Positive Control comprises a mixture of synthetic oligonucleotides representing the mutations detected by the therascreen® EGFR RGQ PCR Kit. Detection of the Positive Control confirms the proper functioning of each of the reaction mixes in the Kit. Of the multiplex reaction mixes (exon 19 Deletions, G719X, Insertions), one mutation from each is included in the Positive Control. # Negative Control (NTC): The therascreen® EGFR RGQ PCR Kit contains nuclease-free water to be used as a No Template Control (NTC) reaction. The NTC serves as a negative control and assesses potential contamination during assay setup. # Instrument and Software The Rotor-Gene Q (RGQ) MDx instrument is a real-time PCR analyzer designed for thermocycling and real-time detection of amplified DNA. The RGQ MDx instrument controls and monitors PCR reactions and includes the software that determines mutation PMA P120022: FDA Summary of Safety and Effectiveness Data {4} status based upon PCR results. It incorporates a centrifugal rotor design for thermal cycling during PCR reactions where each tube spins in a chamber of moving air. 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 for a total of 40 cycles for each PCR run. In the RGQ MDx Instrument, samples are excited from the bottom of the chamber by a light emitting diode. Energy is transmitted through the thin walls at the base of the 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 past the excitation/detection optics every 150 milliseconds. The fluorescence signals monitor the progress of the PCR reactions. The instrument is capable of supporting up to six optical channels (six excitation sources and six detection filters), however only two of these channels (the yellow HEX and green FAM channels) are used with the therascreen® EGFR RGQ PCR Kit. The therascreen® EGFR Assay Package consists of two templates: the "therascreen EGFR Control Run Locked Template" (for DNA sample assessment) and the "therascreen EGFR Locked Template" (for detection of EGFR mutations). These templates contain the PCR run parameters, assess run validity and calculate the results. The same run parameters are used for both the DNA sample assessment with the Control Reaction Mix and for detection of EGFR mutations using the mutation assays. The RGQ MDx instrument software supports real-time analysis procedures. The software determines Ct values, calculates $\Delta \mathrm{Ct}$ values, and compares these to the mutation-specific cut-off values incorporated into the software as described above. A system of Flags/Warnings is embedded within the software in order to inform the user of potential problems with the assay and to indicate non-valid test runs or non-valid samples within a valid test run (inappropriate level of DNA or Internal Control failure). No results are reported for invalid runs or for non-valid samples. The $\Delta \mathrm{Ct}$ values will be reported to the users; however, users do not have access to the raw data. ## Interpretation of Results The Ct for the Control Reaction reflects the total amount of amplifiable EGFR template in the sample, while the Ct for each mutation-specific reaction reflects the amount of EGFR mutation within the sample. The difference in Ct values ($\Delta \mathrm{Ct}$) between the Control Reaction and the mutation-specific reaction (mutation assay) indicates the proportion of mutation within the sample. The $\Delta \mathrm{Ct}$ value reduces and approaches to 0 as the proportion of mutant DNA in the samples increases, a $\Delta \mathrm{Ct}$ value may be negative owing to individual assay variability. The $\Delta \mathrm{Ct}$ value increases as the proportion of mutant DNA in the sample decreases. When the $\Delta \mathrm{Ct}$ value exceeds the $\Delta \mathrm{Ct}$ cut-off values for all seven mutation assays, the assay reports "No Mutation Detected". For each sample, a calculation is performed by the RGQ software to determine the $\Delta \mathrm{Ct}$ value (FAM channel) for each of the 7 mutation-specific reactions: $$ \Delta \mathrm{Ct} = [\text{Mutation Reaction Ct Value}] - [\text{Control Reaction Ct Value}] $$ PMA P120022: FDA Summary of Safety and Effectiveness Data {5} Each sample will have seven possible $\Delta C_t$ values (one per mutation). These values are compared to pre-established specifications (cut-off values) incorporated into the RGQ software to determine whether a sample is "Mutation Detected" or "No Mutation Detected", and which mutation(s), if any, is present. When the mutation assay $\Delta C_t$ value is less than or equal to the cut-off value for that reaction, the sample is EGFR mutation-positive. The assay results will be displayed as "Mutation Detected" (MD), "No Mutation Detected" (NMD), "Invalid" or, if a run control fails, "Run Control Failed". For samples with "Mutation Detected" results, specific mutations are reported, i.e., exon 19 Deletions (Del), L858R, T790M, L861Q, G719X, S768I, or exon 20 Insertions (Ins). It is possible for a mutation positive sample to have more than one out of the seven possible EGFR mutations reported. The multiplex assays, i.e., Del, G719X, and Ins, do not distinguish among the mutations they detect. Mutation Assay Cut-Offs (ΔCt) | Mutation Assay | T790M | Del* | L858R | L861Q | G719X | S768I | Ins** | | --- | --- | --- | --- | --- | --- | --- | --- | | Cut-Off (ΔCt) | ≤7.40 | ≤8.00 | ≤8.90 | ≤8.90 | ≤8.90 | ≤8.90 | ≤8.00 | * Exon 19 Deletions ** Exon 20 Insertions ## VI. ALTERNATIVE PRACTICES AND PROCEDURES There are no other FDA-cleared or approved alternatives for EGFR mutation testing of formalin-fixed, paraffin-embedded (FFPE) NSCLC tissue for the selection of patients who are eligible for first-line treatment with GILOTRIF™ (afatinib). ## VII. MARKETING HISTORY The QIAGEN therascreen® EGFR 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 as expected or failure to correctly interpret test results may lead to incorrect EGFR test results and subsequently improper patient management decisions in non-small cell lung cancer (NSCLC) treatment. For the specific adverse events related to GILOTRIF™ (afatinib) that occurred in the clinical studies, please see Section X below. ## IX. SUMMARY OF PRECLINICAL STUDIES ### A. Laboratory Studies PMA P120022: FDA Summary of Safety and Effectiveness Data {6} The specific performance characteristics of the QIAGEN therascreen® EGFR RGQ PCR Kit (henceforth referred to as EGFR Kit) were determined by studies using formalin-fixed, paraffin-embedded (FFPE) tissue specimens collected from NSCLC patients, and FFPE human cell lines (FFPE cell lines). The FFPE cell lines were generated using a lung carcinoma cell line (A549) to produce cell lines harboring the desired specific EGFR mutations. Bi-directional Sanger sequencing and massively parallel sequencing were used to select the specimens for the following studies, and to determine the percentage of mutation in NSCLC FFPE samples. The similarity between FFPE clinical specimens and FFPE cell lines was demonstrated by comparing assay amplification efficiencies (AE) between the two sample types, and by assessing the Limit of Detection (LoD) for the specific mutation assays. FFPE cell lines were sectioned and processed similar to FFPE clinical specimens. DNA was extracted and tested according the instructions for use. 1. Correlation with Reference Method To demonstrate the accuracy of the EGFR Kit relative to bi-directional Sanger sequencing, accuracy studies were conducted with two sets of samples (a) procured clinical specimens and (b) clinical trial specimens from the Phase 3 study for afatinib, the 1200.32 clinical trial. a. Procured Specimens The procured specimen study was a blinded study using procured FFPE clinical specimens from NSCLC patients. EGFR testing was performed on DNA samples extracted from 373 specimens with bi-directional Sanger sequencing results for exons 18, 19, 20 and 21 that had been blinded to the operators. A total of 366/373 samples (98.1%) produced valid EGFR Kit results and the bi-directional Sanger sequencing results were then unblinded. Analysis of concordance and discordant results were performed on this data set. Concordant results were obtained with 343/366 (93.7%) of procured samples, while discordant results were obtained with 23/366 (6.3%) of procured samples. Any discordant results and an equal number of concordant samples were subject to discordant analyses. The overall results are shown below. PMA P120022: FDA Summary of Safety and Effectiveness Data Page 7 {7} EGFR Kit Compared to Sanger Sequencing – Procured Specimens | | Mutation Call by Bi-Directional Sanger Sequencing | | | | | | | | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | Del | Del & L858R | G719X | Ins | L858R | L861Q | T790M | T790M & L858R | S768I | WT | Total | | EGFR Kit Call | Del | 52 | 3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 7 | 62 | | | Del & L868R | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 2 | | | G719X | 0 | 0 | 4 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 5 | | | Ins | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 3 | 4 | | | L858R | 0 | 0 | 0 | 0 | 58 | 0 | 0 | 0 | 0 | 7 | 65 | | | L861Q | 0 | 0 | 0 | 0 | 0 | 2 | 0 | 0 | 0 | 0 | 2 | | | T790M | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | | | T790M & L868R | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 1 | | | S768I | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | | | WT | 0 | 0 | 1 | 0 | 3 | 0 | 0 | 0 | 0 | 220 | 224 | | | Total | 52 | 3 | 5 | 1 | 62 | 2 | 1 | 1 | 0 | 239 | 366 | Procured samples with both Sanger and EGFR Kit valid results were analyzed to assess the overall percent agreement (OPA), positive percent agreement (PPA) and negative percent agreement (NPA) based on agreement between the two methods for overall mutation status, i.e. that the samples either have a positive mutation, regardless of which mutation it is, or have a no mutation detected result. Additional agreement analyses were also conducted for each of the 7 EGFR mutation assays in the EGFR Kit (data not shown). These percentages, together with the corresponding two-sided exact 95% confidence intervals (95% CI) were reported. The results demonstrate a PPA of 96.9%, a NPA of 92.1% and an OPA of 93.7%. EGFR Kit vs. Sanger Concordance in Procured Specimens | Measure of Agreement | Percent Agreement % (N) | 95% CI* | | --- | --- | --- | | Positive Percent Agreement (PPA) | 96.9% (123/127) | 92.1, 99.1 | | Negative Percent Agreement (NPA) | 92.1% (220/239) | 87.9, 95.1 | | Overall Percent Agreement (OPA) | 93.7% (343/366) | 90.7, 96.0 | *95% CI is calculated based on exact method. For the 23 overall mutation status discordant results, 4 (17.4%) sample gave wild-type (i.e., no mutation detected) results by the EGFR Kit and gave mutation detected results by Sanger sequencing, while 19 (82.6%) samples gave mutation detected results by the EGFR Kit and gave wild-type results by Sanger sequencing. PMA P120022: FDA Summary of Safety and Effectiveness Data {8} b. Phase 3 Clinical Trial Specimens This was a blinded study using FFPE clinical specimens from a subpopulation of patients in the 1200.32 clinical trial. The baseline clinical and demographic characteristics of the patients whose specimens were available for this retrospective testing were comparable to those of otherwise eligible patients whose specimens were not available for retesting. EGFR testing was performed on DNA samples extracted from 360 specimens with bi-directional Sanger sequencing results for exons 18, 19, 20 and 21 that had been blinded to the operators. Concordant results were obtained with 332/360 (92.2%) of samples, while discordant results were obtained with 28/360 (7.8%) of samples. Discordant results were not further evaluated or re-tested as the quantity of DNA remaining was very low. The overall results are shown below. EGFR Kit Compared to Sanger Sequencing – the 1200.32 Clinical Trial Specimens | EGFR Kit Call | Sanger Mutation Status | | --- | --- | | Del | Del & T790M | G719X | Insertions | L858R | L861Q | T790M | T790M & L858R | S768I | L858R & S768I | G719X & S768I | WT | Total | | Deletions | 76 | - | - | - | 1 | - | - | - | - | - | - | 12 | | Del & T790M | 2 | 1 | - | - | - | - | - | - | - | - | - | 0 | | G719X | - | - | 1 | - | - | - | - | - | - | - | - | 1 | | Insertions | - | - | - | 1 | - | - | - | - | - | - | - | 3 | | L858R | - | - | - | - | 69 | - | - | - | - | - | - | 10 | | L861Q | - | - | - | - | - | 1 | - | - | - | - | - | 1 | | T790M | - | - | - | - | - | - | - | - | - | - | - | 0 | | T790M & L858R | - | - | - | - | - | - | - | 1 | - | - | - | 0 | | S768I | - | - | - | - | - | - | - | - | 1 | - | - | 0 | | L858R & S768I | - | - | - | - | - | - | - | - | - | 2 | - | 0 | | G719X & S768I | - | - | - | - | - | - | - | - | - | - | 1 | 0 | | WT | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 175 | | Total | 79 | 1 | 1 | 1 | 70 | 1 | 0 | 1 | 1 | 2 | 1 | 202 | PMA P120022: FDA Summary of Safety and Effectiveness Data Page 9 {9} Samples with both Sanger and EGFR Kit valid results were analyzed to assess the OPA, PPA and NPA based on agreement between the two methods for overall mutation status. These percentages, together with the corresponding two-sided exact 95% CI are summarized below. Additional agreement analyses were also conducted for each of the 7 mutation assays in the EGFR Kit (data not shown). These percentages, together with the corresponding two-sided exact 95% CI were reported. The results demonstrate a PPA of 99.4%, a NPA of 86.6% and an OPA of 92.2%. ## EGFR Kit vs. Sanger Concordance in the 1200.32 Clinical Trial Specimens | Measure of Agreement | Percent Agreement % (N) | 95% CI* | | --- | --- | --- | | Positive Percent Agreement (PPA) | 99.4% (157/158) | 96.5, 100.0 | | Negative Percent Agreement (NPA) | 86.6% (175/202) | 81.2, 91.0 | | Overall Percent Agreement (OPA) | 92.2% (332/360) | 89.0, 94.8 | *95% CI is calculated based on exact method. For the 28 overall mutation status discordant results, 1 (3.6%) sample with an EGFR mutation as determined by Sanger sequencing gave a wild-type (i.e., no mutation detected) result by the EGFR Kit, while 27 (96.4%) samples gave mutation detected results by the EGFR Kit and gave wild-type results by Sanger sequencing. ## 2. Analytical Sensitivity ### a. Analytical Sensitivity – Limit of Blank (LoB) To assess performance of the EGFR Kit in the absence of template and to ensure that a blank sample or a sample with wild-type DNA does not generate an analytical signal that might indicate a low concentration of mutation, samples with no template and NSCLC FFPE EGFR wild-type cell line DNA were evaluated. The results demonstrated no positive mutation calls with No Template Control (NTC) samples and FFPE wild-type samples i. Limit of Blank (LoB) no template – The No Template Control (NTC, nuclease-free water) as supplied in the EGFR Kit was run as the sample with no template. All NTC tested results in each mutation assay (n=14/14 per mutation assay) were valid and were “No Mutation Detected”. ii. Limit of Blank (LoB) FFPE Specimens – High input DNA from NSCLC FFPE EGFR wild-type cell line was tested (n=60). All valid results (n=59) were “No Mutation Detected”. One PMA P120022: FDA Summary of Safety and Effectiveness Data Page 10 {10} replicate of the 60 tested was invalid according to the run criteria of the EGFR Kit and was excluded from analysis. b. **Analytical Sensitivity – Limit of Detection (LoD)** The LoD is the minimum percentage of mutant DNA that can be detected in a background of wild-type DNA when the total amplifiable DNA (within the input range) produced correct mutation calls at 95% for each mutation positive sample (C95). The DNA input working range for the therascreen® EGFR RGQ PCR Kit is based on the Control Reaction Ct value at the pre-specified range of 23.70 to 31.10 (also refer to section 2c below), which is used to indicate the amount of amplifiable DNA in a sample. The EGFR Kit does not use a specific concentration of DNA as determined by spectrophotometry. Multiple studies were conducted sequentially to estimate and determine LoD for each EGFR mutation. The initial study estimated the LoD across a broad dilution series (eight replicates per dilution). Based on the results of the initial study, subsequent studies were conducted to determine the LoD across a narrow dilution series (24 replicates per dilution). For each EGFR mutation, percentages of correct calls were assessed across dilution levels. The LoD was determined at both the high DNA input (Control Ct ~25.50) and at low DNA input (Control Ct ~30.1) levels. The NSCLC FFPE clinical specimens or FFPE cell lines were used in the LoD study. The mutant DNA, extracted from the FFPE clinical specimens or FFPE cell lines, was diluted in a background of wild-type DNA in order to create a series of samples containing different percentages of mutant DNA. At each dilution level (% mutation), 24 replicates were evaluated using multiple EGFR Kit lots. For each mutation, the rate of correct calls by % mutation, mean control Ct value, mean mutation Ct value, and mean ΔCt value were tabulated by % mutation. An example of such table for the exon 21 L858R mutation is shown below. PMA P120022: FDA Summary of Safety and Effectiveness Data Page 11 {11} Rate of Correct Mutation Calls by % Mutation for Exon 21 L858R Mutation | L858R | % Mutation (Dilution) | Rate of Correct Calls | Mean Control Ct Value (Range) | Mean Mutation Ct Value (Range) | Mean ΔCt Value (Range) | | --- | --- | --- | --- | --- | --- | | Low DNA Input | 17.06 | 24/24 | 30.29 (29.92-30.57) | 35.14 (34.44-37.41) | 4.85 (4.05-7.26) | | | 12.15 | 24/24 | 30.21 (29.59-30.74) | 35.57 (34.95-37.49) | 5.36 (4.46-7.19) | | | 8.69 | 24/24 | 30.24 (29.60-30.97) | 36.10 (35.04-38.12) | 5.86 (4.76-8.16) | | | 6.21 | 23/24 | 30.17 (29.90-30.60) | 36.62 (35.13-39.29) | 6.45 (4.53-9.34) | | | 4.46 | 20/24 | 30.21 (29.62-30.95) | 37.39 (35.62-39.68) | 7.19 (5.35-9.71) | | | 3.15 | 17/24 | 30.31 (29.78-30.70) | 37.44 (36.14-38.43) | 7.16 (5.75-8.65) | | | 2.25 | 12/24 | 30.13 (29.65-30.66) | 38.15 (36.39-39.51) | 8.02 (6.49-9.76) | | High DNA Input | 1.55 | 24/24 | 26.29 (25.97-26.71) | 33.61 (32.97-34.75) | 7.31 (6.42-8.38) | | | 1.10 | 21/24 | 25.62 (25.29-26.07) | 33.89 (33.21-34.97) | 8.27 (7.21-9.29) | | | 0.79 | 22/24 | 25.67 (25.31-26.03) | 34.04 (33.18-34.74) | 8.37 (7.15-9.30) | | | 0.56 | 19/24 | 26.00 (25.64-26.35) | 34.58 (33.67-35.87) | 8.59 (7.68-9.66) | | | 0.40 | 9/24 | 25.68 (25.40-26.03) | 34.76 (34.09-35.44) | 9.08 (8.35-9.90) | | | 0.29 | 8/24 | 25.70 (25.38-26.06) | 34.83 (34.05-35.73) | 9.13 (8.20-9.96) | | | 0.21 | 3/24 | 25.69 (25.37-26.04) | 34.91 (34.00-35.69) | 9.23 (8.39-9.89) | Utilizing the mutational call data, the logistic regression analysis was applied to each mutation individually using each of the low and high input DNA datasets. In these analysis models, the predicted probability of detection of 95% (p = 0.95) on the y-axis would intercept the fitted curve to read out a mutation percentage on the x-axis [log2 (% mutation dilution)]. Hence, the LoD value was determined as the % mutation dilution which gave an estimated probability of detection of 95% (C95). The LoD was determined for each EGFR mutation at either the low or high DNA input levels (data not shown). The final LoD claims listed in tables below indicate the percentage of mutation which gave a predicted probability of correct calls of 95% for each of the 21 mutations. The final LoD claims for EGFR mutations were fully supported by results from the reproducibility study conducted at low DNA input (refer to Repeatability and Reproducibility section below). Among the 21 EGFR mutations detected by the EGFR Kit, safety and efficacy of GILOTRIF™ (afatinib) have been established for the 15 mutations listed in the first table below, but have not been established for the 6 mutations listed in the second table below. Refer to GILOTRIF™ (afatinib) drug labeling for more details. PMA P120022: FDA Summary of Safety and Effectiveness Data {12} Sensitivity of the therascreen® EGFR RGQ PCR Kit – Safety and Efficacy of GILOTRIF™ (afatinib) Established | Exon | Mutation | Cosmic ID | Base Change | Sample Type^ | Final LoD Claim (% Mutation) | | --- | --- | --- | --- | --- | --- | | 19 | Deletions | 6220 | 2238_2255del18 | CL | 2.7 | | | | 6223* | 2235_2249del15 | CL+CS | 6.4 | | | | 6225* | 2236_2250del15 | CL+CS | 6.5$ | | | | 6254 | 2239_2253del15 | CS | 10.2$ | | | | 6255 | 2239_2256del18 | CS | 0.81$ | | | | 12369 | 2240_2254del15 | CS | 4.94 | | | | 12370 | 2240_2257del18 | CS | 8.1 | | | | 12382 | 2239_2248TTAAGAGAAG>C | CS | 1.45$ | | | | 12383 | 2239_2251>C | CS | 4.58 | | | | 12384 | 2237_2255>T | CS | 7.54$ | | | | 12387 | 2239_2258>CA | CL | 4.91 | | | | 12419 | 2238_2252>GCA | CL | 16.87 | | | | 12422 | 2238_2248>GC | CL | 3.24 | | | | 13551 | 2235_2252>AAT | CL | 4.24 | | 21 | L858R | 6224* | 2573T>G | CL+CS | 5.94 | Sensitivity of the therascreen® EGFR RGQ PCR Kit – Safety and Efficacy of GILOTRIF™ (afatinib) Not Established | Exon | Mutation | Cosmic ID | Base Change | Sample Type^ | Final LoD Claim (% Mutation) | | --- | --- | --- | --- | --- | --- | | 20 | T790M | 6240* | 2369C>T | CL+CS | 17.5% | | 21 | L861Q | 6213 | 2582T>A | CL | 9.24$ | | 18 | G719A | 6239 | 2156G>C | CL | 32.5$ | | 20 | S768I | 6241 | 2303G>T | CL | 7.66 | | 20 | Insertions | 12377 | 2319_2320insCAC | CL | 3.72$ | | | | 12378 | 2310_2311insGGT | CL | 19.96$ | ^ CS denotes FFPE clinical specimen; CL denotes FFPE cell line. * LoD was determined using both FFPE clinical specimens and FFPE cell line at low DNA input level for these 4 EGFR mutations, which covers 68.71% of the reported EGFR mutations. $ Final LoD % mutation claims are based on results from reproducibility study. The final LoD claims for EGFR mutations were fully supported by results from the reproducibility study conducted at low DNA input (refer to Repeatability and Reproducibility section below). The LoD values determined at the low DNA input were higher than those determined at high DNA input; this difference is expected as there is more abundant mutation DNA template at the higher DNA input level. The data overall support the final LoD claims listed above at the percent mutation which gave a predicted probability of correct calls of 95%. Although the analytical sensitivity studies have been conducted, the therascreen® PMA P120022: FDA Summary of Safety and Effectiveness Data Page 13 {13} EGFR RGQ PCR Kit is for the qualitative detection of the EGFR mutation and is not intended to for quantitative measurements. 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 range for use in assessing DNA sample validity. A set of 417 sectioned FFPE clinical specimen blocks were assayed and characterized using the EGFR Kit. Among these, 400 samples were wild-type and 17 samples were of mutation positive as determined by bi-directional Sanger sequencing. EGFR mutations representing five out of the seven mutation assays in the EGFR Kit were covered in this study. Samples for the remaining two mutation assays (i.e., G719X and S768I) were not available. Distribution of Control assay Ct values were tested for normality using graphical inspection and a standard test for normality. Data was not normally distributed therefore non-parametric methods were used. The nonparametric one-sided tolerance intervals used for each boundary were selected at 90% coverage with 99% confidence. The boundaries selected were further rounded following considerations in accordance with the user needs and risk management. The final Control Reaction Ct Working range selected for the EGFR Kit is determined to be 23.70 to 31.10 Ct. ii. Correlation to Relative Fluorescence Increase (RFI) – During development, the optimal analysis settings for the ‘Green’ (FAM) and ‘Yellow’ (HEX) Channels were determined to set thresholds, i.e., the mean fluorescence values which a curve must cross to be distinguished from background fluorescence as positive amplification. The thresholds at 0.075 for the ‘Green’ (FAM) channel and at 0.02 for the ‘Yellow’ (HEX) channel were chosen and applied during the analysis of 235 FFPE replicates, 35 NTC replicates and 35 Positive Control replicates. The analysis parameters were found to be suitable prior to assessing Ct values, and were configured in the locked templates contained within assay packages (i.e., operators would be unable to alter). iii. ΔCt Cut-offs – A risk-based approach was taken with regard to false positive rates when setting the assay cut-offs, and estimated LoB values were used as one component in developing cut-off values. The raw data plots and/or box plots of ΔCt were used as a visual aid based on distribution and separation of ΔCt values. The red-dashed line shows the cut-offs, above that are the mutation negative samples, and below that the mutation-positive samples. PMA P120022: FDA Summary of Safety and Effectiveness Data Page 14 {14} There is no boxplot for S768I mutation-negative samples as no amplification was observed with wild-type samples in this assay, therefore no $\Delta \mathrm{Ct}$ value could be calculated. There are no boxplots displayed for G719X and S768I, as there were no mutation positive samples available for G719X and S768I due to the rarity of the samples. For the T790M assay, a cut-off of 7.4 was chosen to enable greater sensitivity of the assay whilst still ensuring no false positive or negative calls would result. For the exon 19 Deletions assay, a cut-off of 8.0 was selected to enable greater sensitivity of the assay with a false positive rate of $0.7\%$. Refer to section V above on Interpretation of Results for the cut-off values selected.  Cut-off $\Delta \mathrm{Ct}$ Values by Mutation # 3. Analytical Sensitivity – Effect of DNA Input on $\Delta \mathrm{Ct}$ The DNA input level is defined as the total quantity of amplifiable EGFR DNA in a sample and is determined by the Ct values from the Control Reaction. When samples at different total DNA levels contain the same proportion of mutant DNA, it is expected that the measured $\Delta \mathrm{Ct}$ values will remain consistent. The objective of the study was to demonstrate that the performance of the EGFR Kit is consistent over the total DNA input (Control Ct) range of the assay. Nineteen (19) of the mutations detected by the EGFR Kit were tested, including all 7 EGFR mutation assays. DNA extracted from 19 FFPE cell lines was used to prepare pools of DNA at the lower end of the Control Reaction working range. For each point within the working range sufficient material was prepared to carry out 6 replicate tests. The target Ct were set for dilution 1 (100% or undiluted) for each mutation of Ct at $\sim 24.70$. For some mutations it was not possible to achieve a Ct of 24.70; thus, the lowest possible Ct was used. This pool of DNA was used to generate 6 equally spaced dilution levels across and PMA P120022: FDA Summary of Safety and Effectiveness Data {15} beyond the working range, resulting a dilution at approximately 1:3 ratio (i.e., each subsequent dilution level contained approximately 3 fold less DNA). The final dilution point was outside of the lowest DNA input level of the working range, Ct ~32-33 Ct (0.41%). For each of the EGFR mutations tested, the mean control Ct, mutant Ct and ΔCt values for each mutation assay at each dilution were displayed in tables below, as a function of the Control Ct ranges for the dilutions. Overall, the ΔCt values measured at different total DNA input levels were consistent across the working range of the EGFR Kit, and passed the pre-specified acceptance criteria for the study. Mean Control Ct Values across the Control Ct Range – FFPE cell lines | Sample Mutation | Relative DNA Input Level | | | | | | | --- | --- | --- | --- | --- | --- | --- | | | 0.41% | 1.23% | 3.70% | 11.11% | 33.33% | 100.00% | | WT | 33.34 | 31.38 | 29.76 | 28.22 | 26.09 | 24.74 | | T790M | 32.98 | 31.05 | 29.34 | 27.62 | 26.06 | 24.39 | | Del6210 | 33.53 | 31.28 | 29.64 | 27.90 | 26.48 | 24.65 | | Del6218 | 32.82 | 31.24 | 29.42 | 27.78 | 26.32 | 24.68 | | Del6220 | 33.84 | 31.81 | 30.19 | 28.38 | 26.59 | 24.97 | | Del6223 | 33.93 | 31.14 | 29.45 | 27.82 | 26.19 | 24.42 | | Del6225 | 32.83 | 31.03 | 29.51 | 27.80 | 26.18 | 24.48 | | Del12367 | 33.50 | 31.40 | 29.64 | 28.05 | 26.41 | 24.79 | | Del12387 | 33.31 | 31.47 | 29.70 | 28.01 | 26.41 | 24.73 | | Del12422 | 33.42 | 31.48 | 29.91 | 28.17 | 26.35 | 24.78 | | Del12419 | 33.42 | 31.19 | 29.35 | 27.78 | 26.17 | 24.58 | | Del12728 | 33.19 | 31.32 | 29.69 | 28.04 | 26.45 | 24.78 | | Del13551 | 32.86 | 30.98 | 29.51 | 27.89 | 26.22 | 24.62 | | L858R | 32.93 | 31.34 | 29.61 | 27.85 | 26.26 | 24.64 | | L861Q | 33.25 | 31.67 | 30.00 | 28.42 | 26.56 | 24.99 | | G719A | 33.26 | 31.63 | 29.97 | 28.32 | 26.88 | 25.00 | | S768I | 33.13 | 31.33 | 29.79 | 27.96 | 26.43 | 24.74 | | Ins12377 | 33.68 | 31.63 | 29.64 | 27.87 | 26.26 | 24.72 | | Ins12378 | 33.47 | 31.23 | 29.64 | 27.83 | 26.31 | 24.58 | PMA P120022: FDA Summary of Safety and Effectiveness Data Page 16 {16} Mean Mutant Ct Values across the Control Ct Range – FFPE cell lines | Sample Mutation | Relative DNA Input Level | | | | | | | --- | --- | --- | --- | --- | --- | --- | | | 0.41% | 1.23% | 3.70% | 11.11% | 33.33% | 100.00% | | T790M | 35.12 | 32.87 | 31.31 | 29.60 | 27.97 | 26.31 | | Del6210 | 36.18 | 34.29 | 32.50 | 30.91 | 29.19 | 27.60 | | Del6218 | 37.75 | 35.34 | 33.62 | 31.77 | 30.22 | 28.56 | | Del6220 | 33.00 | 30.98 | 29.39 | 27.52 | 25.90 | 24.22 | | Del6223 | 36.00 | 32.87 | 31.24 | 29.63 | 27.96 | 26.42 | | Del6225 | 34.54 | 32.70 | 31.23 | 29.53 | 27.91 | 26.27 | | Del12367 | 34.35 | 32.30 | 30.56 | 28.98 | 27.34 | 25.70 | | Del12387 | 35.09 | 33.20 | 31.46 | 29.82 | 28.19 | 26.63 | | Del12422 | 36.10 | 34.05 | 32.29 | 30.62 | 28.95 | 27.39 | | Del12419 | 37.92 | 36.12 | 34.32 | 32.71 | 31.12 | 29.68 | | Del12728 | 33.56 | 31.84 | 30.18 | 28.62 | 26.88 | 25.23 | | Del13551 | 33.32 | 31.68 | 30.24 | 28.43 | 26.76 | 25.24 | | L858R | 32.97 | 31.40 | 29.76 | 28.07 | 26.47 | 24.98 | | L861Q | 34.45 | 33.09 | 31.43 | 29.64 | 28.00 | 26.45 | | G719A | 36.22 | 34.19 | 32.25 | 30.58 | 28.96 | 27.34 | | S768I | 36.23 | 34.17 | 32.42 | 30.70 | 29.14 | 27.60 | | Ins12377 | 35.75 | 33.75 | 31.63 | 30.19 | 28.62 | 26.93 | | Ins12378 | 33.71 | 31.71 | 29.91 | 28.21 | 26.55 | 24.76 | PMA P120022: FDA Summary of Safety and Effectiveness Data Page 17 {17} Mean ΔCt Values across the Control Ct Range – FFPE cell lines | Sample Mutation | Relative DNA Input Level | | | | | | | --- | --- | --- | --- | --- | --- | --- | | | 0.41% | 1.23% | 3.70% | 11.11% | 33.33% | 100.00% | | T790M | 2.14 | 1.82 | 1.97 | 1.98 | 1.91 | 1.92 | | Del6210 | 2.64 | 2.96 | 2.86 | 3.01 | 2.71 | 2.95 | | Del6218 | 4.93 | 4.10 | 4.20 | 4.00 | 3.90 | 3.89 | | Del6220 | -0.84 | -0.84 | -0.80 | -0.87 | -0.69 | -0.75 | | Del6223 | 2.08 | 1.73 | 1.79 | 1.81 | 1.77 | 2.00 | | Del6225 | 1.71 | 1.67 | 1.72 | 1.74 | 1.73 | 1.79 | | Del12367 | 0.85 | 0.90 | 0.92 | 0.93 | 0.94 | 0.90 | | Del12387 | 1.78 | 1.72 | 1.75 | 1.81 | 1.78 | 1.89 | | Del12422 | 2.68 | 2.56 | 2.39 | 2.45 | 2.60 | 2.61 | | Del12419 | 4.60 | 4.93 | 4.98 | 4.93 | 4.96 | 5.11 | | Del12728 | 0.37 | 0.52 | 0.49 | 0.57 | 0.43 | 0.45 | | Del13551 | 0.45 | 0.70 | 0.73 | 0.53 | 0.53 | 0.62 | | L858R | 0.04 | 0.06 | 0.15 | 0.23 | 0.21 | 0.34 | | L861Q | 1.19 | 1.42 | 1.43 | 1.22 | 1.44 | 1.46 | | G719A | 2.96 | 2.56 | 2.28 | 2.26 | 2.08 | 2.34 | | S768I | 3.10 | 2.84 | 2.63 | 2.74 | 2.71 | 2.85 | | Ins12377 | 2.07 | 2.12 | 1.98 | 2.33 | 2.35 | 2.21 | | Ins12378 | 0.23 | 0.48 | 0.28 | 0.37 | 0.24 | 0.21 | The dilution levels were actually different for Ins12376 than the other mutation samples, and these are reported in a separate table below. Dilution points are labeled as relative percent DNA input level, with the undiluted FFPE cell line specified as 100%. PMA P120022: FDA Summary of Safety and Effectiveness Data Page 18 {18} Mean Control, Mutant and ΔCt Values for Insertions 12376 across the Control Ct Range – FFPE cell lines | Sample Mutation | Relative DNA Input Level | | | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | | | | 1.80% | 4.01% | 8.62% | 20.04% | 45.04% | 100.00% | | Ins12376 | Control Ct | 31.69 | 30.48 | 29.25 | 27.96 | 26.84 | 25.65 | | Ins12376 | Mutant Ct | 34.86 | 33.62 | 32.06 | 30.95 | 29.78 | 28.75 | | Ins12376 | Delta Ct | 3.16 | 3.13 | 2.81 | 3.00 | 2.94 | 3.10 | ## 4. Linearity ### a. Amplification Efficiency as a Function of DNA Input The linearity and amplification efficiency of PCR for each mutation assay, relative to the Control Reaction, across the working range of the EGFR Kit was investigated. Amplification efficiency was calculated using linear regression with assay Ct as the response variable and $\log_2$ relative DNA input level as the explanatory variable for 19 EGFR mutations detected by the EGFR Kit including all the 7 EGFR mutation assays, and the Control Reaction. $$ \text{Amplification Efficiency} = \left[2^{**}(-1 / \text{slope})\right] - 1 $$ (** indicates to the power of) EGFR mutations were tested targeting the lower end of the Control Reaction working range (~25Ct, high DNA input) and were serially diluted with ATE buffer, effectively diluting the input DNA and mutant DNA equally. The final dilution point was outside of the lowest DNA input level of the working range, Ct ~32-33Ct. The largest difference in the amplification efficiencies (AE) between the Control Reaction and a mutant reaction was observed for the exon 19 deletion DEL6218 (mean difference in efficiencies approximately 12.2%). The amplification efficiency of the control compared to the mutation assay indicates that the $\Delta$Ct, and thus mutation call, is consistent across the working range of the assay. A summary of the data is shown below. PMA P120022: FDA Summary of Safety and Effectiveness Data {19} Assay Linearity as a Function of DNA Input | FFPE Cell Line | Assay Ct | Intercept | Intercept Standard Error | Slope Estimate | Slope Standard Error | Slope Two-Sided 95% CI | Amplification Efficiency(AE) | Difference in AE | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | DEL12367 | Control Ct | 31.868 | 0.058 | -1.083 | 0.015 | -1.114,-1.052 | 0.896 | . | | | Mutant Ct | 32.756 | 0.068 | -1.076 | 0.018 | -1.112,-1.04 | 0.904 | -0.008 | | DEL12387 | Control Ct | 31.828 | 0.047 | -1.077 | 0.012 | -1.102,-1.053 | 0.903 | . | | | Mutant Ct | 33.579 | 0.056 | -1.062 | 0.015 | -1.092,-1.033 | 0.920 | -0.017 | | DEL12419 | Control Ct | 31.686 | 0.073 | -1.097 | 0.019 | -1.135,-1.058 | 0.882 | . | | | Mutant Ct | 36.430 | 0.071 | -1.040 | 0.018 | -1.077,-1.002 | 0.948 | -0.066 | | DEL12422 | Control Ct | 31.935 | 0.050 | -1.088 | 0.013 | -1.114,-1.061 | 0.891 | . | | | Mutant Ct | 34.490 | 0.127 | -1.091 | 0.033 | -1.159,-1.023 | 0.888 | 0.003 | | DEL12728 | Control Ct | 31.731 | 0.036 | -1.052 | 0.009 | -1.071,-1.033 | 0.933 | . | | | Mutant Ct | 32.190 | 0.069 | -1.047 | 0.018 | -1.084,-1.01 | 0.939 | -0.006 | | DEL13551 | Control Ct | 31.441 | 0.040 | -1.029 | 0.010 | -1.05,-1.008 | 0.961 | . | | | Mutant Ct | 32.028 | 0.051 | -1.026 | 0.013 | -1.053,-0.999 | 0.965 | -0.003 | PMA P120022: FDA Summary of Safety and Effectiveness Data Page 20 {20} Assay Linearity as a Function of DNA Input (continued) | FFPE Cell Line | Assay Ct | Intercept | Intercept Standard Error | Slope Estimate | Slope Standard Error | Slope Two-Sided 95% CI | Amplification Efficiency(AE) | Difference in AE | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | DEL6210 | Control Ct | 31.850 | 0.080 | -1.093 | 0.021 | -1.135,-1.051 | 0.885 | . | | | Mutant Ct | 34.666 | 0.088 | -1.077 | 0.023 | -1.124,-1.031 | 0.903 | -0.018 | | DEL6218 | Control Ct | 31.470 | 0.031 | -1.030 | 0.008 | -1.046,-1.014 | 0.960 | . | | | Mutant Ct | 35.931 | 0.105 | -1.139 | 0.027 | -1.194,-1.083 | 0.838 | 0.122 | | DEL6220 | Control Ct | 32.283 | 0.054 | -1.114 | 0.014 | -1.143,-1.085 | 0.863 | . | | | Mutant Ct | 31.446 | 0.057 | -1.099 | 0.015 | -1.129,-1.068 | 0.879 | -0.016 | | DEL6223 | Control Ct | 31.917 | 0.098 | -1.153 | 0.026 | -1.206,-1.101 | 0.824 | . | | | Mutant Ct | 33.791 | 0.154 | -1.158 | 0.040 | -1.24,-1.076 | 0.820 | 0.004 | | DEL6225 | Control Ct | 31.436 | 0.034 | -1.045 | 0.009 | -1.063,-1.027 | 0.941 | . | | | Mutant Ct | 33.133 | 0.070 | -1.034 | 0.018 | -1.071,-0.997 | 0.955 | -0.014 | | G719A | Control Ct | 31.941 | 0.047 | -1.031 | 0.012 | -1.056,-1.007 | 0.958 | . | | | Mutant Ct | 34.571 | 0.077 | -1.113 | 0.020 | -1.154,-1.072 | 0.864 | 0.094 | | INS12376 | Control Ct | 32.541 | 0.043 | -1.043 | 0.010 | -1.064,-1.022 | 0.944 | . | | | Mutant Ct | 35.631 | 0.100 | -1.060 | 0.024 | -1.108,-1.013 | 0.923 | 0.021 | | INS12377 | Control Ct | 31.996 | 0.074 | -1.130 | 0.019 | -1.169,-1.09 | 0.847 | . | | | Mutant Ct | 34.090 | 0.073 | -1.098 | 0.019 | -1.138,-1.059 | 0.879 | -0.033 | | INS12378 | Control Ct | 31.791 | 0.061 | -1.100 | 0.016 | -1.133,-1.067 | 0.878 | . | | | Mutant Ct | 32.132 | 0.036 | -1.117 | 0.010 | -1.137,-1.097 | 0.860 | 0.018 | | L858R | Control Ct | 31.593 | 0.026 | -1.053 | 0.007 | -1.067,-1.039 | 0.931 | . | | | Mutant Ct | 31.667 | 0.023 | -1.017 | 0.006 | -1.029,-1.005 | 0.977 | -0.046 | | L861Q | Control Ct | 31.963 | 0.060 | -1.049 | 0.016 | -1.081,-1.017 | 0.936 | . | | | Mutant Ct | 33.267 | 0.064 | -1.028 | 0.017 | -1.062,-0.994 | 0.962 | -0.026 | PMA P120022: FDA Summary of Safety and Effectiveness Data {21} Assay Linearity as a Function of DNA Input (continued) | FFPE Cell Line | Assay Ct | Intercept | Intercept Standard Error | Slope Estimate | Slope Standard Error | Slope Two-Sided 95% CI | Amplification Efficiency(AE) | Difference in AE | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | S768I | Control Ct | 31.721 | 0.037 | -1.054 | 0.010 | -1.073,-1.034 | 0.931 | . | | | Mutant Ct | 34.609 | 0.088 | -1.081 | 0.023 | -1.129,-1.034 | 0.898 | 0.032 | | T790M | Control Ct | 31.455 | 0.050 | -1.075 | 0.013 | -1.102,-1.049 | 0.905 | . | | | Mutant Ct | 33.450 | 0.080 | -1.090 | 0.021 | -1.133,-1.048 | 0.889 | 0.017 | | WT | Control Ct | 31.840 | 0.049 | -1.089 | 0.013 | -1.115,-1.063 | 0.890 | . | b. Amplification Efficiency as a Function % Mutation The objective of this study was to evaluate the linearity of each mutant assay across the working range of the assay, when the total amount of DNA is held constant but the percentage of mutant DNA is varied. To maintain an equivalent Control Ct across the dilution series, EGFR mutation positive FFPE cell line DNA was diluted with the wild-type FFPE cell line DNA. Dilution series at both the high DNA input (Control Ct ~26) and low DNA input (Control Ct ~29-30) were tested. The dilution factor used was dependent on the ΔCt value generated by the undiluted FFPE cell line. Different cell lines generated different ΔCt values. Combined with different cut-offs for each mutation assay, each cell line required a slightly different dilution series to ensure that the final dilution point was beyond the assay cut-off. For each EGFR mutation test, pools of DNA sufficient for 6 replicates at each dilution level were prepared. The Ct and ΔCt data for each mutation at each dilution level were calculated. The Control Ct values corresponding to either ~26Ct or ~29-30 Ct were consistent over the dilution series of each mutation. A linear regression model was fitted to estimate the difference in mean ΔCt between the two DNA input levels. A plot of the ΔCt values was generated showing the data for both high and low DNA input levels on the same plot. The slope and 95% confidence intervals (95% CI) were reported. The study results are summarized in table below, which showed that dilution of mutations in a background of a constant amount of total DNA resulted in amplification efficiencies ("ΔCt Amplification Efficiency" column) that are mostly comparable (± 10%) to the mutation amplification efficiency determined in the above linearity study ("Mutation Amplification Efficiency" column). Amplification efficiencies differing close to or greater than 10% are noted for Del6220, Del6223, G719A, Ins12377, and L861Q mutations. PMA P120022: FDA Summary of Safety and Effectiveness Data {22} Assay Linearity as a Function of % Mutation | FFPE Cell Line | R-Square | Root Mean Square Error | Intercept Estimate | Slope Estimate | Difference In Estimated Mean ΔCt Between DNA Input Levels | Difference In Estimated Mean ΔCt Two-Sided 95% CI | ΔCt Amplification Efficiency | Control Amplification Efficiency | Mutation Amplification Efficiency | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | Del12367 | 0.975 | 0.373 | 0.064 | 1.013 | 0.018 | -0.192, 0.227 | 0.983 | 0.896 | 0.904 | | Del12387 | 0.970 | 0.376 | -0.096 | 1.019 | 0.177 | -0.040, 0.393 | 0.974 | 0.903 | 0.920 | | Del12419 | 0.945 | 0.277 | -0.454 | 1.100 | -0.162 | -0.323, -0.001 | 0.878 | 0.882 | 0.948 | | Del12422 | 0.876 | 0.626 | -0.009 | 1.066 | -0.478 | -0.844, -0.112 | 0.916 | 0.891 | 0.888 | | Del12728 | 0.945 | 0.650 | -0.596 | 0.976 | 1.134 | 0.756, 1.512 | 1.035 | 0.933 | 0.939 | | Del13551 | 0.979 | 0.373 | -0.394 | 1.027 | 0.473 | 0.265, 0.681 | 0.963 | 0.961 | 0.965 | | Del6210 | 0.888 | 0.571 | -0.222 | 1.057 | 0.007 | -0.325, 0.339 | 0.926 | 0.885 | 0.903 | | Del6218 | 0.914 | 0.418 | -0.786 | 1.117 | 0.325 | 0.082, 0.568 | 0.860 | 0.960 | 0.838 | | Del6220 | 0.989 | 0.309 | -0.351 | 1.015 | 0.341 | 0.160, 0.521 | 0.980 | 0.863 | 0.879 | | Del6223 | 0.966 | 0.368 | 0.855 | 1.008 | -0.954 | -1.195, -0.713 | 0.989 | 0.824 | 0.820 | | Del6225 | 0.975 | 0.336 | -0.287 | 0.992 | 0.119 | -0.074, 0.312 | 1.011 | 0.941 | 0.955 | | G719A | 0.947 | 0.520 | -0.083 | 1.033 | 0.310 | 0.021, 0.599 | 0.957 | 0.958 | 0.864 | | Ins12376 | 0.931 | 0.383 | 0.306 | 0.955 | -0.081 | -0.305, 0.143 | 1.066 | 0.944 | 0.923 | | Ins12377 | 0.957 | 0.353 | 0.078 | 1.016 | -0.174 | -0.381, 0.033 | 0.978 | 0.847 | 0.879 | | Ins12378 | 0.987 | 0.312 | -0.079 | 1.122 | -0.243 | -0.420, -0.065 | 0.855 | 0.878 | 0.860 | | L858R | 0.982 | 0.395 | -0.053 | 1.067 | 0.435 | 0.213, 0.656 | 0.915 | 0.931 | 0.977 | | L861Q | 0.943 | 0.634 | -0.365 | 1.122 | -0.104 | -0.460, 0.251 | 0.855 | 0.936 | 0.962 | | S768I | 0.950 | 0.458 | 0.459 | 1.048 | -0.619 | -0.917, -0.322 | 0.938 | 0.931 | 0.898 | | T790M | 0.949 | 0.487 | -0.050 | 1.146 | -0.308 | -0.579, -0.037 | 0.831 | 0.905 | 0.889 | ## 5. Analytical Specificity ### a. Primer and Probe Specificity The primers and probes have been designed to avoid any known EGFR polymorphisms. A specificity analysis was conducted using the Basic Local Alignment Search Tool (BLAST) to ensure that the primers used in the EGFR Kit would amplify only human EGFR sequences and not sequences from other species or non-EGFR human sequences (e.g., pseudogenes). No non-specific amplification is predicted from non-EGFR genes. In addition, alignments of pairs of oligonucleotides (primers, probes, and templates) used in the EGFR Kit were performed to ensure there is no unexpected binding that could lead to non-specific amplification. There was no significant homology between the various reagents. ### b. Cross Reactivity to Other EGFR Mutations Cross-reactivity of the therascreen® EGFR RGQ PCR Kit to other EGFR mutations was observed in the Phase 3 clinical trial specimens, FFPE cell lines and EGFR plasmids. The therascreen® EGFR RGQ PCR Kit gave "Mutation Detected" results for the following EGFR mutations in the PMA P120022: FDA Summary of Safety and Effectiveness Data {23} specific sample types indicated in the table below. Analytical performance of the therascreen® EGFR RGQ PCR Kit in detecting these mutations has not been evaluated for its intended use. ## Mutations Determined to Cross-React with the therascreen® EGFR RGQ PCR Kit | Exon 19 – Mutation Assay Del | | | | --- | --- | --- | | Mutation | COSMIC ID* | Sample Type | | 2237_2251del15 | 12678 | FFPE Clinical Trial Specimen* | | 2239_2247del9 | 6218 | FFPE Cell Line | | 2236_2253del18 | 12728 | FFPE Cell Line | | 2237_2254del18 | 12367 | FFPE Cell Line | | 2240_2251del12 | 6210 | FFPE Cell Line | | Exon 18 – Mutation Assay G719X | | | | G719S | 6252 | FFPE Clinical Trial Specimen# | | G719C | 6253 | Plasmid | | Exon 20 – Mutation Assay Ins | | | | Insertion | 12376 | FFPE Cell Line | Catalogue of Somatic Mutations in Cancer (COSMIC), http://www.sanger.ac.uk/genetics/CGP/cosmic ^The exon 19 deletion 12678 was observed in non-randomized study population of the 1200.32 clinical trial. #The exon 18 G719S mutation was observed in randomized study population of the 1200.32 clinical trial. ## c. Cross-Reactivity / Exclusivity The therascreen® EGFR RGQ PCR Kit is comprised of 8 separate reaction mixes: one single Control Reaction that detects a non-polymorphic region of the EGFR gene and seven mutation assays that detect EGFR mutations. There is no reaction that specifically measures the wild-type EGFR sequence at exons 18, 19, 20 or 21. The EGFR Kit “No Mutation Detected” result is determined from the absence of any positive mutation results. The objective of this study is to demonstrate the amount of non-specific amplification, or cross reactivity that occurs in each reaction with excess amounts of EGFR wild-type DNA to ensure no false positive results occur. Similarly, non-specific amplification of EGFR mutations does not result in erroneous mutation calls in the presence of excess amounts of mutant DNA. Since the DNA input for this assay is based on the Control Ct range (23.70 – 31.10), the highest concentration of DNA input is based on having a Control Ct value of approximately 25. FFPE cell lines were used for this evaluation. i. **Non-Specific Amplification/Cross Reactivity: Wild-type EGFR DNA** – To address the amount of non-specific amplification of wild-type EGFR DNA by reaction mixes designed to amplify PMA P120022: FDA Summary of Safety and Effectiveness Data Page 24 {24} specific mutations, 60 replicates of wild-type FFPE cell line DNA at approximately the highest concentration of amplifiable DNA input level were evaluated using the EGFR Kit. The Control Ct values were approximately 25 Ct. These results demonstrated that the lowest $\Delta$ Ct values exceeded the established cut-offs indicating that non-specific amplification was not observed. The lowest $\Delta$ Ct values observed for each reaction are shown below. Lowest Mean $\Delta$ Ct Observed for Wild-type Samples in Mutant Reactions | Assay | Cut-off | Lowest ΔCt Observed | | --- | --- | --- | | T790M | ≤ 7.4 | 12.77 | | Deletions | ≤ 8.0 | 13.25 | | L858R | ≤ 8.9 | * | | L861Q | ≤ 8.9 | 13.34 | | G719X | ≤ 8.9 | * | | S768I | ≤ 8.9 | * | | Insertions | ≤ 8.0 | * | *denotes no ΔCt value. ii. Non-Specific Amplification/Exclusivity: Mutation positive EGFR DNA – The exclusivity of the EGFR Kit is intended to discriminate between mutation negative and mutation positive status. Mutant samples with a high concentration of input DNA were tested against all reaction mixes by preparing DNA samples from FFPE cell lines (19 EGFR mutations detectable by the EGFR Kit) such that the Control Reaction Ct value corresponded to approximately 25. Sixty (60) replicates of each mutation sample were evaluated. The percentage of mutation in the sample was governed by the percentage of mutant in the cell line DNA. The minimum $\Delta$ Ct for each mutant is presented in the table below and demonstrates that there is no impact due to the cross reactivity between mutant reactions as the minimum $\Delta$ Ct values were all higher than the respective assay cut-off values for all non-matching reaction mixes and mutant DNA samples. PMA P120022: FDA Summary of Safety and Effectiveness Data {25} Cross-Reactivity between Mutation Assays Using FFPE Cell Line DNA | | Cut-Off | Mutant Reaction | | | | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | Deletions | G719X | Insertions | L858R | L861Q | S768I | T790M | | | Mutant DNA | DEL12367 | 8.0 | 0.31 | * | * | * | * | * | 12.07 | | | DEL12387 | 8.0 | 1.02 | * | * | * | * | * | 11.89 | | | DEL12419 | 8.0 | 4.19 | * | * | 13.57 | * | 13.96 | 11.67 | | | DEL12422 | 8.0 | 2.03 | * | * | * | * | * | 12.89 | | | DEL12728 | 8.0 | 0.31 | * | * | * | * | * | 13.03 | | | DEL13551 | 8.0 | -0.16 | * | * | * | * | * | 12.09 | | | DEL6210 | 8.0 | 2.16 | * | * | * | 13.87 | * | 10.93 | | | DEL6218 | 8.0 | 2.83 | * | * | 12.79 | 10.2 | * | 10.9 | | | DEL6220 | 8.0 | -1.01 | * | * | 11.46 | * | * | 11.37 | | | DEL6223 | 8.0 | 1.27 | * | * | 14.34 | * | * | 12.7 | | | DEL6225 | 8.0 | 0.24 | * | * | * | * | * | 12.52 | | | G719A | 8.9 | 11.91 | 1.62 | * | * | 13 | * | 12.47 | | | INS12376 | 8.0 | 13.21 | * | 2.09 | * | * | * | 11.82 | | | INS12377 | 8.0 | 12.27 | * | 1.5 | * | 13.11 | * | 11.93 | | | INS12378 | 8.0 | 13.71 | * | -0.61 | * | * | * | 11.09 | | | L858R | 8.9 | 13.68 | * | 14.2 | -0.83 | 14.39 | * | 12.46 | | | L861Q | 8.9 | 10.36 | * | * | * | 0.88 | * | 13.27 | | | S768I | 8.9 | 13.81 | * | 13.8 | 10.93 | 13.48 | 2.46 | 11.49 | | | T790M | 7.4 | 14.43 | * | 14.34 | * | 14.39 | * | 1.59 | Minimum ΔCts from matched reactions are shaded. Blank cells indicate no cross reactivity * indicates no ΔCt value ## 6. Interference – Effects of Necrotic Tissue To evaluate the potential interference of necrotic tissue content in NSCLC FFPE specimens on the performance of the EGFR Kit, FFPE clinical specimens from the 1200.32 clinical trial with both EGFR Kit test results and Sager sequencing results were analyzed. A total of 66 EGFR mutant specimens (i.e., 30 exon 19 deletion, 32 L858R, 1 G719X, 1 exon 20 insertion, 1 L861Q, and 1 S768I specimens) and 81 wild-type specimens were evaluated. Percent necrosis, as identified by a pathologist, varied from 0-50% for both mutant and wild-type FFPE specimens. For both mutant and wild-type FFPE specimens, all except 2 samples have EGFR Kit results that matched the expected Sanger sequencing results. The two mismatched results were from one wild-type and one mutant samples with less than 10% necrotic content; thus, it is unlikely that necrosis was the reason for the discordant result. The overall percent correct calls were 98.5% (65/66) and 98.8% (80/81) for mutant and wild-type FFPE specimen, respectively. The results support the use of the EGFR Kit with NSCLC FFPE specimen with necrotic tissue content up to 50%. ## 7. Interference – Exogenous Substance To evaluate the impact of interfering substance on performance of the EGFR Kit, potentially interfering substances present in the DNA extraction process, were tested at 10x concentration in mutant and wild-type samples with a target Ct value of approximately 27.5. The impact of each substance on the ΔCt PMA P120022: FDA Summary of Safety and Effectiveness Data {26} values and mutation status of the samples was assessed. The substances tested were (1) paraffin wax, (2) xylene, (3) ethanol, and (4) Proteinase K. The difference between the $\Delta C_t$ of samples with interferent was compared to samples without interferent according to statistical methods outlined in CLSI guidance document EP7-A2. For mutant samples, of the 672 replicates tested (7 mutation assays x 8 substances x 3 levels of interfering substance x 4 replicates), there were no invalid or indeterminate results and no false mutation negative results. For wild-type samples, 0 of the 96 replicates tested were invalid or indeterminate and there were no false mutation detected results. The results demonstrated that these substances did not interfere with the therascreen® EGFR RGQ PCR Kit call results. ## 8. Repeatability and Reproducibility The repeatability and reproducibility of the EGFR Kit was investigated by testing DNA extracted from NSCLC FFPE clinical specimens or FFPE cell lines, representing all 7 mutation assays in the EGFR Kit. NSCLC wild-type FFPE clinical specimens were also included in the study. Reproducibility was conducted across three sites (i.e., United Kingdom, Germany, and USA). At each site, samples were tested in duplicate (for within-run repeatability assessment), on 2 different RGQ instruments, using 2 operators and 2 EGFR Kit lots (3 lots across 3 sites) over a total of 16 days. Reproducibility for each individual mutation was conducted over non-consecutive days at each site. This resulted in a total of 32 data points for each test sample at each site, and a total of 96 data points for each EGFR mutation per mutation level. One lot of QIAGEN QIAamp® DSP DNA FFPE Tissue Kit was used to extract DNA from FFPE samples. Samples were prepared to have low DNA input levels where a control Ct value of approximately 30.10 was targeted. A sample pool for each mutation was created from a single specimen, and the wild-type pool comprised 5 different specimens combined. For each sample, the proportions of correct mutation calls and the lower one-sided exact 95% confidence intervals (95% CI) were reported in table below. There were no "Mutation Detected" results in 84 valid tests of wild-type sample, producing 100% correct calls. The percentage of correct calls ranged from 96% -100% for mutant samples tested at 1-3x LoD across sites. PMA P120022: FDA Summary of Safety and Effectiveness Data Page 27 {27} Assay Reproducibility – Proportion of Correct Calls for EGFR Mutation Tested | Exon | Mutation | COSMIC ID | % Mutation Tested | % Mutation Tested Relative to Final LoD Claim | Number of Valid Results (N) | Correct Calls (N) | % Correct Calls | % Correct Call Lower One-Sided 95% CI | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | Wild-type | NA | | | 84 | 84 | 100.00 | 96.50 | | 19 | Deletions | 6220* | 5.69% | 2-3x LoD | 96 | 96 | 100.00 | 96.93 | | | | 6223 | 15.99% | 2-3x LoD | 95 | 95 | 100.00 | 96.90 | | | | 6225 | 7.06% | 1-2x LoD | 95 | 91 | 95.79 | 90.62 | | | | 6254 | 10.02% | LoD | 92 | 92 | 100.00 | 96.80 | | | | 6255 | 0.81% | LoD | 96 | 94 | 97.92 | 93.59 | | | | 12369 | 9.29% | 1-2x LoD | 95 | 95 | 100.00 | 96.90 | | | | 12370 | 8.06% | LoD | 63^ | 62 | 98.41 | 92.69 | | | | 12382 | 1.45% | LoD | 95 | 92 | 96.84 | 92.04 | | | | 12383 | 8.43% | 1-2x LoD | 93 | 93 | 100.00 | 96.83 | | | | 12384 | 7.54% | LoD | 92 | 92 | 100.00 | 96.80 | | | | 12387* | 9.53% | 1-2x LoD | 95 | 95 | 100.00 | 96.90 | | | | 12419* | 28.75% | 1-2x LoD | 83 | 83 | 100.00 | 96.46 | | | | 12422 | 7.85% | 2-3x LoD | 94 | 94 | 100.00 | 96.86 | | | | 13551* | 11.12% | 2-3x LoD | 95 | 95 | 100.00 | 96.90 | | 21 | L858R | 6224 | 5.77% | LoD | 92 | 92 | 100.00 | 96.80 | | 20 | T790M# | 6240 | 34.02% | 1-2xLoD | 94 | 94 | 100.00 | 96.86 | | 21 | L861Q# | 6213 | 9.24% | LoD | 84 | 83 | 98.81 | 94.48 | | 18 | G719A# | 6239 | 32.50% | LoD | 78 | 77 | 98.72 | 94.06 | | 20 | S768I# | 6241* | 11.57% | 1-2x LoD | 82 | 82 | 100.00 | 96.41 | | 20 | Insertions# | 12377* | 10.45% | 2-3x LoD | 93 | 93 | 100.00 | 96.83 | | | | 12378* | 19.96% | LoD | 92 | 92 | 100.00 | 96.80 | ^Control Ct for deletion 12370 at LoD dropped out of the working range and no data could be generated at one site (n=32). The missing data from this site for this mutation was not retested due to lack of sample availability. *Reproducibility for these mutations was conducted using FFPE cell lines. #Safety and efficacy of GILOTRIF™ (afatinib) has not been established for patients with these EGFR mutations. Refer to GILOTRIF™ (afatinib) drug labeling for more details. A variance component analysis was used to estimate the standard deviation and 95% confidence intervals for within-run, between-run, between-day, between-lot and between-site variability. These estimates were reported along with %CV and the number of observations for the mean of ΔCt, Control Ct, and Mutant Ct values. Results by variance components and total variance are presented below for ΔCt. The column “N” included the number of data points that generated a ΔCt value. Across all variance components, the total coefficient of variation (CV) was ≤14.11% in all EGFR mutations tested. Across all mutant panel members, the %CV was in general &lt;6% for between-lots, between lots, between days, and between runs. The %CV for within-run (repeatability) ranged from 5.99% to 13.49%. PMA P120022: FDA Summary of Safety and Effectiveness Data {28} Overall Mean, Standard Deviation (SD), and Coefficient of Variation (% CV) for $\Delta \mathrm{Ct}$ - Between Lot, Between Site, and Total Variance | Exon | Mutation | COSMIC ID | LoD Level Tested | N | Mean | Between Lot (SD) | Between Lot (%CV) | Between Site (SD) | Between Site (%CV) | Total (SD) | Total (%CV) | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | 19 | Deletions | 6220 | 2-3x LoD | 96 | 4.57 | 0.00 | 0.00 | 0.07 | 1.64 | 0.62 | 13.56 | | | | 6223 | 2-3x LoD | 95 | 3.59 | 0.00 | 0.00 | 0.20 | 5.46 | 0.39 | 10.95 | | | | 6225 | 1-2x LoD | 94 | 6.06 | 0.35 | 5.79 | 0.00 | 0.00 | 0.86 | 14.11 | | | | 6254 | LoD | 92 | 4.91 | 0.00 | 0.00 | 0.05 | 1.03 | 0.47 | 9.48 | | | | 6255 | LoD | 96 | 6.11 | 0.00 | 0.00 | 0.00 | 0.00 | 0.68 | 11.20 | | | | 12369 | 1-2x LoD | 95 | 4.83 | 0.10 | 2.02 | 0.17 | 3.60 | 0.51 | 10.55 | | | | 12370 | LoD | 62 | 6.15 | 0.00 | 0.00 | 0.00 | 0.00 | 0.64 | 10.48 | | | | 12382 | LoD | 94 | 6.25 | 0.00 | 0.00 | 0.04 | 0.66 | 0.79 | 12.61 | | | | 12383 | 1-2x LoD | 93 | 4.71 | 0.00 | 0.00 | 0.09 | 1.85 | 0.40 | 8.58 | | | | 12384 | LoD | 92 | 4.31 | 0.00 | 0.00 | 0.00 | 0.00 | 0.46 | 10.62 | | | | 12387 | 1-2x LoD | 95 | 4.81 | 0.00 | 0.00 | 0.02 | 0.33 | 0.58 | 12.13 | | | | 12419 | 1-2x LoD | 83 | 5.64 | 0.00 | 0.00 | 0.17 | 3.07 | 0.52 | 9.16 | | | | 12422 | 2-3x LoD | 94 | 4.07 | 0.00 | 0.00 | 0.10 | 2.49 | 0.33 | 8.00 | | | | 13551 | 2-3x LoD | 95 | 4.46 | 0.03 | 0.59 | 0.12 | 2.74 | 0.60 | 13.39 | | 21 | L858R | 6224 | LoD | 92 | 4.77 | 0.00 | 0.00 | 0.19 | 4.00 | 0.55 | 11.59 | | 20 | T790M# | 6240 | 1-2xLoD | 94 | 4.74 | 0.00 | 0.00 | 0.25 | 5.22 | 0.49 | 10.34 | | 21 | L861Q# | 6213 | LoD | 84 | 6.04 | 0.00 | 0.00 | 0.11 | 1.79 | 0.76 | 12.56 | | 18 | G719A# | 6239 | LoD | 78 | 5.73 | 0.00 | 0.00 | 0.11 | 1.86 | 0.63 | 11.06 | | 20 | S768I# | 6241 | 1-2x LoD | 82 | 5.30 | 0.00 | 0.00 | 0.12 | 2.28 | 0.55 | 10.31 | | 20 | Insertions# | 12377 | 2-3x LoD | 93 | 4.14 | 0.00 | 0.00 | 0.00 | 0.00 | 0.42 | 10.05 | | | | 12378 | LoD | 92 | 4.9 | 0.00 | 0.00 | 0.00 | 0.00 | 0.57 | 11.56 | #Safety and efficacy of GILOTRIFTM (afatinib) has not been established for patients with these EGFR mutations. Refer to GILOTRIFTM (afatinib) drug labeling for more details. PMA P120022: FDA Summary of Safety and Effectiveness Data {29} Overall Mean, Standard Deviation (SD), and Coefficient of Variation (% CV) for ΔCt – Within Run, Between Run, and Between Day | Exon | Mutation | COSMIC ID | LoD Level Tested | N | Mean | Within Run (SD) | Within Run (%CV) | Between Run (SD) | Between Run (%CV) | Between Day (SD) | Between Day (%CV) | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | 19 | Deletions | 6220 | 2-3x LoD | 96 | 4.57 | 0.62 | 13.49 | 0.00 | 0.00 | 0.00 | 0.00 | | | | 6223 | 2-3x LoD | 95 | 3.59 | 0.35 | 9.73 | 0.00 | 0.00 | 0.08 | 2.37 | | | | 6225 | 1-2x LoD | 94 | 6.06 | 0.78 | 12.91 | 0.00 | 0.00 | 0.21 | 3.47 | | | | 6254 | LoD | 92 | 4.91 | 0.45 | 9.11 | 0.1 | 1.95 | 0.08 | 1.6 | | | | 6255 | LoD | 96 | 6.11 | 0.68 | 11.20 | 0.00 | 0.00 | 0.00 | 0.00 | | | | 12369 | 1-2x LoD | 95 | 4.83 | 0.48 | 9.90 | 0.00 | 0.00 | 0.00 | 0.00 | | | | 12370 | LoD | 62 | 6.15 | 0.5 | 8.10 | 0.41 | 6.71 | 0.00 | 0.00 | | | | 12382 | LoD | 94 | 6.25 | 0.79 | 12.60 | 0.00 | 0.00 | 0.00 | 0.00 | | | | 12383 | 1-2x LoD | 93 | 4.71 | 0.40 | 8.39 | 0.04 | 0.95 | 0.00 | 0.00 | | | | 12384 | LoD | 92 | 4.31 | 0.36 | 8.43 | 0.23 | 5.40 | 0.17 | 3.97 | | | | 12387 | 1-2x LoD | 95 | 4.81 | 0.57 | 11.91 | 0.11 | 2.3 | 0.00 | 0.00 | | | | 12419 | 1-2x LoD | 83 | 5.64 | 0.4 | 7.05 | 0.32 | 5.63 | 0.00 | 0.00 | | | | 12422 | 2-3x LoD | 94 | 4.07 | 0.28 | 6.98 | 0.14 | 3.37 | 0.00 | 0.00 | | | | 13551 | 2-3x LoD | 95 | 4.46 | 0.55 | 12.42 | 0.20 | 4.52 | 0.00 | 0.00 | | 21 | L858R | 6224 | LoD | 92 | 4.77 | 0.41 | 8.65 | 0.34 | 7.04 | 0.02 | 0.47 | | 20 | T790M# | 6240 | 1-2xLoD | 94 | 4.74 | 0.28 | 5.99 | 0.29 | 6.08 | 0.20 | 4.21 | | 21 | L861Q# | 6213 | LoD | 84 | 6.04 | 0.66 | 10.95 | 0.36 | 5.89 | 0.08 | 1.32 | | 18 | G719A# | 6239 | LoD | 78 | 5.73 | 0.41 | 7.17 | 0.48 | 8.33 | 0.00 | 0.00 | | 20 | S768I# | 6241 | 1-2x LoD | 82 | 5.30 | 0.48 | 9.02 | 0.20 | 3.72 | 0.15 | 2.82 | | 20 | Insertions# | 12377 | 2-3x LoD | 93 | 4.14 | 0.42 | 10.05 | 0.00 | 0.00 | 0.00 | 0.00 | | | | 12378 | LoD | 92 | 4.90 | 0.48 | 9.83 | 0.30 | 6.10 | 0.05 | 1.08 | #Safety and efficacy of GILOTRIFTM (afatinib) has not been established for patients with these EGFR mutations. Refer to GILOTRIFTM (afatinib) drug labeling for more details. ## 9. Lot-to-Lot Reproducibility This study addressed the potential for lot-to-lot variability to impact the mutation detection by the EGFR Kit. A maximum of three FFPE clinical specimens or FFPE cell lines representing each of the seven mutation assays of the EGFR Kit were used in this study, along with 21 wild-type samples. The EGFR Kit test system utilizes two separate kits: (1) The QIAamp® DSP DNA FFPE Tissue Kit for isolation of DNA from NSCLC FFPE specimens, and (2) the therascreen® EGFR RGQ PCR Kit for the amplification and detection of the isolated DNA for its EGFR mutation status. FFPE mutant and wild-type specimens were tested with 3 lots of the QIAamp® DSP DNA FFPE Tissue Kit and 3 lots of the therascreen® EGFR RGQ PCR Kit. Sections from each FFPE clinical specimen or FFPE cell line were extracted in triplicate, and extracted DNA was tested with the Control assay and the corresponding mutation assay. The samples were tested in duplicate giving 18 results for each sample per EGFR Kit. This produced a maximum of 54 data points in total per mutation assay. The exceptions to this were S768I, which only had one representative sample and gave 18 data points, and the Insertions, which had two representative samples and gave a total of 36 data points. Only valid results were included in the analysis. The test results were analyzed for PMA P120022: FDA Summary of Safety and Effectiveness Data {30} each mutation assay and also by the kit lots. The overall percentage of correct calls across lots for EGFR mutation assay was 97.8% (317/324) and that for wild-type samples was 100% (379/379). The Control Ct, Mutant Ct and ΔCt values for each specimen were summarized across lots and no trend in Ct or ΔCt values were observed. ## 10. Specimen Handling – Reproducibility The objective of this study was to assess sample handling variability as part of the EGFR Kit test system process. The reproducibility of the QIAamp® DSP DNA FFPE Tissue Kit was examined using sections taken from three FFPE specimen blocks, one containing an exon 19 deletion mutation (2235-2249 del15), one containing the exon 21 L858R mutation (2573T&gt;G), and one that is wild-type. From the mutant FFPE clinical specimens, 48 sequential FFPE slide sections were randomized and divided into four batches where one batch of slide sections was used per site and one batch was kept as a contingency batch. For the wild-type FFPE clinical specimen, 36 sequential FFPE slide sections were randomized and divided into three batches where one batch of slide sections was used per site. For each specimen, extractions were carried out in duplicate at each test site and tested on three non-consecutive days over a period of six days across three sites, yielding a total of 18 data points per specimen. At each site, two operators conducted the testing using one lot of the QIAamp® DSP DNA FFPE Tissue Kit (one lot per site, 3 lots total) in combination with the same lot of the EGFR Kit reagents across sites. One RGQ instrument was used to conduct the testing at Site 1 and two RGQ instruments were used at Sites 2 and 3. The acceptance criterion was that each clinical specimen should give the correct mutation call at least 17 out of 18 times. All mutant and wild-type specimen results were valid and yielded the expected call result (correct call =100%, 18/18 for each specimen), supporting the reproducibility and repeatability for the therascreen® EGFR RGQ PCR Kit at the pre-analytical step of DNA isolation ## 11. Guard banding The objective of the guard banding studies was to establish the robustness of the EGFR Kit. The following studies were conducted to assess proteinase K digestion step for the QIAamp® DSP DNA FFPE Tissue Kit and the RGQ PCR cycling parameters for the therascreen® EGFR RGQ PCR Kit. ### a. Proteinase K This study was designed to determine the effect on the mutation Ct and ΔCt when varying the 56°C and 90°C incubation times and temperatures during the extraction of FFPE samples. FFPE sample extractions were performed using the QIAamp® DSP DNA FFPE Tissue Kit according to the Instruction For Use (IFU) provided with the kit. According to the IFU, the extraction process has two incubation steps: 1) Incubate at 56°C ± 3°C for ≥1 hour, and 2) Incubate at 90°C ± 5°C for 1 hour ±5 minutes. PMA P120022: FDA Summary of Safety and Effectiveness Data Page 31 {31} A total of 8 samples representing the Control and the seven mutation assays were tested in singlet at 26 experimental conditions, including variation of length of time and the temperature of the two incubation steps using a central composite design. FFPE clinical specimens were used for the exon 19 deletion assay and the exon 21 L858R assay, while FFPE cell lines were used for the other 5 mutation assays. A total of 187 mutation detected calls of repeated sample testing at varying cycling tolerances were reported, with 5 of the 187 as false positive calls. The conditions which resulted in these false positive calls were all ±5°C away from the standard conditions for both incubation steps. However, the data generated shows that varying the incubation temperatures by ±3°C can be tolerated with a maximum shift of 0.40 ΔCt observed across all the variables, from the standard condition. The results indicate that varying the incubation temperatures by ±5°C cannot be tolerated; however, varying the time by ±10 minutes can be tolerated. Thus, the EGFR Kit IFU recommendations are incubation at 56°C ± 3°C for 1 hour ± 5 minutes, followed by incubation at 90°C ± 3°C for 1 hour ± 5 minutes. b. Thermal Cycling Profile The study objective was to determine effect of mutation calling by varying the thermal cycling profile of the RGQ instrument, using FFPE clinical samples. Three FFPE clinical specimens, one containing an exon 19 deletion mutation (Del6223 or 2235-2249 del15), one containing an L858R mutation (2573T&gt;G), and one that is wild-type, were processed to isolate DNA on the same day using the QIAamp® DSP DNA FFPE Tissue Kit. The standard cycling condition for the therascreen® EGFR RGQ PCR Kit is as follows: - Activation: 95°C for 15 minutes - Denaturation: 95°C for 30 seconds - Annealing: 60°C for 1 minute The thermal cycling profile was guard banded by varying ±2°C, in 1°C steps, during activation, denaturation and annealing steps. A full 5 by 5 factorial design was used to test a total of 25 PCR conditions using 5 RGQ instruments over 4 days. Each extraction sample was tested in triplicate using the control assay to derive a mean Ct, and further normalized to a target Control Ct ~30. Three replicates of each specimen pool were then tested for each condition using a single reagent lot of the EGFR Kit. For each condition with temperatures within ±1°C from the standard condition, the mutation positive ΔCt should be ±1.5 Ct from the standard condition. All observed mutation calls matched the expected results when the cycling temperature shifts by ±1°C. Across all variables, a maximum ΔCt shift of 0.74 was observed for the exon 19 deletion assay and a maximum ΔCt shift of 0.46 was observed for the exon 21 L858R mutation assay. All PMA P120022: FDA Summary of Safety and Effectiveness Data Page 32 {32} wild-type samples showed amplification in the Control assay, and no positive mutation calls were made. The results showed that the therascreen® EGFR RGQ PCR Kit is able to tolerate variations of ± 1°C of the thermal cycling profile activation, denaturation and annealing temperatures. c. PCR Amplification Mix The study objective was to determine effect of mutation calling and tolerance of the EGFR Kit to volumetric variations of each PCR components in the PCR amplification mixture. Three FFPE clinical specimens, one containing an exon 19 deletion mutation (Del 6223, also known as 2235-2249 del15), one containing an L858R mutation (2573T&gt;G), and one that is wild-type. The same DNA samples prepared for the thermal cycling guard banding study above were utilized, with the target Control Ct ~30. The volume of each PCR component (PCR mix, Taq polymerase, Master mix and DNA sample) was varied incrementally up to ±20% while keeping all other components at the standard volumes as stated in the IFU. A total of 22 test conditions were…