LeukoStrat CDx FLT3 Mutation Assay

P160040 · Invivoscribe Technologies, Inc. · OWD · Apr 28, 2017 · Pathology

Device Facts

• Record ID: P160040
• Device Name: LeukoStrat CDx FLT3 Mutation Assay
• Applicant: Invivoscribe Technologies, Inc.
• Product Code: OWD · Pathology
• Decision Date: Apr 28, 2017
• Decision: APPR
• Device Class: Class 3
• Attributes: Expedited Review

Intended Use

The LeukoStrat® CDx FLT3 Mutation Assay is a PCR-based, in vitro diagnostic test designed to detect internal tandem duplication (ITD) mutations and the tyrosine kinase domain (TKD) mutations D835 and I836, in the FLT3 gene in genomic DNA extracted from mononuclear cells obtained from PB or BM aspirates of patients diagnosed with acute myelogenous leukemia (AML). The LeukoStrat® CDx FLT3 Mutation Assay is used as an aid in the selection of patients with AML for whom midostaurin (RYDAPT®) treatment is being considered. The LeukoStrat® CDx FLT3 Mutation Assay is to be performed only at Laboratory for Personalized Molecular Medicine (LabPMM) LLC, a single laboratory site located at 6330 Nancy Ridge Dr., San Diego, CA 92121.

Device Story

Assay detects FLT3 ITD and TKD (D835, I836) mutations in genomic DNA from mononuclear cells isolated from peripheral blood or bone marrow aspirates. Process: DNA extraction (QIAGEN DSP kits); PCR amplification using fluorescently labeled primers; EcoRV restriction digest for TKD; capillary electrophoresis on ABI 3500xL Genetic Analyzers. Software calculates mutant:wild-type signal ratio (SR). Results reported as Positive (SR ≥0.05) or Negative. Performed exclusively at LabPMM, San Diego. Output informs physician selection of AML patients for midostaurin (RYDAPT®) therapy. Benefits include targeted treatment for poor-prognosis FLT3-mutated AML; risks include potential false negatives (missing treatment benefit) or false positives (unnecessary drug exposure).

Clinical Evidence

Retrospective bridging study using 1,058 samples from the Phase 3 RATIFY (A2301) clinical trial. Compared LeukoStrat® CDx results to clinical trial assay (CTA). Overall percent agreement (OPA) was 97.3% (with invalids) to 98.4% (without invalids). PPA 97.2%, NPA 97.3%. Clinical efficacy (OS) in CDx-positive population (HR 0.67) was consistent with the overall trial results (HR 0.77). Bench testing established LoD (3-35% depending on mutation size), analytical specificity, and reproducibility.

Technological Characteristics

PCR-based in vitro diagnostic; uses fluorescently labeled primers and EcoRV restriction enzyme digestion. Analyzed via capillary electrophoresis on ABI 3500xL Genetic Analyzers. Software: FLT3 Mutation Analysis GeneMapper Software, Version 1.5. Requires mononuclear cell isolation via Ficoll gradient. DNA extraction via QIAGEN DSP blood mini kits. Standardized to 10 ng/μL DNA input.

Indications for Use

Indicated for patients diagnosed with acute myelogenous leukemia (AML) to detect FLT3 ITD and TKD (D835, I836) mutations to aid in selecting candidates for midostaurin (RYDAPT®) treatment.

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.

Submission Summary (Full Text)

{0}

SUMMARY OF SAFETY AND EFFECTIVENESS DATA (SSED)

I. GENERAL INFORMATION

Device Generic Name: Somatic Mutation PCR test

Device Trade Name: LeukoStrat® CDx FLT3 Mutation Assay

Device Procode: OWD

Applicant's Name and Address: Invivoscribe Technologies, Inc.
6330 Nancy Ridge Drive, Suite 106,
San Diego, CA 92121

Date(s) of Panel Recommendation: None

Premarket Approval Application (PMA) Number: P160040

Date of FDA Notice of Approval: April 28, 2017

Priority Review: Granted priority review status on October 21, 2016, because the device addresses an unmet medical need as demonstrated by significant meaningful advantage.

II. INDICATIONS FOR USE

The LeukoStrat® CDx FLT3 Mutation Assay is a PCR-based, in vitro diagnostic test designed to detect internal tandem duplication (ITD) mutations and the tyrosine kinase domain (TKD) mutations D835 and I836, in the FLT3 gene in genomic DNA extracted from mononuclear cells obtained from PB or BM aspirates of patients diagnosed with acute myelogenous leukemia (AML).

The LeukoStrat® CDx FLT3 Mutation Assay is used as an aid in the selection of patients with AML for whom midostaurin (RYDAPT®) treatment is being considered.

The LeukoStrat® CDx FLT3 Mutation Assay is to be performed only at Laboratory for Personalized Molecular Medicine (LabPMM) LLC, a single laboratory site located at 6330 Nancy Ridge Dr., San Diego, CA 92121.

III. CONTRAINDICATIONS

There are no known contraindications.

IV. WARNINGS AND PRECAUTIONS

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 1

{1}

The warnings and precautions can be found in the LeukoStrat® CDx FLT3 Mutation Assay labeling.

V. DEVICE DESCRIPTION

The LeukoStrat® CDx FLT3 Mutation Assay (CDx) is an in vitro diagnostic device performed in a single laboratory, LabPMM, in San Diego, CA. The assay includes reagents, software, instruments and procedures for testing DNA extracted from mononuclear cells isolated from patient specimens to determine if FLT3 ITDs or TKD mutations are present. The Kit contains sufficient reagents for 150 reactions.

Device Instrumentation and Software

The FLT3 Mutation Assay to be used on serialized Applied Biosystems 3500xL Genetic Analyzers which is configured with FLT3 Mutation Analysis GeneMapper Software, Version 1.5. The serial numbers for the 4 instruments are S/N 21106-231, S/N 23321-020, S/N 23321-030, and S/N 25351-081.

Specimen Preparation

Peripheral blood (PB) specimens (≥1mL) or bone marrow aspirate (BM) specimens (≥0.25mL) are collected in tubes with sodium heparin anticoagulant. (Frozen blood is unacceptable as a sample for testing.) Blood and BM specimens are received on cold packs, the mononuclear cells are separated from the specimens via passage over a Ficoll column, and then counted to ensure that the extraction columns are not overloaded. The DNA is extracted from the mononuclear cells using QIAGEN DSP blood mini kits. After DNA isolation, the DNA is quantified. DNA samples ≥10.5 ng/μL must be diluted to 10 ng/μL in AE buffer.

PCR Amplification and Detection

Fms-like tyrosine kinase 3 (FLT3) is a gene that encodes for a tyrosine kinase that activates pathways in hematopoietic cells. FLT3 ITDs are caused by duplication and insertion of a portion of the FLT3 gene that includes the region in and around the juxtamembrane (JM) region of the FLT3 gene. These mutations vary in both the location and the length of the inserted duplicated DNA sequence. FLT3 TKD mutations detected by the LeukoStrat® CDx FLT3 Mutation Assay are caused by nucleic acid substitutions and/or deletions that result in a change in the EcoRV restriction enzyme cut site. This includes TKD mutations at D835 and I836. The FLT3-ITD and TKD mutations result in constitutive autophosphorylation and activation of FLT3.

FLT3-ITD: The LeukoStrat® CDx FLT3 Mutation Assay uses primers that are in the JM region. The forward and reverse PCR primers are fluorescently labeled with different fluorophores that serve to confirm the presence of sample signal. Wild type FLT3 alleles will amplify and produce a product measured at 327±1 bp as measured by the LeukoStrat® CDx FLT3 Mutation Assay, while alleles that contain ITD mutations will produce a product that exceeds the wild-type result by at least 3 bp. The LeukoStrat® CDx FLT3 Mutation Assay is validated to detect ITD containing amplicons approximately up to 650 bp (i.e., ITDs up to approximately 279 bp).

PMA P160040: FDA Summary of Safety and Effectiveness Data

{2}

FLT3-TKD: Wild-type alleles of the FLT3 gene include an EcoRV restriction digest site. When a substitution and/or deletion change occurs, the restriction digest recognition site disappears, and the EcoRV endonuclease is unable to identify and digest the DNA at this site. The LeukoStrat® CDx FLT3 Mutation Assay uses primers that lie on either side of the TKD region. The FLT3 target region is amplified using PCR and then an EcoRV restriction digest is performed. One of the PCR primers is labeled with a fluorophore and the other contains an EcoRV restriction site, so both wild type and mutant alleles are digested. The digestion pattern identifies loss of the normal gene sequence and ensures that digestion occurred. Wild type alleles of the FLT3 gene yield digestion products of $79 \pm 1$ bp whereas mutant alleles yield products of $125 \pm 1$ (deletion) or $127 \pm 1$ (substitution) bp from the original undigested amplicon product of $145 \pm 1$ (deletion) or $147 \pm 1$ (substitution) bp as measured by the LeukoStrat® CDx FLT3 Mutation Assay.

# Interpretation of Results

Following amplification, the resulting amplicons are analyzed via capillary electrophoresis. The assay measures the ratio of signals from mutation against a background of signal from wild type. The mutant:wild-type signal ratio (SR) is calculated by the FLT3 Mutation Analysis Software and automatically evaluated against the cut-off (medical decision point) of 0.05. The SR is the peak area of the mutant signal, if present, divided by the peak area of the wild-type signal. If the mutant:wild type SR for either ITD or TKD in a valid sample is at or above the cut-off of 0.05, the result will be interpreted and reported as Positive (mutation detected). Below the cut-off, the result will be interpreted and reported as Negative. The mutation status of a sample is defined by the rules below in Table 1.

Table 1. Determination of Sample FLT3 Mutation Status

| Scenario | ITD Result | TKD Result | Sample Mutation Status |
| --- | --- | --- | --- |
| 1 | Positive (SR ≥0.05) | Positive (SR ≥0.05) | Positive |
| 2 | Negative (SR <0.05) | Negative (SR <0.05) | Negative |
| 3 | Invalid | Invalid | Invalid |
| 4 | Positive (SR ≥0.05) | Negative (SR <0.05) | Positive |
| 5 | Negative (SR <0.05) | Positive (SR ≥0.05) | Positive |
| 6 | Positive (SR ≥0.05) | Invalid | Positive |
| 7 | Negative (SR <0.05) | Invalid | Invalid |
| 8 | Invalid | Positive (SR ≥0.05) | Positive |
| 9 | Invalid | Negative (SR <0.05) | Invalid |

# Controls and Run Validity

# Positive Controls

The FLT3-ITD positive control is designed to yield a signal ratio of between 0.05 and 0.11. The control is produced by mixing gDNA from a cell line (MV4-11) containing a FLT3-ITD mutation with DNA from a cell line (Jurkat) without a FLT3 ITD mutation.

The FLT3-TKD positive control is designed to yield a signal ratio of between 0.05 and 0.11. The control is produced by mixing gDNA from a cell line containing a FLT3-TKD

PMA P160040: FDA Summary of Safety and Effectiveness Data

{3}

mutation (TVI-TKD cell line) with DNA from a cell line without a FLT3 ITD mutation (Jurkat cell line).

The positive controls are meant to ensure the assay is properly evaluating positive ITD and TKD mutated samples.

## No Template Control

Sterile water serves as a negative control to assess if any samples were free from contamination.

## Extraction Control

Extraction control is gDNA isolated from a cell line without either a FLT3-ITD or a FLT3-TKD mutation (Jurkat cell line). This control ensures the extraction step functioned properly and works as a negative control. If an Extraction Control does not meet its validity criteria for signal strength, signal ratio, and/or peak size, all samples associated with that Extraction Control will be labeled Fail.

The FLT3 Mutation Analysis Software automatically evaluates the results against the criteria shown in Table 2 for peak size when using the ITD or TKD PCR reaction.

Table 2. Run Validity Criteria

| Run Validity Control | ITD Requirements | TKD Requirements |
| --- | --- | --- |
| Positive Control | Valid (meets signal strength and peak size specifications)
Mutant:wild type signal ratio ≥0.05 and ≤0.11 | Valid (meets signal strength and peak size specifications)
Mutant:wild type signal ratio ≥0.05 and ≤0.11 |
| No Template Control | No peaks in the region of interest | No peaks in the region of interest |

## VI. ALTERNATIVE PRACTICES AND PROCEDURES

There are no other FDA approved or cleared alternatives for testing of FLT3 mutations in human specimens.

## VII. MARKETING HISTORY

The LeukoStrat® CDx FLT3 Mutation Assay 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 FLT3 mutation results and subsequently, improper patient management decisions in AML treatment. Patients with a false negative assay result may not be considered for treatment with RYDAPT® (midostaurin). Patients with a false

PMA P160040: FDA Summary of Safety and Effectiveness Data

{4}

positive assay result may receive treatment with RYDAPT® (midostaurin) for which there is no expectation of benefit. There is also a risk of delayed results, which may lead to delay in treatment.

For the specific adverse events related to midostaurin (RYDAPT®) (midostaurin) that occurred in the clinical studies, refer to Section X below.

# IX. SUMMARY OF NONCLINICAL STUDIES

# A. Laboratory Studies

# 1. Correlation to a Reference Method (Accuracy)

All Evaluable Set: The accuracy of the LeukoStrat® CDx FLT3 Mutation Assay was determined by comparing the results of the LeukoStrat® CDx FLT3 Mutation Assay to a validated high throughput sequencing method using specimens from the clinical trial. The samples for the method comparison study were a subset of the FLT3 bridging study samples evaluated by a clinical trial assay (CTA)&gt; the samples included all available and evaluable FLT3 mutation positive (CTA+) samples and approximately 300 FLT3 mutation negative (CTA-) samples. The negative sample subset was selected by a randomization algorithm with the proportion from each CTA laboratory test site matching the proportion from that site in the overall A2301 study. After accounting for specimens with valid results, 505 CTA+ specimens were included and 263 CTA- specimens for a total of 768 patient specimens. Four of these contained low DNA quantity and were tested on deviation. Of the 764 results, 487 were FLT3 positive by both assays and 230 were negative by both as summarized in Table 2. Percent positive agreement (PPA)&lt; negative percent agreement (NPA) and overall agreement (OA) with and without the invalid results are shown in Table 3.

Table 1. Concordance between CDx and High throughput Sequencing for All Samples

| CDx* | Sequencing | | |
| --- | --- | --- | --- |
| | Positive | Negative | Total |
| Positive | 487 | 6 | 493 |
| Negative | 31 | 230 | 261 |
| Invalid | 7 | 3 | 10 |
| Total | 525 | 239 | 764 |

Table 2. Agreement between CDx and High throughput Sequencing

| Agreement Measure | Without CDx Invalid | | With CDx Invalid | |
| --- | --- | --- | --- | --- |
| | Percent Agreement (N) | 95% CI(1) | Percent Agreement (N) | 95% CI(1) |
| PPA | 94.0% (487/518) | (91.6%, 95.9%) | 92.8% (487/525) | (90.2%, 94.8%) |
| NPA | 97.5% (230/236) | (94.5%, 99.1%) | 96.2% (230/239) | (93.0%, 98.3%) |
| OPA | 95.1% (717/754) | (93.3%, 96.5%) | 93.8% (717/764) | (91.9%, 95.4%) |

PMA P160040: FDA Summary of Safety and Effectiveness Data

{5}

# Results by Mutation Type

FLT3-ITD: The ITD detected population refers to the samples that harbor only ITD mutations based on Sequencing. Among the 378 ITD samples, $64\%$ showed only one (1) ITD variant with the remaining containing multiple ITD mutations. The ITD insert length ranged from 3 bp to 209 bp. Most samples with ITD mutations were of insert lengths less than 100 bp ( $&gt;85\%$ ). Thirty-seven (37) of the ITD samples contained insert lengths greater than or equal to 100 bp. The size distribution of the ITDs is shown below in Figure 2. The number of specimens (y axis) is shown as a function of the insert size on the Y-axis.

Figure 1: Distribution of ITD Insert Sizes by CDx (381 Positive Samples, N Insert Sizes = 554, Mean=55.6).

Nine patients failed to yield valid ITD results with the CDx. There were 57 discordant results among the 764 samples analyzed. Of the 57 discordant results, 50 showed low variant fraction reads by Sequencing and the CDx reported mutation negative based on the clinical cut-off (SR=0.05). The point estimates of PPA, NPA and OPA are $86.8\%$ , $97.3\%$ , and $91.4\%$ respectively including the CDx invalids (Table 4a). Without the CDx invalids, the PPA, NPA and OPA are at or above $88\%$ .

Table 3a. Agreement on ITD between CDx and High throughput Sequencing for FLT3-ITD results

| Agreement Measure | Without CDx Invalid | | With CDx Invalid | |
| --- | --- | --- | --- | --- |
| | Percent Agreement (N) | 95% CI(1) | Percent Agreement (N) | 95% CI(1) |
| PPA | 88.0% (375/426) | (84.6%, 91.0%) | 86.8% (375/432) | (83.2%, 89.9%) |
| NPA | 98.2% (323/329) | (96.1%, 99.3%) | 97.3% (323/332) | (94.9%, 98.8%) |
| OPA | 92.5% (698/755) | (90.3%, 94.2%) | 91.4% (698/764) | (89.1%, 93.3%) |

FLT3-TKD: The TKD detected population refers to samples that harbor only TKD mutations, based on high throughput sequencing. Among the 94 TKD samples, $79\%$ contained one (1) TKD variant (substitution or deletion) while $20/94$ ( $21\%$ ) contained two TKD variants. As expected, the single nucleotide substitution at codon D835 was the predominant mutation, mainly as D835Y. The D835H, D835V

PMA P160040: FDA Summary of Safety and Effectiveness Data

{6}

and I836S mutations were also observed at lower prevalence. Thirteen percent (13%) of the TKD positive samples contained the deletion variant at I836 as either a deletion only or deletion plus substitution.

One hundred thirty-seven (92.6%) of 148 samples, identified as positive for a TKD mutation by sequencing, were identified as TKD positive by the CDx assay. Six hundred eleven (98.2%) of the 616 samples identified as TKD negative by sequencing, were TKD negative by the CDx assay. Eight patient sampless yielded an invalid TKD result by the CDx and 8 of the 764 samples tested were discordant. The 8 discordant results showed low variant fraction reads by high throughput sequencing and the SR for the CDx reported mutation negative result found to be below the assay cut-off. Results for TKD agreement with and without invalids are summarized in 4b.

Table 4b. Agreement on TKD between CDx and High throughput Sequencing for FLT3-TKD

| Agreement Measure | Without CDx Invalid | | With CDx Invalid | |
| --- | --- | --- | --- | --- |
| | Percent Agreement (N) | 95% CI(1) | Percent Agreement (N) | 95% CI(1) |
| PPA | 94.5% (137/145) | (89.4%, 97.6%) | 92.6% (137/148) | (87.1%, 96.2%) |
| NPA | 100.0% (611/611) | (99.4%, 100.0%) | 99.2% (611/616) | (98.1%, 99.7%) |
| OPA | 98.9% (748/756) | (97.9%, 99.5%) | 97.9% (748/764) | (96.6%, 98.8%) |

# Results by Specimen Type

Results were analyzed separately for peripheral blood and bone marrow and demonstrated to be comparable. The Results are shown Tables 5a through 5d

Table 5a. Agreement between CDx and High throughput Sequencing (ITD BM)

| Agreement | CDx ITD Conditional on Sequencing ITD | | Sequencing ITD Conditional on CDx ITD | |
| --- | --- | --- | --- | --- |
| | % (n/N) | 95% CI (1) | % (n/N) | 95% CI (1) |
| PPA | 87.3% (309/354) | 83.4%, 90.6% | 98.4% (309/314) | 96.3%, 99.5% |
| NPA | 98.2% (274/279) | 95.9%, 99.4% | 85.9% (274/319) | 81.6%, 89.5% |
| OPA | 92.1% (583/633) | 89.7%, 94.1% | 92.1% (583/633) | 89.7%, 94.1% |

(1) The 95% CI was calculated using the Exact (Clopper-Pearson) method

PMA P160040: FDA Summary of Safety and Effectiveness Data

{7}

Table 5b. Agreement between CDx and High throughput Sequencing (ITD PB)

| Agreement | CDx ITD Conditional on Sequencing ITD | | Sequencing ITD Conditional on CDx ITD | |
| --- | --- | --- | --- | --- |
| | % (n/N) | 95% CI (1) | % (n/N) | 95% CI (1) |
| PPA | 90.4% (66/73) | 81.2%, 96.1% | 98.5% (66/67) | 92.0%, 100% |
| NPA | 98.0% (49/50) | 89.4%, 99.9% | 87.5% (49/56) | 75.9%, 94.8% |
| OPA | 93.5% (115/123) | 89.6%, 97.2% | 93.5% (115/123) | 89.6%, 97.2% |

(1) The 95% CI was calculated using the Exact (Clopper-Pearson) method

Table 5c. Agreement between CDx and High throughput Sequencing (TKD BM)

| Agreement | CDx ITD Conditional on Sequencing ITD | | 4 Sequencing ITD Conditional on CDx ITD | |
| --- | --- | --- | --- | --- |
| | % (n/N) | 95% CI (1) | % (n/N) | 95% CI (1) |
| PPA | 95.8% (115/120) | 90.5%, 98.6% | 100% (115/115) | 96.8%, 100% |
| NPA | 100% (511/511) | 99.3%, 100% | 99.0% (511/516) | 97.8%, 99.7% |
| OPA | 99.2% (626/631) | 98.2%, 99.7% | 99.2% (626/631) | 98.2%, 99.7% |

(1) The 95% CI was calculated using the Exact (Clopper-Pearson) method

Table 5d. Agreement between CDx and High throughput Sequencing (TKD PB)

| Agreement | CDx ITD Conditional on Sequencing ITD | | Sequencing ITD Conditional on CDx ITD | |
| --- | --- | --- | --- | --- |
| | % (n/N) | 95% CI (1) | % (n/N) | 95% CI (1) |
| PPA | 88.0% (22/25) | 68.8%, 97.5% | 100% (22/22) | 84.6%, 100% |
| NPA | 100% (99/99) | 96.3%, 100% | 97.1% (99/102) | 91.6%, 99.4% |
| OPA | 97.6% (121/124) | 93.1%, 99.5% | 97.6% (121/124) | 93.1%, 99.5% |

(1) The 95% CI was calculated using the Exact (Clopper-Pearson) method

PMA P160040: FDA Summary of Safety and Effectiveness Data

{8}

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 9

## 2. Analytical Sensitivity

The detection capability of the LeukoStrat® CDx FLT3 Mutation Assay was evaluated.

## Limit of blank (LoB)

The limit of blank of the assay was assessed using 100% wild-type DNA from the Nalm6 wild type cell line. The study was conducted with DNA from 66 replicates diluted to 10 ng/μl using the ITD or TKD reaction. Under all conditions, the SR was 0.00 for the ITD assay and between 0.00 and 0.01 in the TKD assay. Thus, the false positive rates are 0% for both ITD and TKD.

## Limit of detection (LoD)

LoD of the assay was evaluated in two studies. The first study used contrived samples created by blending cell lines with leukocyte-depleted whole blood. Cell line samples were used to represent three ITD insert sizes: 30 bp insert, 126 bp insert, and a 279 bp insert as described in Table 7. Additional cell lines (TVI-TKD) containing either the D835 mutation or the I836 TKD mutation were also assessed. DNA was diluted to 5 ng/μL, 10 ng/μL, and 15 ng/μL and tested at two allelic ratios for each cell line as shown in the Table 8. Extraction Controls were diluted and tested only at 10 ng/μL as Negative Controls. Each mutation was tested in a set of at least 22 replicates in an effort to achieve a target of &gt;95% of valid results per dilution level. The study was designed to detect the lowest mutant/wild-type allelic ratio (AR) in which a specimen was still called positive because it was above the SR 0.05 that could be detected above the LoB in at least 95% of cases.

The LoD for the LeukoStrat® CDx FLT3 Mutation Assay was 3% for ITD insert of 30 bp, 5% for 126 bp ITD and 33% for 279 bp ITD. The LoD for the TKD mutation at codon D835 was 5%. While the LoD for the larger ITDs was significant, as shown in Figure 1, the majority (95%) of patients had ITDs &lt; 120.

Table 8. Limit of Detection with Cell Lines

| FLT 3 | AR | DNA Dilution | Sample Size | Valid N | Valid N >LoB | % >LoB | Mean SR | Std Dev | Median SR | Min-Max SR | % CV |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 30 bp ITD | 0.03 | 5 ng/μL | 33 | 22 | 21 | 95.5% | 0.028 | 0.007 | 0.03 | 0.00 - 0.03 | 24.7 |
| | | 10 ng/μL | 33 | 33 | 33 | 100.% | 0.029 | 0.003 | 0.03 | 0.02 - 0.03 | 10.0 |
| | | 15 ng/μL | 33 | 33 | 33 | 100.% | 0.028 | 0.004 | 0.03 | 0.02 - 0.03 | 13.9 |
| | 0.05 | 5 ng/μL | 33 | 30 | 30 | 100.% | 0.036 | 0.005 | 0.04 | 0.03 - 0.04 | 13.5 |
| | | 10 ng/μL | 33 | 33 | 33 | 100.% | 0.037 | 0.006 | 0.04 | 0.03 - 0.05 | 15.8 |
| | | 15 ng/μL | 33 | 33 | 33 | 100.% | 0.035 | 0.005 | 0.03 | 0.03 - 0.04 | 14.6 |

{9}

| D835 | 0.05 | 5 ng/μL | 33 | 33 | 33 | 100.% | 0.020 | 0.003 | 0.02 | 0.01 - 0.03 | 15.0 |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | 10 ng/μL | 33 | 33 | 33 | 100.% | 0.020 | 0.002 | 0.02 | 0.01 - 0.02 | 8.8 |
| | | 15 ng/μL | 33 | 33 | 33 | 100.% | 0.020 | 0.000 | 0.02 | 0.02 - 0.02 | 0.0 |
| | 0.08 | 5 ng/μL | 33 | 33 | 33 | 100.% | 0.032 | 0.004 | 0.03 | 0.03 - 0.04 | 13.4 |
| | | 10 ng/μL | 33 | 33 | 33 | 100.% | 0.034 | 0.005 | 0.03 | 0.03 - 0.04 | 14.6 |
| | | 15 ng/μL | 33 | 33 | 33 | 100.% | 0.038 | 0.004 | 0.04 | 0.03 - 0.04 | 11.6 |
| 126bpITD | 0.05 | 5 ng/μL | 22 | 22 | 22 | 100.% | 0.043 | 0.005 | 0.04 | 0.04 - 0.05 | 11.0 |
| | | 10 ng/μL | 22 | 22 | 22 | 100.% | 0.039 | 0.004 | 0.04 | 0.03 - 0.04 | 9.1 |
| | | 15 ng/μL | 22 | 22 | 22 | 100.% | 0.030 | 0.000 | 0.03 | 0.03 - 0.03 | 0.0 |
| 279bpITD | 0.30 | 5 ng/μL | 22 | 19 | 19 | 86.3% | 0.091 | 0.014 | 0.09 | 0.08 - 0.14 | 15.8 |
| | | 10 ng/μL | 22 | 22 | 20 | 90.9% | 0.063 | 0.021 | 0.07 | 0.00 - 0.08 | 34.0 |
| | | 15 ng/μL | 22 | 22 | 21 | 95.5% | 0.049 | 0.013 | 0.05 | 0.00 - 0.06 | 26.6 |
| | 0.33(repeat) | 5 ng/μL | 22 | 22 | 22 | 100.% | 0.118 | 0.012 | 0.12 | 0.09 - 0.13 | 9.8 |
| | | 10 ng/μL | 22 | 22 | 22 | 100.% | 0.117 | 0.014 | 0.115 | 0.09 - 0.14 | 11.5 |
| | | 15 ng/μL | 22 | 22 | 21 | 95.5% | 0.154 | 0.262 | 0.10 | 0.00 - 1.32 | 170.3 |

A second study with clinical specimens was conducted to confirm the LoD observations obtained with cell lines. Five clinical samples were diluted with clinical negative samples in order to yield a targeted signal ratio (TSR) within the linear range of an appropriate cell line standard curve. Each specimen was diluted to 5 levels representing a low negative (LN), high negative (HN), near the cut-off (CO), a low positive (LP), and a medium positive (MP). These linear range samples were tested in the FLT3 Mutation Assay and an average SR value was determined. Each clinical LoD sample dilution was tested 20 times for each dilution level over four nonconsecutive days (5 replicates per day) by one operator using one equipment set. The allelic ratio (AR) of each clinical LoD sample dilution was calculated using the AR estimated from the cell line standard curves. The ARs of the clinical LoD samples were estimated based on the study meeting the following acceptance criteria:

- The SR and AR where FLT3 mutations can be detected above the limit of blank (LoB) in $\geq 95\%$ of replicates (Analytical LoD).
- The AR near the clinical cut-off, a SR of 0.04 - 0.06 (Cut-off).

PMA P160040: FDA Summary of Safety and Effectiveness Data

{10}

- The AR and SR that is detected above the clinical cut-off in $\geq 95\%$ of replicates (Above Cut-off).

The AR is representative of multiple peaks and therefore does not directly indicate the percentage of mutant in the sample. Study results show that the assay LoD was determined as 0.144 AR (12.6%) for the I836 mutation, 0.089 AR (8.2%) for the D835 mutation, 0.107 AR (9.7%) for the small ITD, 0.189 AR (15.9%) for the medium ITD, and 0.539 AR (35%) for the large ITD. Results are summarized for each of the 5 dilution levels in Table 9. The %CV for the I836 mutation was 5.7%, 4.5% for the D835 mutation, 7.1% for the 24 bp ITD, 7.1% for the 66 bp ITD, and 25.6% for the 217 bp ITD. Thus, all FLT3 mutations met the acceptance criterion of 25% CV except the large ITD.

Table 9a. SR, AR and LoD per each Sample and Dilution Level

| Sample ID | Mutation | Level | TSR | SR Mean | AR of Blend | Valid N | N (%) SR > LoB | N (%) SR > 0.05 | *Classification |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| FPB006 | TKD I836 | LN | 0.02 | 0.02 | 0.039 | 20 | 20 (100.0) | 0 | Analytical LoD |
| | | HN | 0.03 | 0.03 | 0.057 | 20 | 20 (100.0) | 0 | - |
| | | CO | 0.05 | 0.05 | 0.094 | 20 | 20 (100.0) | 16 (80.0%) | Cut-off |
| | | LP | 0.08 | 0.07 | 0.144 | 20 | 20 (100.0) | 20 (100.0) | Above Cut-off |
| | | MP | 0.13 | 0.12 | 0.224 | 20 | 20 (100.0) | 20 (100.0) | - |
| TBM232 | TKD D835 | LN | 0.02 | 0.02 | 0.044 | 20 | 20 (100.0) | 0 | Analytical LoD |
| | | HN | 0.03 | 0.03 | 0.065 | 20 | 20 (100.0) | 0 | - |
| | | CO | 0.05 | 0.05 | 0.107 | 20 | 20 (100.0) | 19 (95.0) | Cut-off |
| | | LP | 0.08 | 0.08 | 0.165 | 20 | 20 (100.0) | 20 (100.0) | Above Cut-off |
| | | MP | 0.13 | 0.15 | 0.257 | 20 | 20 (100.0) | 20 (100.0) | - |
| TBM119 | ITD 24 bp | LN | 0.01 | 0.02 | 0.023 | 20 | 20 (100.0) | 0 | Analytical LoD |
| | | HN | 0.02 | 0.03 | 0.047 | 20 | 20 (100.0) | 0 | - |
| | | CO | 0.04 | 0.05 | 0.089 | 20 | 20 (100.0) | 20 (100.0) | Cut-off |
| | | LP | 0.07 | 0.08 | 0.152 | 20 | 19 (95.0) | 19 (95.0) | Above Cut-off |
| | | MP | 0.13 | 0.13 | 0.269 | 20 | 20 (100.0) | 20 (100.0) | - |
| TPB161 | ITD 66 bp | LN | 0.02 | 0.02 | 0.045 | 20 | 20 (100.0) | 0 | Analytical LoD |
| | | HN | 0.03 | 0.03 | 0.066 | 20 | 20 (100.0) | 0 | - |
| | | CO | 0.05 | 0.05 | 0.110 | 20 | 20 (100.0) | 18 (90.0) | Cut-off |
| | | LP | 0.09 | 0.08 | 0.189 | 20 | 20 (100.0) | 20 (100.0) | Above Cut-off |
| | | MP | 0.14 | 0.13 | 0.280 | 20 | 20 (100.0) | 20 (100.0) | - |
| TPB329 | ITD 217 bp | LN | 0.01 | 0 | 0.073 | 20 | 2 (10.0) | 0 | Analytical LoD |
| | | HN | 0.02 | 0.02 | 0.147 | 20 | 15 (75.0) | 0 | - |
| | | CO | 0.04 | 0.04 | 0.276 | 20 | 20 (100.0) | 9 (45.0) | Cut-off |
| | | LP | 0.08 | 0.08 | 0.539 | 20 | 19 (95.0) | 19 (95.0) | Above Cut-off |
| | | MP | 0.13 | 0.13 | 0.838 | 20 | 20 (100.0) | 20 (100.0) | - |
| True Neg ITD | None | TN | N/A | 0 | 0 | 20 | 0 | 0 | N/A |
| TrueNeg TKD | None | TN | N/A | 0 | 0 | 20 | 0 | 0 | N/A |

PMA P160040: FDA Summary of Safety and Effectiveness Data

{11}

*Classifications are defined as 1: Analytical LoD = lowest AR where samples were detected 95% of the time above the LoB, 2: cut-off is AR where samples were near SR 0.05, and 3: above Cut-off = lowest AR where samples could be detected 95% of the time above SR 0.05.

The AR does not represent the percent mutant in a specimen. The AR value was converted to % mutant. The LoD in terms of % mutation is shown in Table 9b below.

Table 9b. Analytical Sensitivity Allelic Ratio and % Mutant
| Sample ID | Mutation | Mutation Classification | Above Cut-off 95% SR≥ 0.05 | | |
| --- | --- | --- | --- | --- | --- |
| | | | AR | SR | %Mut |
| FPB006 | TKD I836 | TKD I836 Deletion | 0.144 | 0.07 | 12.6 |
| TBM232 | TKD D835 | TKD D835 Substitution | 0.089 | 0.05 | 8.2 |
| TBM119 | ITD 24bp | Small ITD Insert <30bp | 0.107 | 0.05 | 9.7 |
| TPB161 | ITD 66bp | Medium ITD Insert 30-100bp | 0.189 | 0.08 | 15.9 |
| TPB329 | ITD 217bp | Large ITD Insert ~200bp | 0.539 | 0.08 | 35.0 |

## 3. Analytical Specificity

### Cross-Reactivity

To assess the potential for amplification of non-specific products from human genomic DNA, in silico analysis of the PCR primers used in the assay was performed. No non-specific products were predicted for the assay.

### Interfering Substances

Various endogenous and exogenous substances listed in Table 10 were evaluated for interference of the LeukoStrat® CDx FLT3 Mutation Assay. Thirty-seven clinical specimens were spiked with a mixture of exogenous and endogenous materials and compared to the same specimens spiked with only the solvents carrying the materials. The study consisted of 11 BM and 26 PB specimens with 15 FLT3 negative and 22 FLT3 positive specimens (14 ITDs and 8 TKDs).

Potential interfering substances were added to clinical specimens (both positive and negative for ITD and TKD mutations) and tested for mutation status based on mutant: wild type SR against the cutoff of 0.05 (negative &lt; 0.05, positive ≥ 0.05). Multiple substances (except wash buffer) were pooled and tested together due to limitations of sample volume and assay throughput. The substances were tested at the recommended testing levels per CLSI EP7-A2. No interference was observed based on mutation status using any of the potential interfering substances. The LeukoStrat®

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 12

{12}

CDx FLT3 Mutation Assay is capable of detecting ITD mutations sized 18 bp to 114 bp and TKD mutations in the presence the interferents shown in the table below.

Table 10. Potential Interfering Substances Tested
| B (Blank) | I (Interferent) |
| --- | --- |
| • Water
• Dimethylsulfoxide (DMSO)
• Tris(hydroxymethyl)amino methane (Tris) | • 0.8 mg/mL sodium heparin in water
• 60 mg/mL human serum albumin in water
• 24 μg/mL cytarabine in water
• 180 ng/mL daunorubicin in DMSO
• 0.19 mg/mL bilirubin in Tris
• 5% v/v Intralipid
• 2 mg/mL hemoglobin |

## 4. Reproducibility

The reproducibility and precision of the LeukoStrat® CDx FLT3 Mutation Assay was assessed in two studies, the first using cell line specimens and the second using clinical specimens.

### a. Reagent lot and instrument variability

In the first study, each operator tested a nine-member test panel using three reagent lots and instruments. The capillary electrophoresis detection portion of the assay was performed on each of three instruments. All test steps utilizing different combinations of reagents and equipment were performed in parallel. Contrived samples comprised of blended cultured cell lines added to leukocyte depleted whole blood (from normal healthy donors) were tested as shown in Figure 2. The samples consisted of ITD samples containing inserts of 30 bp and 126 bp and the D835 TKD mutation. Samples represented low (near cutoff), mid, and high (100% mutant cell line) mutant:wild type SRs for small internal tandem duplication (ITD) insert, large ITD insert, and tyrosine kinase domain (TKD) mutation as shown in Figure 2. The study was conducted with 27 replicates per level. The acceptance criteria were that the total coefficient of variation (CV) shall be ≤25% for each panel member with a mean mutant:wild type SR ≥0.05.

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 13

{13}

Figure 2. Workflow diagram, testing performed for each panel member by one operator. QIA=QIAcube; TC=thermal cycler; ABI=ABI 3500xL Genetic Analyzer. Extracted DNA from QIAcube instruments (3 replicates per reagent lot) was amplified together on 1 thermal cycler; each amplified plate was then tested on each of 3 CE instruments (to generate 9 mutant:wild type SRs per reagent lot).

The results showed that all mutant cells at each level tested positive (SR&gt;0.05) and the percent coefficient of variation (%CV) was within the pre-specified acceptance criterion of 25% for this assay. For the ITD mutation samples, the SR %CV ranged from 3.0% to 8.4%. For the TKD mutation samples, the SR %CV ranged from 5.4% to 10.6%. Overall, the LeukoStrat® CDx FLT3 Mutation Assay lot-to-lot and instrument-to-instrument reproducibility has a SR %CV of less than 11% as summarized in Table 11.

Table 11. Components of Variance per Positive Panel Member, Mutant:Wild Type SR

| PM | Description | Sample Size | SR Average | Variation due to | | | Total | Total CV, % |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | Reagent Lot | Instrument | Random Error | | |
| PM1 | 30 bp ITD, low | 27 | 0.063 | 0.004 (61.67%) | 0.001 (2.08%) | 0.003 (36.25%) | 0.005 | 7.4 |
| PM2 | 30 bp ITD, mid, | 27 | 0.918 | 0.000 (0.00%) | 0.000 (0.00%) | 0.045 (100.0%) | 0.045 | 4.9 |
| PM3 | 30 bp ITD, high | 27 | 1.746 | 0.032 (33.77%) | 0.000 (0.00%) | 0.045 (66.23%) | 0.055 | 3.0 |
| PM4 | 126 bp, low | 27 | 0.076 | 0.005 (58.60%) | 0.001 (1.36%) | 0.004 (40.05%) | 0.007 | 8.4 |
| PM5 | 126 bp, mid | 27 | 0.144 | 0.008 (50.42%) | 0.001 (1.12%) | 0.008 (48.46%) | 0.011 | 7.0 |
| PM6 | 126 bp, high | 27 | 0.379 | 0.008 (20.86%) | 0.005 (7.41%) | 0.014 (71.73%) | 0.017 | 4.2 |

PMA P160040: FDA Summary of Safety and Effectiveness Data

{14}

| PM | Description | Sample Size | SR Average | Variation due to | | | Total | Total CV, % |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | Reagent Lot | Instrument | Random Error | | |
| PM7 | D835, low | 27 | 0.071 | 0.002 (25.00%) | 0.001 (2.08%) | 0.003 (72.92%) | 0.004 | 5.4 |
| PM8 | D835, mid | 27 | 0.157 | 0.004 (4.56%) | 0.000 (0.00%) | 0.016 (95.44%) | 0.017 | 10.6 |
| PM9 | D835, high | 27 | 0.309 | 0.000 (0.00%) | 0.005 (4.83%) | 0.021 (95.17%) | 0.022 | 7.1 |

Variation is expressed in standard deviations. Values in parentheses are percentage of total variation for each source of variability. $\mathrm{CV} =$ coefficient of variation; $\mathrm{PM} =$ panel member.

## b. Operator-to-operator and Between-run variability

Variability was evaluated for precision of the LeukoStrat® CDx FLT3 Mutation Assay. The same nine panel members tested in the lot-to-lot study were tested each by three different operators, using one reagent set and one instrument set, over 15 runs. Each of the operators processed a total of 10 replicates of a test panel member through mononuclear cells (MNC) collection (30 replicates total/panel member). For each panel member, MNCs were isolated over a one day processing period. Following accumulation of MNCs, samples were tested with the remainder of the assay over 31 calendar days. Acceptance criteria identified by the company are that the total CV shall be $\leq 25\%$ for each panel member with a mean mutant:wild type SR $\geq 0.05$ .

The results showed that the percent coefficient of variation (%CV) for all ITD and TKD mutant cells at each level tested positive (SR&gt;0.05) was within the acceptance criterion of $25\%$ . For the ITD mutation samples, the SR %CV ranged from $6.6\%$ to $13.3\%$ . For the TKD mutation samples, the SR %CV ranged from $7.9\%$ to $9.3\%$ . Overall, the LeukoStrat® CDx FLT3 Mutation Assay lot-to-lot and instrument-to-instrument reproducibility has a SR %CV of less than $11\%$ as summarized in Table 12.

Table 12. Components of Variance per Panel Member for Between-Operator and Between Run

| PM | Description | Sample Size | SR Avg. | Variation due to | | | Total Variation | Total % CV |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | Operator | Run Day | Random Error | | |
| PM1 | 30 bp ITD, low | 30 | 0.079 | 0.000 (0.31%) | 0.005 (50.61%) | 0.005 (49.08%) | 0.007 | 9.4 |
| PM2 | 30 bp ITD, mid | 30 | 0.709 | 0.024 (16.70%) | 0.028 (23.09%) | 0.045 (60.21%) | 0.058 | 8.2 |
| PM3 | 30 bp, high | 30 | 1.609 | 0.029 (7.38%) | 0.092 (76.01%) | 0.043 (16.61%) | 0.106 | 6.6 |
| PM4 | 126 bp, low | 30 | 0.077 | 0.002 (4.39%) | 0.007 (51.72%) | 0.007 (43.89%) | 0.010 | 13.3 |

PMA P160040: FDA Summary of Safety and Effectiveness Data

{15}

| PM | Description | Sample Size | SR Avg. | Variation due to | | | Total Variation | Total % CV |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | Operator | Run Day | Random Error | | |
| PM5 | 126 bp, mid | 30 | 0.154 | 0.008 (19.19%) | 0.013 (56.83%) | 0.009 (23.98%) | 0.018 | 11.6 |
| PM6 | 126 bp | 30 | 0.338 | 0.012 (16.52%) | 0.025 (65.49%) | 0.013 (17.99%) | 0.030 | 9.0 |
| PM7 | D835, low | 30 | 0.090 | 0.000 (0.00%) | 0.005 (33.11%) | 0.007 (66.89%) | 0.008 | 9.3 |
| PM8 | D835, mid | 30 | 0.156 | 0.000 (0.00%) | 0.008 (42.67%) | 0.009 (57.33%) | 0.012 | 7.9 |
| PM9 | D835, high | 28 | 0.304 | 0.012 (22.35%) | 0.018 (48.48%) | 0.014 (29.17%) | 0.025 | 8.4 |

In the second study, reproducibility was assessed using clinical DNA samples from 8 clinical samples (4 PB and 4 BM) with ITD lengths of 21 bp, 24 bp, 66 bp, 90 bp and 217 bp, TKD D835 substitution, TKD I836 deletion, and FLT3 negative samples. DNA from FLT3 negative clinical specimens was pooled and used to dilute the FLT3 positive samples in order to achieve three target SR levels near the assay's clinical cut-off (i.e., high negative, low positive, and moderate positive). Five FLT3 positive clinical samples originated from PB and two from BM. Three replicates of 5 ITD positive, 2 TKD positive and one pooled true negative sample were tested by three different operators/instrument sets using 1 reagent lot over five non-consecutive days at three dilution levels for positive samples and neat for the negative. Each operator tested 15 replicates total per level for a total of 45 replicates per dilution level.

The total $\% \mathrm{CV}$ of all mutation types and levels are shown in Table 13 and the $\% \mathrm{CV}$ for all mutation types, except the long ITD insert (217 bp) sample, ranged from $4.2\%$ to $16.1\%$ . The sample with a 217 bp mutation $\% \mathrm{CV}$ ranged from $26.9\%$ to $27.2\%$ . The low positive (LP) dilution level $\% \mathrm{CV}$ was $26.9\%$ for 217 bp, therefore failing the study acceptance criteria of $\leq 25\%$ CV for SR. Results show that acceptance criteria were met for both D835 and I836 TKD mutations and for ITD mutations up to 217 bp. Variation for the 217 bp ITD mutation exceeded $25\%$ , thus indicating greater imprecision around the largest ITDs.

Table 13. Components of Variance per Mutation Type and Dilution Level

| Sample ID | Mut Type | Dilution Level | Mean SR | SR Variation Due to | | | Total Variation | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | Operator/ Instrument SD (%) | Run Day SD (%) | Random Error SD (%) | SD | % CV |
| FPB006 | TKD I836 | HN | 0.03 | 0.000 (3.22%) | 0.000 (0.00%) | 0.002 (96.78%) | 0.002 | 7.1 |
| | | LP | 0.077 | 0.001 (2.60%) | 0.000 (0.00%) | 0.005 (97.40%) | 0.005 | 5.9 |

PMA P160040: FDA Summary of Safety and Effectiveness Data

{16}

| Sample ID | Mut Type | Dilution Level | Mean SR | SR Variation Due to | | | Total Variation | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | Operator/ Instrument SD (%) | Run Day SD (%) | Random Error SD (%) | SD | % CV |
| | | MP | 0.132 | 0.002 (6.67%) | 0.003 (17.43%) | 0.005 (75.90%) | 0.006 | 4.6 |
| TPB262 | TKD D835 | HN | 0.04 | 0.001 (7.13%) | 0.000 (0.00%) | 0.002 (92.87%) | 0.002 | 5.3 |
| | | LP | 0.08 | 0.002 (14.02%) | 0.001 (2.47%) | 0.004 (83.51%) | 0.004 | 5.3 |
| | | MP | 0.165 | 0.003 (16.28%) | 0.000 (0.00%) | 0.007 (83.72%) | 0.007 | 4.2 |
| TBM163 | ITD 21 bp | HN | 0.03 | 0.000 (0.00%) | 0.000 (0.00%) | 0.001 (100.0%) | 0.001 | 5 |
| | | LP | 0.074 | 0.000 (0.00%) | 0.002 (8.08%) | 0.005 (91.92%) | 0.005 | 7.2 |
| | | MP | 0.133 | 0.002 (14.5%) | 0.000 (0.00%) | 0.005 (85.54%) | 0.006 | 4.4 |
| TBM119 | ITD 24 bp | HN | 0.029 | 0.000 (0.00%) | 0.000 (0.00%) | 0.004 (100.0%) | 0.004 | 15.2 |
| | | LP | 0.07 | 0.000 (0.00%) | 0.000 (0.92%) | 0.004 (99.08%) | 0.004 | 5.3 |
| | | MP | 0.147 | 0.002 (8.20%) | 0.001 (3.28%) | 0.006 (88.52%) | 0.007 | 4.5 |
| TPB161 | ITD 66 bp | HN | 0.029 | 0.001 (4.28%) | 0.000 (0.00%) | 0.005 (95.72%) | 0.005 | 16.1 |
| | | LP | 0.083 | 0.000 (0.00%) | 0.001 (1.13%) | 0.007 (98.87%) | 0.007 | 8 |
| | | MP | 0.185 | 0.000 (0.00%) | 0.000 (0.00%) | 0.010 (100.0%) | 0.01 | 5.3 |
| TPB267 | ITD 90 bp | HN | 0.03 | 0.001 (5.15%) | 0.000 (0.00%) | 0.003 (94.85%) | 0.003 | 10.1 |
| | | LP | 0.091 | 0.004 (25.23%) | 0.002 (8.42%) | 0.007 (66.35%) | 0.008 | 8.5 |
| | | MP | 0.206 | 0.013 (44.26%) | 0.005 (7.34%) | 0.013 (48.40%) | 0.019 | 8.5 |
| TPB329 | ITD 217 bp | HN | 0.032 | 0.001 (0.90%) | 0.002 (7.20%) | 0.008 (91.90%) | 0.009 | 27.2 |
| | | LP | 0.079 | 0.013 (31.42%) | 0.009 (14.86%) | 0.017 (53.71%) | 0.023 | 26.9 |
| | | MP | 0.162 | 0.029 (36.75%) | 0.015 (9.86%) | 0.035 (53.39%) | 0.047 | 27.2 |

Controls: Precision of the controls was also assessed and obtained from the above described studies. Results for the ITD and TKD positive controls for the lot-to-lot and operator-to-operator demonstrated that the between lot and between operator results were within 0.01 SR difference for ITD positive controls. For TKD positive control, the results were within 0.01 SR for between lot and .04 SR for between-operator. Data demonstrates that the precision on the controls was acceptable.

PMA P160040: FDA Summary of Safety and Effectiveness Data

{17}

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 18

5. **Carryover (CO) and Cross-Contamination (CC)**

To assess for potential carryover and cross-contamination testing of the FLT3 Mutation Assay was performed over 4 ITD runs and 4 TKD runs. Samples were positioned throughout DNA extraction, amplification, digestion, and capillary electrophoresis plates to challenge carryover and cross-contamination conditions which could be expected to cause false positive results. No Template Control carryover testing was performed over 2 ITD runs and 2 TKD runs. On 96 well plates, alternating columns of positive amplicon and water were tested via capillary electrophoresis in order to capture carryover from positive samples into adjacent water filled wells.

For carryover and cross-contamination challenges, DNA was directly extracted from three cell lines (i.e., ITD positive, TKD positive, and wild type) over 7 QIAcube runs. Cell line positions on the QIAcube alternated between mutant positive and wild type. The isolated DNA was tested on 4 ITD and 4 TKD CE runs on the ABI3500xL instruments in the order of QIAcube extractions.

For NTC carryover challenge, water was carried through both the DNA amplification process and the EcoRV digestion. Columns of water samples were positioned adjacent to columns of positive amplicon samples in each of the 96-well plates and tested as 2 ITD runs and 2 TKD runs.

No carryover or cross-contamination was observed in the wild type (mutation negative) samples and no carryover was detected in the NTC (water) samples.

6. **Specimen Volume**

Two specimen volume studies using contrived samples or clinical specimens were carried out to determine the range of sample amounts necessary to perform the LeukoStrat® CDx FLT3 Mutation Assay.

a. **Specimen Volume 1 (Contrived Samples)**

The objective of this study was to demonstrate equivalency when using volumes of blood ranging from 0.9 to 3.3 mL in the FLT3 Mutation Assay. The assay is intended to use 3 mL PB specimen or 0.75 mL BM aspirate, when available. However, a minimum specimen volume of 1 mL PB and 0.25 mL BM aspirate allows testing of specimens with low volume.

Over 4 days, one operator processed 20 replicates of contrived test samples, consisting of FLT3-ITD mutant cells and FLT3 wild-type cultured cells, with the exceptions to sample and RPMI volumes listed below. There were 4 conditions each day, 5 replicates each. The DNA was extracted, and then frozen at -20°C until the PCR step of the assay.

- Condition 1: 0.9 mL sample, 5.1 mL RPMI
- Condition 2: 1 mL sample, 5 mL RPMI
- Condition 3: 3.3 mL sample, 2.7 mL RPMI
- Condition 4: 1 mL sample, 3 mL RPMI

{18}

- Days 1 and 2: 5 replicates each of Conditions 1-4 on Low Challenge sample
- Days 3 and 4: 5 replicates each of Conditions 1-4 on High Challenge sample
- Low Challenge: 6 million cultured cells/ mL
- High Challenge: 100 million cultured cells/ mL

Mutant: Wild Type SR, DNA yield, and MNC counts were evaluated among the different specimen volumes. Means among the different sample volumes were compared using analysis of variance (ANOVA). For each sample, the mean values of SRs, DNA yield, and MNC concentration should not show a difference that is statistically significant, i.e., P-value should exceed 0.05 using ANOVA.

SRs were consistent across the 4 test conditions. This was the anticipated result when using different initial blood volumes over the Ficoll gradient, and the study supports that 3 mL PB can be processed over the Ficoll gradient. While there were some instances of statistically significant differences in mean SRs, the differences had no practical impact. The mean differences were small, and the clinical interpretation of the assay result would not change.

## b. Specimen Volume 2 (Clinical Specimens)

The purpose of the study was to evaluate the variability, if any, in SR when processing different specimen volumes over the Ficoll gradient. Four clinical specimens, representing both ITD and TKD mutations and both specimen matrices were tested in triplicate under each of the following conditions:

- Condition 1: 0.9 mL PB and 5.1 mL RPMI; OR 0.22 mL BM and 5.78 mL RPMI
- Condition 2: 1 mL PB and 5 mL RPMI; OR 0.25 mL BM and 5.75 mL RPMI
- Condition 3: 3.3 mL PB and 2.7 mL RPMI; OR 0.83 mL BM and 5.17 mL RPMI

Specimens were tested at 90% minimum volume, nominal minimum volume, and 110% maximum volume in triplicate. Mutant: wild type SRs were compared and no changes in mean mutant: wild type SR were observed, demonstrating that LeukoStrat® CDx FLT3 Mutation Assay results were not affected by specimen volume and that a range of specimen volumes is acceptable for the assay.

## 7. Guard Band Studies

### a. PCR Guard Band

The objective of this study was to evaluate the robustness of PCR to changes in temperature in the FLT3 Mutation Assay. The annealing, extension, and denaturation temperatures during PCR were tested at the nominal conditions shown in Table 14 and ±1°C FLT3-ITD positive control cell line, TKD positive control cell line, or wild-type cell line were tested using a total of 96 samples. Acceptance criteria were that for samples tested at nominal temperatures, control validity criteria were met. The results demonstrated the test was robust to temperature changes when detecting ITD mutations. However, the TKD was not

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 19

{19}

robust to changes in temperature. The instructions for use indicate that the PCR temperatures must be adhered to.

Table 14. Nominal PCR Conditions
| Step | ITD Program | TKD Program |
| --- | --- | --- |
| 1 | 95°C for 11 minutes | 95°C for 11 minutes |
| 2 | 94°C for 30 seconds | 94°C for 30 seconds |
| 3 | 57°C for 60 seconds | 57°C for 60 seconds |
| 4 | 72°C for 2 minutes | 72°C for 2 minutes |
| 5 | Repeat steps 2 to 4 24 times | Repeat steps 2 to 4 28 times |
| 6 | 94°C for 30 seconds | 94°C for 30 seconds |
| 7 | 60°C for 45 minutes | 60°C for 45 minutes |
| 8 | 4°C forever | 4°C forever |

b. TKD Digestion

The objective of this study was to evaluate the robustness of the tyrosine kinase domain (TKD) mutation digestion step of the assay. Three samples representing a high TKD positive, a low TKD positive, and a TKD negative were tested for 1 hour (nominal condition) or 1.5 hours. Validity criteria were met.

c. Mononuclear Cell Centrifugation Temperature

The objective of this study was to evaluate the robustness of centrifuge temperature in the FLT3 Mutation Assay. Isolating MNCs from peripheral blood or bone marrow aspirate specimens uses centrifugation for the Ficoll gradient and wash steps. The centrifuge chamber temperature setting is 20°C. Two operators processed 16 contrived samples each. Each operator centrifuged half of their samples at 17°C and the other half at 23°C. Only the lower and upper bounds were tested as it was presumed that there was no non-linearity across the range of testing. The contrived samples were comprised of normal blood spiked with the ITD PC cells that have an ITD insertion of 30 bp. MNC counts, DNA yield, and mutant:wild type SR were measured; the values were evaluated to determine if they were equivalent between the two centrifuge temperature conditions. The mean values of MNC concentrations, DNA concentrations, and SR of the samples tested at the 17°C and the 23°C centrifuge temperatures should not show a difference that is statistically significant.

All samples (32/32 ITD, 100%) passed the ITD FLT3 validity criteria. The MNC concentrations, DNA concentrations, and SR from the two centrifuge temperatures were compared to each other. There was no significant difference observed under the tested conditions.

d. FLT3 Mutation Assay Guard Band: MNC Layer Collection

The objective of this study was to evaluate the robustness of cell collection in the assay. Mononuclear cells (MNC) were isolated from peripheral blood or bone marrow aspirate specimens using a Ficoll gradient. The MNC layer was manually

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 20

{20}

collected by aspiration with a transfer pipette. In the assay, the operator collected the MNC layer in up to $3\mathrm{ml}$ . Three operators collected 2.5, 3, and $3.5\mathrm{ml}$ of the MNC layer from contrived samples comprised of normal blood spiked with the 30 bp ITD PC. The SR, DNA Yield, and MNC Concentration mean values should not show a difference that is statistically significant.

Over a two day period, 63 contrived samples were made by 3 operators each processing 21 replicates of $1\mathrm{ml}$ . The replicates were divided among MNC layer collection volumes of 2.5, 3, and $3.5\mathrm{ml}$ (7 replicates each).

Table 15. Operator Variability of Final SR, DNA Yield, and MNC Concentration

| | Operator | MNC Collection Volume (ml) | Mean | StDev | %CV | Max | Min | N |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Final SR | 1 | 2.5 | 0.033 | 0.008 | 23.5 | 0.049 | 0.025 | 7 |
| | | 3 | 0.031 | 0.003 | 9.2 | 0.035 | 0.026 | 7 |
| | | 3.5 | 0.033 | 0.006 | 18.3 | 0.044 | 0.025 | 7 |
| | 2 | 2.5 | 0.031 | 0.012 | 38.1 | 0.054 | 0.015 | 7 |
| | | 3 | 0.024 | 0.005 | 19.0 | 0.034 | 0.020 | 7 |
| | | 3.5 | 0.027 | 0.003 | 9.5 | 0.029 | 0.023 | 7 |
| | 4 | 2.5 | 0.058 | 0.009 | 15.2 | 0.050 | 0.066 | 7 |
| | | 3 | 0.057 | 0.009 | 16.2 | 0.048 | 0.065 | 7 |
| | | 3.5 | 0.055 | 0.005 | 9.6 | 0.051 | 0.060 | 7 |
| DNA Yield (ng/μl) | 1 | 2.5 | 27.5 | 3.8 | 13.7 | 31.2 | 22.0 | 7 |
| | | 3 | 31.1 | 1.7 | 5.5 | 34.1 | 28.8 | 7 |
| | | 3.5 | 29.1 | 5.2 | 18.0 | 34.0 | 18.7 | 7 |
| | 2 | 2.5 | 22.5 | 5.2 | 23.1 | 26.5 | 11.2 | 7 |
| | | 3 | 27.4 | 4.6 | 17.0 | 33.7 | 20.1 | 7 |
| | | 3.5 | 28.8 | 3.7 | 12.8 | 32.8 | 21.8 | 7 |
| | 4 | 2.5 | 54.1 | 8.62 | 15.9 | 46.2 | 62.1 | 7 |
| | | 3 | 61.4 | 9.56 | 15.6 | 52.6 | 70.2 | 7 |
| | | 3.5 | 68.6 | 3.41 | 5.0 | 65.5 | 71.8 | 7 |
| MNC Conc. (x10^3 cells / μl) | 1 | 2.5 | 0.14 | 0.05 | 37.4 | 0.2 | 0.1 | 7 |
| | | 3 | 0.20 | 0 | 0 | 0.2 | 0.2 | 7 |
| | | 3.5 | 0.16 | 0.05 | 34.0 | 0.2 | 0.1 | 7 |
| | 2 | 2.5 | 0.16 | 0.08 | 50.1 | 0.3 | 0.1 | 7 |
| | | 3 | 0.17 | 0.05 | 28.5 | 0.2 | 0.1 | 7 |
| | | 3.5 | 0.21 | 0.04 | 17.6 | 0.3 | 0.2 | 7 |
| | 4 | 2.5 | 0.23 | 0.049 | 21.3 | 0.3 | 0.2 | 7 |
| | | 3 | 0.27 | 0.049 | 18.0 | 0.3 | 0.2 | 7 |
| | | 3.5 | 0.34 | 0.053 | 15.6 | 0.4 | 0.3 | 7 |

PMA P160040: FDA Summary of Safety and Effectiveness Data

{21}

Comparing the SR results across the 3 MNC layer collection volumes showed no significant difference in the 2.5, 3, or 3.5 ml MNC collection volumes (Table 15). The actual difference between the mean signal ratios of the 3 collection volumes was between 0 and 0.004 for samples. These differences in mean signal ratios are well within the signal ratio requirement for a %CV of 25%.

The DNA yield across the 3 MNC layer collection volumes was significantly different in both Batch 1 (operators 1 and 2) and 2 (operator 4) at a 5% level of confidence. In both batches, the 2.5 ml collection volume yielded a DNA concentration significantly lower than one of the other collection volumes. The difference in DNA yields across the 3 MNC layer collection volumes is expected since more cells are collected with the increasing volume. The DNA yields show a general increasing concentration with increasing MNC collection volume.

The MNC concentrations across the 3 MNC layer collection volumes showed no statistically significant difference in Batch 1 (p-value = 0.20) but show a statistically significant difference in Batch 2 (p-value = 0.007 &lt; 0.05). However, like the DNA yield, it is expected that the MNC concentration would increase with the increased volume of MNC layer collected as seen in Batch 2.

Overall, the SRs showed no statistically significant differences between the 2.5, 3, and 3.5 ml collection volumes. Neither DNA yield nor MNC concentration had a significant impact on SR as proven by the similar SRs across the 3 MNC layer collection volumes.

e. DNA Input

The purpose of this study was to provide evidence that demonstrated equivalency when using DNA inputs at 10±3 ng/μL in the assay. Extracted DNA replicates from the Limit of Detection and Dynamic Range study with contrived samples were used by testing only the lowest allelic ratio sample panel members. DNA samples, listed below, were diluted to 7, 10, and 13 ng/μL and tested with the assay along with a single replicate of Negative Control.

- AR 0.03 30 bp ITD (33 replicates at each DNA input level)
- AR 0.05 D835 TKD (33 replicates)
- AR 0.05 126 bp ITD (22 replicates)
- AR 0.33 279 bp ITD (11 replicates)

Acceptance criteria were met for 30 bp ITD, 126 bp ITD, and D835 cells: 1) &gt;93.9% of replicates met sample validity criteria for every sample type and DNA input; 2) overall coefficient of variation (CV) was &lt;20.5% for every sample type; and 3) CV was &lt;21.0% for every sample type when replicates were pooled between 7 and 10 ng/μL and between 13 and 10 ng/μL DNA input. Acceptance criteria were not met for TVI-ITD. While 100% of replicates met sample validity criteria, the overall CV and CV among pooled DNA inputs exceeded 25%.

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 22

{22}

The difference in mean mutant:wild type SRs among DNA inputs did not exceed 0.022, and the differences between means were not significantly different. The assay is able to provide consistent results when challenged with DNA inputs at $10 \pm 3 \mathrm{ng} / \mu \mathrm{L}$.

f. Assay Stopping Points in the FLT3 Mutation Assay

Stopping points experienced during execution of the FLT3 Mutation Assay were evaluated in this study. DNA and amplified samples were incubated at the temperatures and durations sufficient to challenge typical work-flow stopping points. Samples studied were the same as from the precision/reproducibility study. Thus, only contrived samples were investigated. At least $25\%$ extra time was added to the expected typical incubation times. Assay will allow three work flow stopping points:

1. After DNA isolation, samples can be held for up to 7 days at $2^{\circ}$ to $8^{\circ}\mathrm{C}$ before proceeding
2. After PCR, samples can be held for up to 72 hours at $2^{\circ}$ to $8^{\circ}\mathrm{C}$ before proceeding
3. After digestion step of TKD assay, samples can be held for up to 72 hours at $2^{\circ}$ to $8^{\circ}\mathrm{C}$ before proceeding

The sample validity rate was required to be $&gt;95\%$. The qualitative test results were required to have at least $95\%$ agreement with those obtained from the operator-to-operator precision study.

The following assay steps were performed: 1) samples were processed, mononuclear cells (MNCs) were isolated, and DNA was extracted from the isolated MNCs (if time permits, DNA can be quantified); 2) DNA was quantified (if not already done) and diluted to $10\mathrm{ng} / \mu \mathrm{L}$, and the DNA was amplified by PCR; 3) the tyrosine kinase domain (TKD) amplicons were digested with EcoRV restriction enzyme; and 4) the amplicons were analyzed by capillary electrophoresis (CE). The first step (sample processing through DNA extraction) must be completed in one day. While steps 2 to 4 above can be performed in one work shift, for workflow purposes pauses are allowed between steps 1 and 2, between steps 2 and 3, and between steps 3 and 4. All studies were conducted in duplicate.

Stopping Point 1: DNA Extraction

The DNA concentrations from the precision study were recorded and the samples were diluted to $10\mathrm{ng} / \mu \mathrm{l}$ based on those measurements. The concentrated and diluted DNA samples were stored at $2^{\circ}$ to $8^{\circ}\mathrm{C}$ for at least 9 days in duplicates for each of the 9 panel members. After 9 days, the concentrated DNA samples were diluted to $10\mathrm{ng} / \mu \mathrm{l}$, and the assay (amplification through CE) was completed on all DNA samples.

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 23

{23}

Stopping Point 2: Amplification, and Holding Point: CE Preparation (ITD)

An amplified ITD PCR plate from a precision study run was obtained, and the plate was stored at 2° to 8°C for at least 90 hours. After the incubation period, the plate was prepared for CE analysis. The CE plate was stored on a 3500xL Genetic Analyzer for at least 15 hours and analyzed at the end of incubation.

This Stopping Point 2 is also referred to as the ITD Stopping Point, or ISP.

Stopping Points 2 and 3: Amplification and EcoRV Digestion, and Holding Point: CE Preparation (TKD)

An amplified TKD PCR plate from a precision study run was obtained, and the plate was stored at 2° to 8°C for at least 90 hours. After the incubation period, the plate was digested with EcoRV. The digested plate was stored at 2° to 8°C for at least 90 hours.

After the incubation period, the plate was prepared for CE analysis. The CE plate was stored on a 3500xL Genetic Analyzer for at least 15 hours and analyzed at the end of incubation. This Stopping Point, 2 and 3, is also referred to as the TKD Stopping Point, or TSP.

All Stopping Points and Holding Points

The ITD and TKD PCR plates from Stopping Point 1 were stored at 2° to 8°C for at least 90 hours. After the 90 hour incubation period, the ITD plate was prepared for CE analysis. The CE plate was stored on a 3500xL Genetic Analyzer for at least 15 hours and analyzed at the end of incubation. After the 90 hour incubation period, the TKD plate was digested with EcoRV. The digest plate was stored at 2° to 8°C for at least 90 hours. After the incubation period, the plate was prepared for CE analysis. The CE plate was stored on a 3500xL Genetic Analyzer for at least 15 hours and analyzed at the end of incubation.

The All Stopping Points and Holding Points are also referred to as the Cumulative Stopping Points, or CSP.

The qualitative test result (negative/positive) was determined for each sample based on the mutant:wild type signal ratio. A sample was called negative, i.e., mutation not detected because mutant:wild type signal ratio did not exceed the cutoff of 0.05, or positive, i.e., mutation detected because mutant:wild type signal ratio met or exceeded the cutoff of 0.05.

Results for each mutation type and stopping point are summarized in Table 21. All study acceptance criteria were met. Results show that correct calls can be made when incubation times are varied for different potential stopping points during the assay.

8. Stability Studies

a. Assay Kit Stability (shelf-life)

The purpose of this study was to provide objective evidence that the reagent stability requirements of the LeukoStrat® CDx FLT3 Mutation Assay would be

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 24

{24}

met. Each kit lot was functionally tested at time points 0, 2, 3 and 4 months using 2 replicates of the 30 bp ITD and D835 TKD DNA positive control, no template control, extraction control or cell based positive controls in the ITD and TKD portions of the assay. Acceptance criteria were that kits had to yield $\geq 9.5\mathrm{ng} / \mu \mathrm{l}$ upon extraction and meet validity criteria. The study demonstrates that kits are stable for at least 3 months from the date of manufacture when components are stored at the recommended temperatures $(-30^{\circ}\mathrm{C}$ to $-15^{\circ}\mathrm{C}, 2^{\circ}\mathrm{C}$ to $8^{\circ}\mathrm{C},$ or $15^{\circ}\mathrm{C}$ to $30^{\circ}\mathrm{C}.$

## b. Open Kit Stability

To provide proof that all reagents are stable for at least one week after being opened, used and stored per the LeukoStrat® CDx FLT3 Mutation Assay instructions, two ITD Positive Cell Based Controls (ITD CBC---30 bp ITD) and two TKD Positive Cell Based Controls (TKD CBC) were tested in the FLT3 Mutation Assay twice over the span of a week. Three kit lots were tested on Day 1 and Day 8 using kits that were 4 months old. ITD and CBC SR had to be $&gt;0.05$ in the ITD reaction and $&lt;0.05$ for the TKD reaction. TKD CBC SR had to be $&gt;0.05$ in the TKD reaction and $&lt;0.05$ in the ITD reaction.

The ITD CBC SRs and TKD CBC SRs were independent of the time point, Day 1 or Day 8, and size standard solution, verifying that all reagents are stable when opened and used multiple times over eight days.

## c. Capillary Stability

The purpose of this study was to provide objective evidence that the stability requirements of the ABI3500xL Capillary Array would be met. Capillaries' abilities to appropriately separate products were evaluated using sizing accuracy of Map Marker and DS-33 GeneScan Installation Standard fragments; signal sensitivity was evaluated using DS-33 GeneScan Installation Standard and Size Standard fragments. Each capillary array lot was functionally tested at time points 0, 300, 400, and 500 injections to demonstrate performance of the capillary arrays are stable for a maximum of 400 injections.

The report provides objective evidence that three capillary array lots have successfully completed all 4 times points, indicating they are stable for at least 400 injections.

## d. Reagent Stability

i. Master Mixes and Controls

The purpose of this study was to provide evidence that the reagent stability requirements of the LeukoStrat® CDx FLT3 Mutation Assay were met. Invivoscribe manufactured reagents must be stable for at least 6 months when stored under recommended conditions. Three kit lots were tested with each of the assay controls [positive control (PC), extraction control (EC), no template control (NTC)] at 0, 6, and 8 months. The 6 month time point showed $100\%$ agreement with the baseline values for the 54 replicates, for each control (162

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 25

{25}

total replicates) while the 8 month time point had 53/54 (98%) PC replicates and 54/54 replicates of both EC and NTC. Each time point was comprised of 6 runs, 3 ITD and 3 TKD. Each lot of Master Mix was tested with all 3 lots of Controls – 3 vials per Control, 2 replicates per vial. Testing was divided among three operators. The 8 month time point had 53/54 (98%) PC replicates and 54/54 replicates of both EC and NTC meeting the predefined material acceptance criteria of 99% for all controls).

ii. EcoRV, Buffer 3 and BSA Stability

The performance of EcoRV Enzyme, Buffer 3 and 100 µg/mL BSA was evaluated at three time points of 0 months, 6 months, and 8 months. Each reagent was tested together as a set at each time point. Each set consisted of 3 vials with three replicates of each vial that were averaged together to produce an Average Enzyme Activity for the vial. All replicates were above the specification of Average Enzyme Activity per triplicate ≥ 2.55 Units/mL..

This study provides evidence that the EcoRV Enzyme, Buffer 3 and 100 µg/mL of BSA have a shelf life of six months.

iii. Formamide &amp; Dye size standard Stability

Amplicons generated using the FLT3 Mutation Assay must be denatured into single stranded products and sized using ABI3500xL instrument. Highly deionized (HIDI) formamide is used to keep the products single stranded. Dye size standard allows the products to be accurately sized by the instrument. Because HIDI formamide’s and size standard’s stability has a direct impact on the FLT3 Mutation Assay’s performance, this stability study was intended to support HIDI formamide’s and size standard’s stability. This study was carried out using three lots of HIDI formamide and three lots of size standard with only the TKD injections mixed with Map Marker and DS-33 GeneScan Installation Standard for denaturation.

For the time point to be considered acceptable, 3 lot pairs of HIDI formamide and size standard had to meet the validity and acceptance criteria for the expected size fragment and a minimum peak size of 100 RFU for the 20 bp fragment or &gt;250 RFU for all other fragments for each time point (0, 6, and 8 months). This study provided evidence that HIDI formamide and size standards are stable up to at least 6 months by yielding the correct fragment sizes and peak heights.

iv. Taq Polymerase Stability

The purpose of this study was to provide objective evidence that the LeukoStrat® CDx FLT3 LeukoStrat® CDx FLT3 Mutation Assay reagents: Taq Polymerase, Master Mix, Positive Control, Extraction Control and No Template Control are stable for at least 6 months when stored under recommended conditions.

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 26

{26}

Three combinations were constructed and tested at each time point:
- Combination 1: MM1-T1-PC1-EC1-NTC1
- Combination 2: MM2-T2-PC2-EC2-NTC2
- Combination 3: MM3-T3-PC3-EC3-NTC3

For each combination, 6 replicates of each Control were made from a mix of each Master Mix/Taq lot.

Each lot of controls tested has met stability requirements for the LeukoStrat® CDx FLT3 LeukoStrat® CDx FLT3 Mutation Assay over 8 months. Therefore, reagents manufactured at Invivoscribe were stable for 6 months when stored under recommended conditions. The baseline had 52/54 (96%) ITD and 54/54 (100%) TKD replicates that met the acceptance criteria. The 6 and 8 month time points had 100% (54/54) of all control ITD and TKD replicates that met the acceptance criteria.

All time points tested to date following 0 months/baseline, were completed within the 14 day window allowed by the protocol. This study covers time points 0 (baseline), 6, and 8 months.

As of the 8 month time point, the study meets the acceptance criteria of 6 months of stability according to assay requirements with only two invalid results across all time points.

When stored under recommended conditions, the combination of LeukoStrat® CDx FLT3 Mutation Assay reagents (Taq Polymerase, Master Mix, Positive Control, Extraction Control and No Template Control) is stable for 6 months.

e. Specimen Stability

The objective of this study was to evaluate the stability of BM aspirate and PB specimens collected from patients with acute myeloid leukemia (AML). This study presents the FLT3 Mutation Assay results of 52 BM and 100 PB clinical specimens. Specimens were processed to DNA upon receipt (up to 4 days after collection) then stored at 2° to 8°C and processed again at Day 5 and Day 7 after collection. If volume permitted, specimens were processed again on Day 9. DNA isolated from the specimens was stored at -15° to -30°C and tested up to 35 months later once design lock had been achieved and all samples were accrued.

For BM, NPA from Baseline to Day 7 for FLT3 mutation (ITD and TKD mutations combined) was 100% (21/21), and PPA was 100% (30/30). NPA was required to be at least 95% and PPA was required to be at least 90%; therefore, the acceptance criteria for BM were met for this study. For PB, NPA from Baseline to Day 7 for FLT3 mutation (ITD and TKD mutations combined) was 97.9% (47/48), and PPA was 98.0% (50/51). NPA was required to be at least 95% and PPA was required to be at least 90%; therefore, the acceptance criteria for PB were

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 27

{27}

met for this study. The Baseline to Day 9 FLT3 NPA and PPA for BM was 100% (17/17) and 100% (24/24), respectively. The baseline to Day 9 FLT3 NPA and PPA for PB was 96.6% (28/29) and 100% (43/43), respectively.

This study supports the claim that specimens will be analyzable up to 5 days following date of collection if refrigerated at between 2°C to 8°C.

## B. Animal Studies

Not Applicable

## X. SUMMARY OF PRIMARY CLINICAL STUDY(IES)

The safety and effectiveness of the LeukoStrat® CDx FLT3 Mutation Assay were demonstrated in a retrospective analysis of specimens from patients screened for enrollment into the RATIFY clinical trial. The RATIFY (Randomized AML Trial In FLT3 in &lt;60 Year Olds, also known as CPKC412A2301 (or A2301) clinical trial was a Phase 3, randomized, double-blind placebo-controlled trial of 717 patients with newly-diagnosed FLT3-mutated AML to assess the safety and efficacy of RYDAPT (midostaurin) in combination with chemotherapy. A bridging study was conducted to assess clinical agreement between samples with FLT3 status tested with the clinical trial assay (CTA) and the LeukoStrat® CDx FLT3 Mutation Assay in the intent-to-test population. The primary objectives of the study were to 1) establish agreement with respect to selection of FLT3 mutant patients between the CTA and the LeukoStrat® CDx FLT3 Mutation Assay by assessing the overall, positive, and negative percent agreement between the two assays and 2) to estimate midostaurin efficacy in the LeukoStrat® CDx FLT3 Mutation Assay positive population on both overall survival (OS)

## A. Study Design

The RATIFY clinical trial was a multi-site, international, Phase 3, randomized, double-blind trial assessing the use of midostaurin in combination with standard remission induction and consolidation chemotherapy (n = 345), compared to placebo with chemotherapy (n = 335) for treatment of patients with newly-diagnosed FLT3-mutant AML. The original trial design allowed for initial screening for FLT3 mutation in either peripheral blood (PB) or bone marrow (BM) sample of patients with newly-diagnosed AML. Patients were stratified at randomization based on FLT3-ITD allelic ratio &lt; 0.7, FLT3-ITD allelic ratio ≥ 0.7, and FLT3-TKD mutations. The primary endpoint was overall survival.

### 1. Clinical Inclusion and Exclusion Criteria

Enrollment in the RATIFY study was limited to patients who met the following inclusion criteria

- Unequivocal diagnosis of AML (&gt;20% blasts in the bone marrow) excluding acute promyelocytic leukemia without evidence of central nervous system involvement.

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 28

{28}

- Documented FLT3 mutation (ITD or TKD), determined by analysis in a protocol-designated FLT3 screening laboratory.
- Age of subject ≥18 and &lt;60 years.

Patients were not permitted to enroll in the RATIFY study if they met any of the following exclusion criteria:

- Prior chemotherapy for leukemia or myelodysplasia with the following exceptions:
- emergency leukapheresis
- emergency treatment for hyperleukocytosis with hydroxyurea for ≤ 5 days,
- cranial radiation therapy for CNS leukostasis (one dose only),
- growth factor/cytokine support.

- Patients with a history of antecedent myelodysplasia (MDS) who had prior cytotoxic therapy (including azacitidine or decitabine) for MDS.
- Patients who have developed therapy-related AML after prior RT or chemotherapy for another disorder or cancer
- Patients with symptomatic congestive heart failure.
- Bilirubin ≥2.5 x upper limit of normal.
- Pregnant or nursing.

Additional Specimen Inclusion/Exclusion Criteria for the Bridging Study Inclusion Criteria

- DNA isolated from CTA(+) and CTA(-) subjects screened for the RATIFY trial.
- DNA isolated from MNCs frozen in a method shown to adequately preserve the specimen.
- DNA of a quantity of at least 50ng at ≥10ng/μL (defined as ≥9.5ng/μL in the assay IFU).

Exclusion Criteria

- Sample delivered without legible unique identification number.
- Obvious physical damage to sample tube.
- DNA received thawed.

2. Follow-up Schedule

Disease assessment and other clinical assessments were conducted according to schedule. The primary analysis was conducted after a minimum follow-up of approximately 3.5 years after the randomization of the last patient. Efficacy was also assessed by event-free survival (EFS) where an event was defined as a failure to obtain a complete remission (CR) within 60 days of initiation of protocol therapy, or relapse, or death from any cause.

3. Clinical Endpoint

The primary objective of the A2301 study was to determine if the addition of midostaurin to daunorubicin/cytarabine induction, high-dose cytarabine

PMA P160040: FDA Summary of Safety and Effectiveness Data
Page 29

{29}

consolidation, and continuation therapy improved OS in both mutant FLT3-ITD or FLT3-TKD AML patients.

# B. Accountability of PMA Cohort

A total of 3279 patients were screened for the presence of FLT3 mutations by the clinical trial assay (CTA) and enrollment criteria. Of those, 717 patients with FLT3 mutations were enrolled and randomized to receive treatment with chemotherapy plus midostaurin or placebo. Only patients with FLT3 mutations and meeting all trial inclusion criteria were enrolled. The bridging study included the ascertained samples from the RATIFY (A2301) patients who had provided informed consent for LeukStrat CDx FLT3 Mutation Assay (CDx) development. These samples were from the CTA+ patients enrolled in the A2301 study as well as the non-enrolled screen failures. Banked A2301 samples from 520 CTA-positive patients and 578 randomly selected CTA-negative patients were tested by the CDx test. The CTA and the CDx results were used to evaluate the concordance (agreement) of the assays as well as to estimate the midostaurin efficacy in the CDx-positive population. Reasons for unavailability of specimens are shown in Figure 3 below. (Figure 3 also refers to selection of specimens for the method comparison (accuracy/correlation to another method study described above).

Figure 3. AML Specimen Accountability for CDx Bridging Study. For the bridging study, 578 CTA- patients were selected but only 555 CTA- specimens were evaluable by CDx analysis testing due to low sample. For the $\mathrm{CTA + }$ specimens, 498 were evaluable by CDx analysis.

PMA P160040: FDA Summary of Safety and Effectiveness Data

{30}

# C. Study Population Demographics and Baseline Parameters

In the RATIFY trial, 3279 patients were screened to enroll 717 who were positive for FLT3 mutations based on the CTA test. The two treatment groups were generally balanced with respect to the baseline demographics (Tables 23). The median age of study participants was 47 years on the midostaurin arm and 48 years on the placebo arm, with a range of 18-59 years. Race was unknown or not reported in $57\%$ of subjects, but $89\%$ of the subjects for whom race was reported were white.

Distribution of FLT3 based on AR stratification factor was the same across both arms. The distribution of several characteristics for those concordant cases used in the bridging study and by enrollment site and testing site is shown in the Table 16.

Table 16. Comparison of baseline characteristics and sample covariates between the two treatment arms in the (CTA+, CDx+) patients

| Baseline Characteristics (CTA+) | | MIDOSTAURIN N=252 | PLACEBO N=237 | ALL N=489 | p-value (1) |
| --- | --- | --- | --- | --- | --- |
| Region - n (%) | | | | | 0.999 |
| North America | | 101 (40.1) | 95 (40.1) | 196 (40.1) | |
| Non North America | | 151 (59.9) | 142 (59.9) | 293 (59.9) | |
| Country - n | | | | | 0.964 |
| Austria | | 4 (1.6) | 3 (1.3) | 7 (1.4) | |
| Belgium | | 3 (1.2) | 4 (1.7) | 7 (1.4) | |
| Canada | | 2 (0.8) | 5 (2.1) | 7 (1.4) | |
| Czech Republic | | 2 (0.8) | 3 (1.3) | 5 (1.0) | |
| France | | 3 (1.2) | 2 (0.8) | 5 (1.0) | |
| Germany | | 112 (44.4) | 105 (44.3) | 217 (44.4) | |
| Italy | | 18 (7.1) | 15 (6.3) | 33 (6.7) | |
| Netherlands | | 3 (1.2) | 2 (0.8) | 5 (1.0) | |
| Slovakia | | 2 (0.8) | 1 (0.4) | 3 (0.6) | |
| Spain | | 4 (1.6) | 7 (3.0) | 11 (2.2) | |
| United States | | 99 (39.3) | 90 (38.0) | 189 (38.7) | |
| CTA Laboratory | | | | | 0.957 |
| OS | | 101 (40.1) | 95 (40.1) | 196 (40.1) | |
| HA | | 13 (5.2) | 12 (5.1) | 25 (5.1) | |
| UL | | 69 (27.4) | 71 (30.0) | 140 (28.6) | |
| DR | | 36 (14.3) | 28 (11.8) | 64 (13.1) | |
| UR | | 18 (7.1) | 15 (6.3) | 33 (6.7) | |
| NI | | 11 (4.4) | 9 (3.8) | 20 (4.1) | |
| BA | | 2 (0.8) | 3 (1.3) | 5 (1.0) | |
| VA | | 2 (0.8) | 4 (1.7) | 6 (1.2) | |
| Sample Material Banked by the laboratory | | | | | 0.751 |
| | DNA | 11 (4.4) | 9 (3.8) | 20 (4.1) | |
| | MNC | 241 (95.6) | 228 (96.2) | 469 (95.9) | |
| MNC count is greater than 5x10^5 cells | | | | | 0.929 |
| Yes | | 57 (22.6) | 52 (21.9) | 109 (22.3) | |
| No | | 184 (73.0) | 176 (74.3) | 360 (73.6) | |
| Missing | | 11 (4.4) | 9 (3.8) | 20 (4.1) | |
| FLT3 mutation status - n (%) (3) | | | | | 0.539 |
| TKD | | 59 (23.4) | 50 (21.1) | 109 (22.3) | |
| ITD (includes patients with both TKD and ITD) | | 193 (76.6) | 187 (78.9) | 380 (77.7) | |
| | ITD Signal ratio <0.7 | 111 (44.0) | 112 (47.3) | 223 (45.6) | |
| | ITD Signal ratio >=0.7 | 82 (32.5) | 75 (31.6) | 157 (32.1) | |

PMA P160040: FDA Summary of Safety and Effectiveness Data

{31}

D. Safety and Effectiveness Results

1. Safety Results

The safety with respect to treatment with RYDAPT (midostaurin) will not be addressed in detail in the SSE. Briefly, the most frequent (incidence greater than or equal to 20%) adverse drug reactions (ADRs) in the RYDAPT plus chemotherapy arm were febrile neutropenia, nausea, vomiting, mucositis, headache, musculoskeletal pain, petechiae, device-related infection, epistaxis, hyperglycemia and upper respiratory tract infections. Discontinuation due to any adverse reaction occurred in 9% of patients in the RYDAPT arm versus 6% in the placebo arm. The most frequent Grade 3/4 adverse reactions leading to discontinuation in the RYDAPT arm was renal insufficiency (1%). Excluding deaths due to disease progression, no fatal adverse reactions occurred in the study. Refer to the drug label for additional information.

The LeukoStrat™ CDx FLT3 Mutation Assay involves testing on PB and BM specimens. These specimens are routinely removed as part of the practice of medicine for the diagnosis of AML and therefore present no additional safety hazard to the patient being tested. See the Safety Conclusions Section below for further discussion.

2. Effectiveness Results

a. Efficacy based on CTA:

Efficacy was established on the basis of overall survival (OS), measured from the date of randomization until death by any cause. The primary analysis was conducted after a minimum follow-up of approximately 3.5 years after the randomization of the last patient. RYDAPT plus st…