← Product Code PQP · P240010 # MI Cancer Seek (P240010) _Caris Life Sciences · PQP · Nov 5, 2024 · Medical Genetics · APPR_ **Canonical URL:** https://fda.innolitics.com/device/P240010 ## Device Facts - **Applicant:** Caris Life Sciences - **Product Code:** PQP - **Decision Date:** Nov 5, 2024 - **Decision:** APPR - **Device Class:** Class 3 - **Review Panel:** Medical Genetics ## Intended Use MI Cancer Seek is a next-generation sequencing (NGS) based in vitro diagnostic (IVD) device using total nucleic acid (TNA) isolated from formalin-fixed paraffin-embedded (FFPE) tumor tissue specimens for the detection of single nucleotide variants (SNVs) and insertions and deletions (indels) in 228 genes, microsatellite instability (MSI), tumor mutational burden (TMB) in patients with previously diagnosed solid tumors, and copy number amplification (CNA) in one gene in patients with breast cancer. MI Cancer Seek is intended as a companion diagnostic to identify patients who may benefit from treatment with the targeted therapies listed in Table 1 below, in accordance with the approved therapeutic product labeling. Additionally, MI Cancer Seek is intended to provide tumor mutational profiling to be used by qualified healthcare professionals in accordance with professional oncology guidelines for cancer patients with previously diagnosed solid malignant neoplasms. Genomic findings other than those listed in Table 1 are not prescriptive or conclusive for labeled use of any specific therapeutic product. ## Device Story MI Cancer Seek is a single-site NGS assay performed at Caris Life Sciences. It processes TNA extracted from FFPE tumor tissue. The workflow includes automated extraction, library preparation, and sequencing on the Illumina NovaSeq 6000. The bioinformatics pipeline aligns reads to the hg38 reference genome to detect SNVs, indels, MSI, TMB, and ERBB2 amplifications. The device is operated by laboratory personnel at the Caris facility. Results are provided in a report categorized by clinical significance (Level 1 CDx, Level 2/3 potential significance). Healthcare professionals use these findings to guide treatment decisions for patients with solid tumors, matching genomic alterations to FDA-approved targeted therapies. The device benefits patients by identifying potential therapeutic options based on their tumor's molecular profile. ## Clinical Evidence Clinical evidence consists of eight non-inferiority concordance studies comparing MI Cancer Seek to FDA-approved comparator companion diagnostics (CCDs) using remnant FFPE samples. Studies evaluated MSI (n=401), BRAF V600E/K (n=334), BRAF V600E in CRC (n=358), KRAS/NRAS (n=286), EGFR (n=328), and PIK3CA (n=356). All studies met pre-defined non-inferiority margins for PPA and NPA. Analytical validation included 16 studies covering accuracy, LoD, precision, and stability. ## Technological Characteristics NGS-based IVD using custom whole exome sequencing panel. Analyzes TNA from FFPE tissue. Employs hybrid capture with 120-nucleotide biotinylated baits. Sequencing performed on Illumina NovaSeq 6000. Bioinformatics pipeline uses Sentieon BWA for alignment and custom software for variant calling (SNVs, indels, MSI, TMB, ERBB2 CNA). Single-site assay performed at Caris Life Sciences. ## Regulatory Identification A next generation sequencing (NGS) oncology panel is a device used for the qualitative detection of germline or somatic variants in one or more cancer-related genes. The device is intended to be used on DNA or RNA isolated from human clinical specimens. ## Reference Devices - FoundationOne CDx - Ventana MMR RxDx Panel - therascreen BRAF V600E RGQ PCR Kit - cobas 4800 BRAF V600 Mutation Test - THXID BRAF Kit - cobas EGFR Mutation Test v2 - ONCO/Reveal Dx Lung & Colon Cancer Assay (O/RDx-LCCA) - Oncomine Dx Target Test - therascreen EGFR RGQ PCR Kit - Praxis Extended RAS Panel - therascreen PIK3CA RGQ PCR Kit ## Submission Summary (Full Text) > This content was OCRed from public FDA records by [Innolitics](https://innolitics.com). If you use, quote, summarize, crawl, or train on this content, cite Innolitics at https://innolitics.com. > > Innolitics is a medical-device software consultancy. We help companies design, build, and clear FDA-regulated software and AI/ML devices, including [a PMA](https://innolitics.com/services/regulatory/), [a 510(k)](https://innolitics.com/services/510ks/), [a SaMD](https://innolitics.com/services/end-to-end-samd/), [an AI/ML medical device](https://innolitics.com/services/medical-imaging-ai-development/), or [an FDA regulatory strategy](https://innolitics.com/services/regulatory/). {0} # SUMMARY OF SAFETY AND EFFECTIVENESS DATA (SSED) ## I. GENERAL INFORMATION Device Generic Name: Next generation sequencing oncology panel, somatic or germline variant detection system Device Trade Name: MI Cancer Seek Device Procode: PQP Applicant’s Name and Address: Caris Life Sciences 4610 South 44th Place Phoenix, AZ 85040 Date(s) of Panel Recommendation: None Premarket Approval Application (PMA) Number: P240010 Date of FDA Notice of Approval: November 5, 2024 ## II. INDICATIONS FOR USE MI Cancer Seek is a next-generation sequencing (NGS) based in vitro diagnostic (IVD) device using total nucleic acid (TNA) isolated from formalin-fixed paraffin-embedded (FFPE) tumor tissue specimens for the detection of single nucleotide variants (SNVs) and insertions and deletions (indels) in 228 genes, microsatellite instability (MSI), tumor mutational burden (TMB) in patients with previously diagnosed solid tumors, and copy number amplification (CNA) in one gene in patients with breast cancer. MI Cancer Seek is intended as a companion diagnostic to identify patients who may benefit from treatment with the targeted therapies listed in Table 1 below, in accordance with the approved therapeutic product labeling. Additionally, MI Cancer Seek is intended to provide tumor mutational profiling to be used by qualified healthcare professionals in accordance with professional oncology guidelines for cancer patients with previously diagnosed solid malignant neoplasms. Genomic findings other than those listed in Table 1 are not prescriptive or conclusive for labeled use of any specific therapeutic product. PMA P240010: FDA Summary of Safety and Effectiveness Data {1} Table 1. MI Cancer Seek Companion Diagnostic Indications | Indication | Biomarker | Therapy | | --- | --- | --- | | Breast Cancer | PIK3CA (C420R; E542K; E545A, E545D [1635G>T only], E545G, E545K, Q546E, Q546R; and H1047L, H1047R, H1047Y) | PIQRAY® (alpelisib) | | Colorectal Cancer (CRC) | KRAS wild-type (absence of mutations in exons 2, 3, and 4) and NRAS wild type (absence of mutations in exons 2, 3, and 4) | VECTIBIX® (panitumumab) | | | BRAF V600E | BRAFTOVI® (encorafenib) in combination with ERBITUX® (cetuximab) | | Melanoma | BRAF V600E | BRAF Inhibitors approved by FDA* | | | BRAF V600E or V600K | MEKINIST® (trametinib) or BRAF/MEK Inhibitor Combinations approved by FDA* | | Non-small cell lung cancer (NSCLC) | EGFR exon 19 deletions and exon 21 L858R alterations | EGFR Tyrosine Kinase Inhibitors approved by FDA* | | Solid Tumors | MSI-H | KEYTRUDA® (pembrolizumab), JEMPERLI (dostarlimab-gxly) | | Endometrial Carcinoma | Not MSI-H | KEYTRUDA® (pembrolizumab) in combination with LENVIMA® (lenvatinib) | | *For the most current information about the device indications for the therapeutic products in this group, go to: https://www.fda.gov/medical-devices/in-vitro-diagnostics/list-cleared-or-approved-companion-diagnostic-devices-in-vitro-and-imaging-tools#Group_Labeling | | | MI Cancer Seek is a single-site assay performed at Caris Life Sciences, Phoenix, AZ. ## III. CONTRAINDICATIONS There are no known contraindications. ## IV. WARNINGS AND PRECAUTIONS The warnings and precautions can be found in the MI Cancer Seek labeling. ## V. DEVICE DESCRIPTION MI Cancer Seek is a single-site assay performed at Caris Life Sciences located at 4610 South 44th Place, Phoenix, AZ 85040. The test includes reagents, software, and procedures for PMA P240010: FDA Summary of Safety and Effectiveness Data {2} testing of total nucleic acid (TNA) from formalin-fixed paraffin-embedded (FFPE) tumor tissue. The test employs a custom whole exome sequencing panel to detect and report SNVs and indels within 228 genes outlined in Table 2 across solid tumors and amplifications in ERBB2 in patients with breast cancer only. The test also detects MSI determined from 3,210 genes and whole exome based TMB in patients with solid tumors. Table 2. MI Cancer Seek Reportable Gene List for SNVs and indels | ABL1 | BRCA1 | CYLD | FGFR1 | JAK2 | MPL | PDGFRA | RAF1 | STAG2 | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | ACVR1 | BRCA2 | DDR2 | FGFR2 | JAK3 | MRE11 | PDGFRB | RASA1 | STAT3 | | AIP | BRIP1 | DICER1 | FGFR3 | KDM5C | MSH2 | PIK3CA | RB1 | STK11 | | AKT1 | BTK | DNMT3A | FGFR4 | KDM6A | MSH3 | PIK3CB | RET | SUFU | | AKT2 | CALR | EGFR | FH | KDR | MSH6 | PIK3R1 | RHOA | TCF7L2 | | AKT3 | CARD11 | EP300 | FLCN | KEAP1 | MTOR | PIK3R2 | RNF43 | TERT | | ALK | CBFB | EPHA2 | FLT1 | KIT | MUTYH | PIM1 | ROS1 | TET2 | | AMER1 | CCND1 | ERBB2 | FLT3 | KLF4 | MYC | PMS2 | RUNX1 | TMEM127 | | APC | CCND2 | ERBB3 | FOXA1 | KMT2A | MYCN | POLD1 | SDHA | TNFAIP3 | | AR | CCND3 | ERBB4 | FOXL2 | KMT2C | MYD88 | POLE | SDHAF2 | TNFRSF14 | | ARAF | CD79B | ERCC2 | FUBP1 | KMT2D | NBN | POT1 | SDHB | TP53 | | ARID1A | CDC73 | ESR1 | GATA3 | KRAS | NF1 | PPP2R1A | SDHC | TRAF7 | | ARID2 | CDH1 | EZH2 | GNA11 | LZTR1 | NF2 | PPP2R2A | SDHD | TSC1 | | ASXL1 | CDK12 | FANCA | GNA13 | MAP2K1 | NFE2L2 | PRDM1 | SETD2 | TSC2 | | ATM | CDK4 | FANCB | GNAQ | MAP2K2 | NFKBIA | PRKACA | SF3B1 | U2AF1 | | ATRX | CDKN1B | FANCC | GNAS | MAP2K4 | NOTCH1 | PRKAR1A | SMAD2 | VHL | | AXIN2 | CDKN2A | FANCD2 | H3F3A | MAP3K1 | NPM1 | PRKDC | SMAD4 | WRN | | B2M | CHEK1 | FANCE | H3F3B | MAPK1 | NRAS | PTCH1 | SMARCA4 | WT1 | | BAP1 | CHEK2 | FANCF | HIST1H3B | MAX | NSD1 | PTEN | SMARCB1 | XPO1 | | BARD1 | CIC | FANCG | HNF1A | MED12 | NSD2 | PTPN11 | SMARCE1 | XRCC1 | | BCL2 | CREBBP | FANCI | HOXB13 | MEF2B | NTHL1 | RAC1 | SMO | | | BCL9 | CSF1R | FANCL | HRAS | MEN1 | NTRK1 | RAD50 | SOCS1 | | | BCOR | CTCF | FANCM | IDH1 | MET | NTRK2 | RAD51B | SOS1 | | | BLM | CTNNA1 | FAS | IDH2 | MITF | NTRK3 | RAD51C | SPEN | | | BMPR1A | CTNNB1 | FAT1 | IRF4 | MLH1 | PALB2 | RAD51D | SPOP | | | BRAF | CXCR4 | FBXW7 | JAK1 | MLH3 | PBRM1 | RAD54L | SRC | | # A. Test Output The test report includes variants reported in the following levels: Level 1: Companion Diagnostic (CDx) Claims noted in Table 1 of the Intended Use Level 2: Cancer Mutations with Evidence of Clinical Significance Level 3: Cancer Mutations with Potential Clinical Significance PMA P240010: FDA Summary of Safety and Effectiveness Data {3} Genomic findings other than those listed in Table 1 of the Companion diagnostic indications table of the intended use statement (i.e., Levels 2 and 3) are not prescriptive or conclusive for labeled use of any specific therapeutic product. ## B. Test Kit Contents The MI Cancer Seek assay consists of five sub-kits that include the critical reagents required to execute the assay workflow: 1. MI Cancer Seek Extraction Kit 2. MI Cancer Seek Quantification Kit 3. MI Cancer Seek Library Prep Kit 4. MI Cancer Seek Sequencing Kit 5. MI Cancer Seek Library Quant Kit All reagents needed to perform the MI Cancer Seek assay are used exclusively at the Caris facility, as this is a single site assay. ## C. Instruments MI Cancer Seek is intended to be performed with serial numbered controlled instruments (Table 3) that are qualified and maintained by Caris’ Quality System. All instruments needed to perform the MI Cancer Seek assay are used exclusively at the Caris facility, as this is a single site assay. Table 2. Overview of MI Cancer Seek Instruments | Instrument | | --- | | Beckman Biomek i7 Automated Workstation | | Promega GloMax® Explorer Microplate Reader | | BioMicroLab Inc BioMicroLab XL 100 | | Agilent Bravo NGS workstation Option B | | Illumina NovaSeq 6000 Sequencing System | | Hamilton LabElite ID Capper | | Life Technologies Fleet Control Software Thermocycler Server | | Life Technologies Qubit 4.0 Fluorometer | | Life Technologies Veriti (Pro) 96 Well Thermal Cycler | | Thermo Fisher Scientific VisionMate Sample Barcode Scanner | | Thermo Fisher Scientific KingFisher Flex | ## D. Principles of Operation The MI Cancer Seek involves the process described below: a. Specimen Requirements, Collection and Preparation for Analysis PMA P240010: FDA Summary of Safety and Effectiveness Data 4 of 85 {4} The MI Cancer Seek requires TNA isolated from FFPE tissue specimens. Formalin-fixed, paraffin-embedded (FFPE) tumor tissue specimens are collected and prepared following standard pathology practice. FFPE specimens may be received as either unstained slides or FFPE tissue blocks. Prior to MI Cancer Seek testing, a Hematoxylin and Eosin (H&E) stained slide is prepared and reviewed by a board-certified pathologist to confirm that invasive cancer is present and ensure there is adequate tissue and tumor content to proceed with the assay. A minimum tissue area of 25 mm² with ≥20% tumor content is required for MI Cancer Seek. When necessary, Caris will perform manual microdissection of specimens to increase cell density for testing and avoid interferents as much as possible. H&E slides will be annotated for manual microdissection and microdissection will be performed to enrich tumor content and/or avoid areas with potential interferents such as necrotic tissue, fat cells, or melanin. 120 mm² tissue area is recommended for dissection as an optimal tissue input into TNA extraction. ## b. Nucleic Acid Extraction Deparaffinization and protease digestion of the dissected tissue is performed, followed by automated TNA extraction using the MI Cancer Seek Extraction Kit with Biomek i7 Automated Liquid Handling Workstation and KingFisher Flex automated extraction instrument. Tissue specimens are scraped into a tube containing lysis buffer and proceed to TNA extraction. The extraction-negative control (EXT NEG), extraction-positive control (HBCTL) and clinical samples are placed onto the Biomek i7 Automated Liquid Handling Workstation for addition of Proteinase K, decrosslinking and lysis incubations. The 96-well plate is placed onto the KingFisher Flex automated extraction instrument to perform the TNA extraction and elution process with the MI Cancer Seek Extraction Kit. After completion of TNA extraction, double-stranded DNA (dsDNA) is quantified using the MI Cancer Seek Quantification Kit and the GloMax Plate Reader. The sample must yield a minimum of 50 ng of DNA to continue workflow (1.43 ng/μL QC Check), while NTC should not exceed 0.5 ng/μL. The optimal or maximum DNA input for the test is 220 ng. ## c. Library Preparation and Enrichment Libraries are prepared through a series of automated processes using the MI Cancer Library Preparation Kit, Bravo liquid handlers, thermocyclers, and GloMax plate readers. The process begins with enzymatic fragmentation of the extracted TNA, followed by cDNA synthesis. cDNA molecules are chemically labelled to later allow for mutation origin differentiation (RNA or DNA-based) by the pipeline. After 3' adenylation, end-repair and ligation, libraries are amplified in a pre-capture polymerase chain reaction (PCR) followed by quantification using the MI Cancer Seek Quantification Kit and GloMax plate reader. Genomic targets are enriched through hybridization with a custom bait panel, covering the whole exome with boosted regions for clinically relevant genes. Baits are 120 nucleotides long, double- PMA P240010: FDA Summary of Safety and Effectiveness Data 5 of 85 {5} stranded and biotinylated. The enriched library is amplified by post-capture PCR and purified. Samples are normalized and pooled through an automated process on the Bravo liquid handlers using the MI Cancer Seek Quantification Kit. The resulting pooled library is quantified using the MI Cancer Seek Library Quantification Kit and is then manually denatured prior to use of the MI Cancer Seek Sequencing Kits for flow cell loading and sequencing on the Illumina NovaSeq 6000 platform. ## d. DNA Sequencing The pooled libraries are subject to sequencing-by-synthesis using the standard NovaSeq6000 workflow and the MI Cancer Seek Sequencing Kits. NovaSeq6000 systems will automatically load pooled libraries onto the flow cell and proceed for cluster generation and sequencing of the prepared library. Base calls are detected from the signal intensity measurements during each cycle. Base Call (BCL) files generated from the Illumina NovaSeq systems are stored on a secure High-Performance Computing (HPC) server on rapid solid-state hard drives. The initial sequencing metrics for the flow cells are reviewed by the NGS bioinformatics pipeline which includes the following run validity criteria, quality control (QC) check: Reads Passing Filter (M), Yield (G) and Q30 scores for each read 1 and read 4 (forward and reverse). The pipeline initiates automatic conversion and de-multiplex using the FPGA-adapted version of bcl2fastq via Illumina DRAGEN devices to convert the Illumina-generated BCL files to FASTQ file formats for downstream analysis. Demultiplexing is performed on the TNA reads, creating sample-specific FASTQ files. Custom software is applied to differentiate RNA and DNA molecules using the synthetic labels added during cDNA synthesis. Once separated, RNA data proceeds through the whole transcriptome pipeline and DNA data proceeds through the whole exome pipeline. DNA and RNA Seq data are aligned to the NCBI hg38reference genome (GRCh38). DNA variants detection (SNVs, indels, TMB, MSI and ERBB2 CNAs) are evaluated separately and the data are stored on secured servers. While the test has the capability to detect alterations in RNA via whole transcriptome sequencing (WTS), the test is only approved to report alterations using genomic DNA consistent with the test's intended use. ## e. Data Analysis and Reporting Sequencing results are analyzed by the MI Cancer Seek NGS Bioinformatic Pipeline, and reportable results passing run, control, and sample validity criteria are provided through Report Generation software. The pipeline performs an assessment of DNA-based NGS data as described below to generate the reportable calls for variants (SNV and indel), ERBB2 CNA, MSI, and TMB. The Report Generation software makes applicable CDx therapy associations based on pipeline results. PMA P240010: FDA Summary of Safety and Effectiveness Data 6 of 85 {6} The pipeline automatically initiates after the sequencer has finished generating the sequencing data in base call (BCL) format. Upon conclusion of copy of BCL data to a secure and designated directory, the pipeline commences data analysis with conversion of BCL files into demultiplexed sample-specific FASTQ files containing both DNA and RNA reads. These FASTQ files are split into DNA and RNA FASTQ file sets using the RNA labels to identify the RNA reads. Each FASTQ pair (one for DNA and RNA each) consists of two FASTQ files: each corresponding to forward and reverse reads due to paired-end sequencing. The DNA and RNA FASTQ files are then processed through WES and WTS workflows, respectively. Results from both WES and WTS workflows are required in order to assess the validity of the control samples and the run. The DNA FASTQ files are aligned with hg38 reference genome, using Sentieon BWA, creating a binary alignment map (BAM) where the duplicates are marked to be excluded from downstream steps. Next, as part of SNVs and indel identification, the BAM file is processed in a series of steps using several in-house and industry-standard tools. These steps include indel realignment and initial variant calling. Subsequently, high quality SNVs and indels are retained and annotated by removal of false positives, flagging low-quality variants and adding annotation including variant clinical impact. Next, ERRB2 CNA assessment is performed using CNV kit. This step uses the DNA BAM and VCF file as inputs and detects any ERBB2 amplifications. The CNV kit uses both the on-target reads and the nonspecifically captured off-target reads along with a segmentation algorithm to infer discrete copy number segments for an accurate amplification status of ERBB2. Subsequently, MSI status of a given sample is determined by calculating the number of observed frame shift mutations among 5,721 specific candidate loci present in the VCF file. Caris MI Cancer Seek CDx test uses tumor-only sequencing for MSI status determination. Finally, as the last step in WES workflow, TMB is calculated for a given sample as the number of variants per Mega bases (total number of variants divided by the size of target region in Mega bases) after filtering out non-coding, low-quality, low-variant frequency (VF) synonymous variants along with variants containing 'gnomAD' and 'Benign' tags. As the last step, the WES and WTS results are combined together and evaluated together to perform the following validity criteria checks: - Run validity on the sequencer output to classify the run as pass/fail - Run Controls QC to classify the run as pass or fail - Analyte Controls QC to determine which capabilities are reportable - DNA and RNA validity to pass/fail the individual samples - Variant calling criteria to determine which biomarkers are reportable for successful samples PMA P240010: FDA Summary of Safety and Effectiveness Data 7 of 85 {7} f. Assay Validity Criteria Validity criteria has been established at various levels and stages to evaluate each run, control sample, clinical sample and individual biomarker. Details are provided below for each type of validity criteria. Table 4 provides an overview of the MI Cancer Seek QC Process. i. Run validity criteria Run validity criteria are evaluated to determine if the sequencing run has passed. If a run fails its validity criteria, all clinical and control samples on that run are marked as fail. As part of this validity check, number of reads passing filter (PF) should be > 16 billion. This is calculated by the Illumina sequencers that perform an internal quality filtering procedure and reads that pass this filter are reported as PF reads. The run yield should be > 2,400 Gb and the percentage (%) of Q30 reads for both R1 and R4 should > 75%. A quality score of 30 (Q30) represents an error rate (for calling the base wrong) of 1 in 1000 (meaning every 1000 bp sequencing read may contain an error), with a corresponding call accuracy of 99.9%. When sequencing quality reaches Q30, virtually all of the reads will be perfect, with no errors or ambiguities. This is why Q30 is considered a benchmark for quality in next-generation sequencing (NGS). Finally, all three run control samples (NTC, HBCTL and POSCTL) need to pass for the run to pass. The validity criteria for these run controls is described in subsection ii. below. ii. Control validity criteria Control validity is used to determine if the control samples have passed or failed. There are three run controls that all have to pass for the run to pass. Validity criteria is established separately for each of them. In order to pass, run controls have to pass both WES and WTS validity criteria which are established separately. The first run control sample, No Template Control (NTC), does not contain any template DNA and is usually replaced with an equal amount of nuclease-free water and hence should ideally receive very few reads. For NTC to pass, the average read depth in WES should be <1x. Also, ratio of mapped reads to NTC and median of reads mapped to clinical samples within that flowcell should be < 5%. For POSCTL, the second run control, average depth over the boosted panel should be at least 100x in WES and it should receive at least 20 million DNA reads to ensure that the biomarkers are called confidently. In addition, the variant allele frequency (VAF) of several reference biomarkers specific to this POSCTL should be in their allowable range of the pipeline. For HBCTL, sample average depth should be at least 100x in WES and it should receive at least 20 million reads. In addition, the variant allele frequency (VAF) of several control biomarkers should be in their intended range. Additionally, for a PMA P240010: FDA Summary of Safety and Effectiveness Data 8 of 85 {8} given set of reference genes, HBCTL should not report any pathogenic variants in them. In addition to these run controls, an analyte control sample is also used (HBCNA), failure of which results in an invalid TMB, MSI and amplification results. For HBCNA to pass, the average WES depth should be $>100\mathrm{x}$ . Also, MSI, TMB and ERBB2 copy values for HBCNA should be in a certain range. In addition, it should also not report amplification in any genes in a reference list. # iii. Sample validity criteria Several WES and WTS validity criteria are also established for classifying a clinical sample as PASS or FAIL. A sample failing any number of these criteria is also referred to as QNS. WES and WTS QC results are recorded separately which are then used to determine which biomarkers are reportable for that sample. On WES side, the average read depth should be at least $100\mathrm{x}$ and the sample should also receive at least 20 million reads. In addition, all cancer-type drug rules genes which are of high clinical significance to different cancer types should have valid results (non-indeterminate results). On WTS side, the clinical sample should have a certain number of total and on-target RNA reads. Table 4. MI Cancer Seek QC Process | QC No. | QC Steps within Assay Workflow | Description | Impact of Failure | | --- | --- | --- | --- | | 1 | MI Cancer Seek Extraction Kit: QC Check for potential cross-contamination during TNA extraction with a negative control. | An extraction-negative control (EXT NEG) is added to each plate and does not contain tissue. The control is expected to not have measurable levels of TNA. The control must have DNA below a threshold DNA concentration measured with the MI Cancer Seek Quantification Kit. | If extraction negative control exceeds threshold, the whole plate will be re-extracted. | | | MI Cancer Seek Extraction Kit: QC Check for sufficient yield during TNA extraction with an extraction positive control. | An extraction positive control (HBCTL) is a kit component and must meet a minimum threshold DNA concentration measured with the MI Cancer Seek Quantification Kit. | If extraction positive control does not exceed threshold, the whole plate will be re-extracted. | | | MI Cancer Seek Extraction Kit: QC Check criteria for sufficient TNA extraction for each specimen. | Each specimen must meet a minimum threshold DNA concentration measured with the MI Cancer Seek Quantification Kit. | If the specimen does not exceed threshold, the sample will be re-extracted, with increased tissue input if possible. | | 2 | Library Pooling QC Check for quantification of final library pool concentration. | Pooled library must be above the lower spec limit and below the upper spec limit DNA concentrations | For pools outside of the concentration range, the Qubit quantification is repeated, or the samples are re-normalized. The re-pool is | | | | in the range of 10-100x. | re-normalized. The sample will be re-extracted. | PMA P240010: FDA Summary of Safety and Effectiveness Data {9} | QC No. | QC Steps within Assay Workflow | Description | Impact of Failure | | --- | --- | --- | --- | | | | measured with the MI Cancer Seek Library Quant Kit. | retested with the MI Cancer Seek Library Quant Kit. If the re-pool fails, remake libraries or re-extract TNA. | | 3 | MI Cancer Seek NGS sequencing metrics run validity criteria for DNA and RNA combined. | Reads Passing Filter (B) > 16 Yield (G) > 2,400 % of Q30 Reads (both R1 and R4) ≥ 75% | Run will be re-sequenced if one of the specified metrics fails. Repeat library preparation if necessary. | | 4 | MI Cancer Seek NGS bioinformatics pipeline QC Check for negative control validity criteria for the run. | The EXT-NEG (NTC) DNA criteria will be below specified thresholds: depth of <1x and ratio of mapped reads <5%. | Repeat library preparation for whole plate using extracted TNA. | | | MI Cancer Seek NGS bioinformatics pipeline QC Check for HBCTL positive control validity criteria for the run. | The HBCTL DNA target variants detected within lower and upper specification limits and validity criteria. | Repeat library preparation for whole plate using extracted TNA. | | | MI Cancer Seek NGS bioinformatics pipeline QC Check for POSCTL positive control DNA validity criteria for the run. | The POSCTL DNA and RNA target variants within lower and upper specification limits and validity criteria. | Repeat library preparation for whole plate using extracted TNA. | | | MI Cancer Seek NGS bioinformatics pipeline QC Check for HBCNA positive control DNA validity criteria for capabilities. | The HBCNA DNA validity criteria average depth ≥100x, DNA capabilities MSI value, TMB value and ERBB2 copy value for CNA will each be between a lower and upper specification limit criteria will be met. | The HBCNA specified capabilities will not be reported for the run. | | | MI Cancer Seek NGS bioinformatics pipeline QC Check for DNA sample validity criteria (also referred to as QNS). | The sample DNA validity criteria will be met including ≥20 M total reads and average depth of ≥100x | If DNA read criteria is not met, then remake library from TNA. | | | MI Cancer Seek NGS bioinformatics pipeline QC Check for Negative drug rules genes. | The sample is Negative for DNA variants if exons associated to the drug decision for the corresponding tumor type and validity criteria. | If validity criteria are not met, then remake the library from TNA. | ## E. Variant calling criteria In addition to the run and clinical sample validity criteria listed above, several variant calling criteria established separately for different biomarkers are used to determine which biomarkers are reportable in a passing sample in a successful run. a. SNV and Indels PMA P240010: FDA Summary of Safety and Effectiveness Data 10 of 85 {10} For SNV and indels to be reported, i.e., receive a “Variant Detected” result, they should achieve ≥20x read depth with at least 5 supporting reads, have a minimum VAF of >5% and should not have any poor-quality flags such as strand bias, false positives due to sequence reads containing residue to RNA or indel repeat of more than 8 nucleotides. A “low coverage” classification is assigned to any gene in the Reportable Gene list (Table 2) if there is no reportable variant from the gene and any exon in that gene has an average depth <100x with the exception of a list of exons in which frequently report depth below threshold and no important variant have been observed. This category, also referred to as “No Call”, does not rule out the presence or absence of a mutation. With low coverage, a call cannot be reliably determined. b. ERBB2 CNA For ERRB2 CNA to be reported, ERBB2 ≥ 6.9 copies should be observed. ERBB2 CNAs may also be reported if ERBB2 has ≥4.1 copies with HER2/CEP17 ratio ≥ 2.1. HER2/CEP17 ratio is a test that measures the number of HER2 gene copies on chromosome 17 (ERBB2) in relation to the number of chromosome 17 centromere (CEP17) copies per nucleus. Samples with ≥ 3.3 copies and ERBB2/CEP17 ≥ 2.1 are deemed to be “Intermediate” for ERBB2 CNAs. Samples with ERBB2 CNA results not covered above are deemed not amplified for ERBB2. Patients with breast cancer whose samples receive ERRB2 CNA “Intermediate” calls should be tested with another FDA approved or cleared test to ascertain ERRB2 CNA status in their tumor. Patient’s specimens with average read depth of < 100x are determined to be “Indeterminate” for ERBB2 CNAs. c. MSI For MSI to be reported as high, MSI-High (MSI-H), the clinical sample should have ≥39 frameshift mutations. MSI is reported as not MSI-H if there are <39 frameshift mutations and the average read depth of a panel of reference genes (N=753) is ≥ 100x. Patient’s specimens with average read depth of < 100x are determined to be “Indeterminate” for MSI. d. TMB TMB is reported quantitatively (mutations/megabase) and therefore has no reportable positive/negative result. Samples with a depth of coverage lower than 100x are considered “Indeterminate” and a result for TMB will not be provided. For samples with sufficient coverage, variants with at least 5% allele frequency identified across the whole exome are filtered to remove low quality, low depth, non-coding, synonymous, and any presumed germline variants which belong to gnomAD (AC>0), dbSNP 151 common. From the filtered list, missense, nonsense, in-frame indel, and frameshift variants in selected coding regions with sufficient depth are counted. The final value is the variant count divided by the length of the assessed genomic regions (25 Mb), presented as mutations/Mb. PMA P240010: FDA Summary of Safety and Effectiveness Data 11 of 85 {11} # F. Variant Classification for Companion Diagnostic Biomarkers Table 5 describes the biomarker rules for CDx claims in Table 1 of the intended use reported by MI Cancer Seek. Table 5. Biomarker Rules for Companion Diagnostic Claims Reported by MI Cancer Seek | Indication | Biomarker | Reportable Mutations | | --- | --- | --- | | Breast Cancer | PIK3CA SNVs | C420R E542K; E545A, E545D [1635G>T only], E545G, E545K, Q546E, Q546R H1047L, H1047R, H1047Y | | Colorectal Cancer (CRC) | KRAS wild-type (exons 2, 3, and 4) and NRAS wild-type (exons 2, 3, and 4) | Wild-type as determined by a “Variant Not Detected” result in the following locations: KRAS Exon 2: G12A, G12C, G12D, G12F, G12N, G12R, G12S, G12V, G12W, G13C, G13D (c.38G>A; c.38_39delinsAT), G13E, G13R, G13V KRAS Exon 3: A59G, A59T, Q61E, Q61H (c.183A>C; c.183A>T), Q61K, Q61L, Q61R KRAS Exon 4: A146P, A146T, A146V, K117N (c.351A>C; c.351A>T) NRAS Exon 2: G12A, G12C, G12D, G12F, G12N, G12R, G12S, G12V, G12W, G13C, G13D, G13E (c.38_39delinsAA; c.38_39delinsAG), G13R, G13V NRAS Exon 3: A59G, A59T, Q61E, Q61H (c.183A>T, c.183A>C), Q61K, Q61L, Q61R NRAS Exon 4: A146V, A146T, A146P, K117N (c.351G>C; c.351G>T) | | | BRAF SNV | V600E | | Melanoma | BRAF SNV | V600E, V600K | | NSCLC | EGFR activating mutations (EGFR exon 19 deletions and exon 21 SNV)* | Exon 21: L858R Exon 19 deletions: E746_S752delinsA, K745_A750del, E746_R748del, T751_E758del, T751_N756del, K757_I759delinsN, L747_K754del, E746_E749delinsQ, R748_T751del, E746_T751del, R748_K754del, P753_I759del, S752_A755del, E746del, R748_E749del, T751_A755del, L858R, E746_A750del, L747_T751del, S752_I759del, E746_T751delinsA, L747_P753delinsS, E746_T751delinsVP, E746_S752delinsV, L747_K754delinsATSPE, E746_T751delinsI, L747_E749del, L747_T751delinsP, L747_A755delinsSKD, L747_A750delinsP, T751_I759delinsN, E746_T751delinsIP, E746_S752delinsC, E746_A750delinsQP, A750_I759delinsSN, E746_T751delinsVA, L747_P753delinsQQ, L747_K754delinsG, L747_A750delinsS, E746_A750delinsIP, E746_T751delinsAA, L747_S752del, A750_I759delinsSSS, L747_A755delinsAT, E746_E749del, | PMA P240010: FDA Summary of Safety and Effectiveness Data {12} | | | E746_K754delinsVSE, K754_D761delinsN, L747_T751delinsN, E746_A750delinsFP, L747_K754delinsPRE, E746_P753delinsVS, L747_S752delinsQ, L747_A755delinsSKG, L747_A755delinsAN, A750_I759delinsGD, L747_K754delinsSPE, E746_A750delinsKP, E746_T751delinsV, I744_E749delinsMKL, L747_K754delinsAISPE, E746_T751delinsL, E746_S752delinsI, L747_A750del, L747_P753delinsQ, E746_L747delinsIP, L747_K754delinsQQ, L747_A755delinsSQQG, L747_A755delinsGN, L747_P753del, E746_A750delinsAP, L747_A755delinsSRD, A750_I759delinsPT, T751_I759delinsS, I744_E749delinsLKR, E746_S752delinsD, E746_A750delinsSP, L747_A755delinsATSPEG, E746_S752delinsL, E746_P753delinsLS, R748_A750del, E746_S752delinsIPVA, A750_I759delinsSQQG, L747_A755delinsNRET, E746_T751delinsKP, E746_T751delinsLP, E746_T751delinsAP, L747_P753delinsT, E749_S752delinsD, L747_A755delinsTKD, L747_S752delinsSRD, E746_A750delinsRP, L747delinsFPSLS, E746_R748delinsIPVAIKE, K754_I759del, R748_A755delinsT, L747_A755delinsGT, L747_K754delinsSPQ, E746_P753delinsIS, E746_A750delinsYP, L747_I759delinsSKANKEL, E746_T751delinsFPS, E746_T751delinsIS, A750_I759delinsGG, E746_P753delinsSS, L747_A755delinsNNNN, E746_A750delinsVP, E746_T751delinsLA, E746_T751delinsQ, L747_S752delinsQH, E746_P753delinsAS, L747_K754delinsNIE, E746_E749delinsK, A750_T751delinsVP, L747_T751delinsS, E749_A755delinsD, L747_P753delinsQT, A750_E758delinsP, L747_K754delinsQE, L747_S752delinsPF | | --- | --- | --- | | Solid Tumors | MSI | MSI-H | | Endometrial Carcinoma | MSI | Not MSI-H | | *More than one cDNA change may be associated with the same protein change. Mutations found in patients with the corresponding indication will be reported as a CDx for the associated therapies in the MI Cancer Seek Intended Use. | | | # VI. ALTERNATIVE PRACTICES AND PROCEDURES There are FDA-approved CDx alternatives for the detection of genetic alterations using FFPE tumor specimens, as listed in Table 1 of the MI Cancer Seek intended use statement. The approved CDx tests are listed in Table 6 below; for additional details see FDA List of Cleared or Approved Companion Diagnostic Devices at https://www.fda.gov/medical-devices/in-vitro-diagnostics/list-cleared-or-approved-companion-diagnostic-devices-in-vitro-and-imaging-tools. Each alternative has its own advantages and disadvantages. A patient should fully discuss these alternatives with his/her physician to select the method that best meets expectations and lifestyle. PMA P240010: FDA Summary of Safety and Effectiveness Data {13} Table 6. List of FDA-Approved CDx Assays for Biomarkers Identified by MI Cancer Seek | Biomarker | Device | Company | Technology | Therapy | Indication | | --- | --- | --- | --- | --- | --- | | BRAF V600E | therascreen BRAF V600E RGQ PCR Kit | QIAGEN GmbH | PCR² | Braftovi (encorafenib) in combination with Erbitux (cetuximab) | Colorectal Cancer | | | cobas 4800 BRAF V600 Mutation Test | Roche Molecular Systems, Inc. | PCR² | Zelboraf (vemurafenib) | Melanoma Cancer | | | cobas 4800 BRAF V600 Mutation Test | Roche Molecular Systems, Inc. | PCR² | Cotellic (cobimetinib) in combination with Zelboraf (vemurafenib) | Melanoma Cancer | | | FoundationOne CDx | Foundation Medicine, Inc. | NGS¹ | Mekinist (trametinib) | Melanoma Cancer | | | THXID BRAF Kit | bioMérieux Inc. | PCR² | Mekinist (trametinib) | Melanoma Cancer | | | THXID BRAF Kit | bioMérieux Inc. | PCR² | Tafinlar (dabrafenib) | Melanoma Cancer | | | THXID BRAF Kit | bioMérieux Inc. | PCR² | Braftovi (encorafenib) in combination with Mektovi (binimetinib) | Melanoma Cancer | | BRAF V600E/ BRAF V600K | FoundationOne CDx | Foundation Medicine, Inc. | NGS¹ | Mekinist (trametinib) or BRAF/MEK Inhibitor Combinations approved by FDA | Melanoma Cancer | | | THXID BRAF Kit | bioMérieux Inc. | PCR² | Mekinist (trametinib) or Braftovi (encorafenib) in combination with Mektovi (binimetinib) | Melanoma Cancer | PMA P240010: FDA Summary of Safety and Effectiveness Data {14} | Biomarker | Device | Company | Technology | Therapy | Indication | | --- | --- | --- | --- | --- | --- | | EGFR – Exon 19 deletions & L858R | cobas EGFR Mutation Test v2 | Roche Molecular Systems, Inc. | PCR2 | EGFR Tyrosine Kinase Inhibitors approved by FDA* | NSCLC | | | FoundationOne CDx | Foundation Medicine, Inc. | NGS1 | EGFR Tyrosine Kinase Inhibitors approved by FDA* | NSCLC | | | ONCO/Reveal Dx Lung & Colon Cancer Assay (O/RDx-LCCA) | Pillar Biosciences, Inc. | NGS1 | EGFR Tyrosine Kinase Inhibitors approved by FDA* | NSCLC | | | Oncomine Dx Target Test | Life Technologies Corporation | NGS1 | Iressa (gefitinib) | NSCLC | | | therascreen EGFR RGQ PCR Kit | Qiagen Manchester, Ltd. | PCR2 | Gilotrif (afatinib) | NSCLC | | | therascreen EGFR RGQ PCR Kit | Qiagen Manchester, Ltd. | PCR2 | Iressa (gefitinib) | NSCLC | | | therascreen EGFR RGQ PCR Kit | Qiagen Manchester, Ltd. | PCR2 | Vizimpro (dacomitinib) | NSCLC | | KRAS and NRAS | Praxis Extended RAS Panel | Illumina | NGS1 | Vectibix (panitumumab) | Colorectal Cancer | | MSI-H/dMMR | FoundationOne CDx | Foundation Medicine, Inc. | NGS1 | Keytruda (pembrolizumab) | Solid Tumor | | | Ventana MMR RxDx Panel | Ventana Medical Systems, Inc. | IHC3 | Keytruda (pembrolizumab) | Solid Tumor | | | Ventana MMR RxDx Panel | Ventana Medical Systems, Inc. | IHC3 | Jemperli (dostarlimab-gxly) | Solid Tumor | PMA P240010: FDA Summary of Safety and Effectiveness Data {15} | Biomarker | Device | Company | Technology | Therapy | Indication | | --- | --- | --- | --- | --- | --- | | Not MSI-H/pMMR | Ventana MMR RxDx Panel | Ventana Medical Systems, Inc. | IHC^{3} | Keytruda (pembrolizumab) in combination with Lenvima (lenvatinib) | Endometrial Cancer | | PIK3CA | FoundationOne CDx | Foundation Medicine, Inc. | NGS^{1} | Piqray (alpelisib) | Breast Cancer | | | therascreen PIK3CA RGQ PCR Kit | QIAGEN GmbH | PCR^{2} | Piqray (alpelisib) | Breast Cancer | | *For the most current information about the device indications for the therapeutic products in this group, go to: https://www.fda.gov/medical-devices/in-vitro-diagnostics/list-cleared-or-approved-companion-diagnostic-devices-in-vitro-and-imaging-tools#Group_Labeling Abbreviations: 1 NGS = Next Generation Sequencing 2 PCR = Polymerase Chain Reaction 3 IHC = Immunohistochemistry | | | | | | ## VII. MARKETING HISTORY Caris Life Sciences indicates the MI Cancer Seek assay was offered as a laboratory developed test (LDT) with the name “MTS LDT” since November 2022. As described in the preamble to FDA’s LDT final rule, FDA uses the term "IVDs offered as LDTs" for IVDs that are manufactured and offered as LDTs by laboratories that are certified under CLIA and that meet the regulatory requirements under CLIA to perform high complexity testing, and used within such laboratories, even if those IVDs do not fall within FDA’s traditional understanding of an LDT because they are not designed, manufactured, and used within a single laboratory. ## VIII. POTENTIAL ADVERSE EFFECTS OF THE DEVICE ON HEALTH Failure of the device to perform as expected or failure to correctly interpret test results may lead to incorrect test results, and subsequently, inappropriate patient management decisions. Patients with false positive results may be inappropriately treated with one of the therapies listed in the above intended use statement without clinical benefit and may experience adverse reactions associated with inappropriate therapy. Patients with false negative results may not be considered for treatment with the indicated therapy. There is also a risk of delayed results, which may lead to delay of treatment with indicated therapy. For the specific adverse events related to the approved therapeutics, please see approved drug product labels which are available at Drugs@FDA. PMA P240010: FDA Summary of Safety and Effectiveness Data {16} PMA P240010: FDA Summary of Safety and Effectiveness Data 17 of 85 # IX. SUMMARY OF NON-CLINICAL STUDIES ## A. Laboratory Studies Caris Life Sciences executed 16 analytical validation studies to support the CDx and tumor profiling claims indicated in the intended use statement. Studies were performed using FFPE specimens and TNA extracted from a wide range of FFPE tissue types having various genomic alterations across a number of genes. Studies included CDx variants, ERBB2 copy number amplification, as well as a broad range of representative SNVs and indels for tumor profiling. Analysis of the genomic signatures for MSI and TMB was also performed. See Section IX-A below for analytical validation study summaries. ## 1. Analytical Accuracy ### a. Comparison to an Orthogonal Method for Tumor Profiling (SNV and indels) and TMB This study demonstrated accuracy of Tumor Profiling variants (SNVs, MNVs and indels) and TMB for the MI Cancer Seek assay by comparing variant calling using two orthogonal validated NGS assays. A total of 500 clinical FFPE samples across 38 tumor types were consecutively enrolled based on tumor type and tissue requirements to support testing with three assays (MI Cancer Seek and two orthogonal NGS assays). To evaluate variant accuracy, samples were tested on an FDA-cleared commercially available Targeted Gene Panel assay and processed per the assay Instructions for Use (IFU) using optimal input (100 ng) where possible. To evaluate TMB accuracy, samples were tested using the MI Cancer Seek and an externally validated WES assay. Samples for MI Cancer Seek were processed using the complete test workflow starting with extraction and the majority were tested at the minimal allowable input (50 ng) to allow for more challenging comparisons. ## Accuracy of SNVs, MNVs, and Indels Concordance for Tumor Profiling A total of 500 samples were tested on both MI Cancer Seek and the Targeted Gene Panel assay, of which 454 samples were analyzed to determine variant accuracy across 131 genes and 527 exons having reportable pathogenic or likely pathogenic variants detected by both assays. Forty-six (46) samples were excluded from the concordance analysis due to either sample quality or sample swaps that occurred during orthogonal method testing. The apparent sample swap for 30 cases involved the Targeted Gene Panel assay, therefore these samples were excluded because definitive evidence for the sample swap outside of the data analysis could not be obtained. The accuracy results are summarized by variant type, and further stratified by FDA's Biomarker Class Level for tumor profiling in Table 7. {17} Table 7: Agreement Summary for SNVs, MNVs, Insertions and Deletions for Tumor Profiling | Variant Type | Total Unique Variants | MI Cancer Seek (+) Comparator (+) | MI Cancer Seek (-) Comparator (-) | MI Cancer Seek (-) Comparator (+) | MI Cancer Seek (-) Comparator (-) | PPA (n/N) [95% CI] | NPA (n/N) [95% CI] | | --- | --- | --- | --- | --- | --- | --- | --- | | All Variants | 170636 | 1177 | 215 | 54 | 77467298 | 95.61% (1177/1231) [94.31-96.69] | 100.00% (77467298/77467513) [100.00-100.00] | | All SNVs | 50425 | 923 | 86 | 46 | 22891895 | 95.25% (923/969) [93.72-96.50] | 100.00% (22891895/22891981) [100.00-100.00] | | Level 1 SNVs | 15 | 67 | 0 | 4 | 6739 | 94.37% (67/71) [86.20-98.44] | 100.00% (6739/6739) [99.95-100.00] | | Level 2 SNVs | 25018 | 617 | 46 | 31 | 11357478 | 95.22% (617/648) [93.28-96.73] | 100.00% (11357478/11357524) [100.00-100.00] | | Level 3 SNVs | 25392 | 239 | 40 | 11 | 11527678 | 95.60% (239/250) [92.26-97.78] | 100.00% (11527678/11527718) [100.00-100.00] | | All MNVs | 4563 | 3 | 12 | 0 | 2071587 | 100.00% (3/3) [29.24-100.00] | 100.00% (2071587/2071599) [100.00-100.00] | | Level 1 MNVs | 29 | 0 | 0 | 0 | 13166 | NA (0/0) [0-0] | 100.00% (13166/13166) [99.97-100.00] | | Level 2 MNVs | 2879 | 3 | 10 | 0 | 1307053 | 100.00% (3/3) [29.24-100.00] | 100.00% (1307053/1307063) [100.00-100.00] | | Level 3 MNVs | 1655 | 0 | 2 | 0 | 751368 | NA (0/0) [0-0] | 100.00% (751368/751370) [100.00-100.00] | | All Insertions | 31090 | 59 | 23 | 0 | 14114778 | 100.00% (59/59) [93.94-100.00] | 100.00% (14114778/14114801) [100.00-100.00] | | Level 1 Insertions | 2 | 0 | 0 | 0 | 908 | NA (0/0) [0-0] | 100.00% (908/908) [99.59-100.00] | | Level 2 Insertions | 16851 | 27 | 12 | 0 | 7650315 | 100.00% (27/27) [87.23-100.00] | 100.00% (7650315/7650327) [100.00-100.00] | | Level 3 Insertions | 14237 | 32 | 11 | 0 | 6463555 | 100.00% (32/32) [89.11-100.00] | 100.00% (6463555/6463566) [100.00-100.00] | PMA P240010: FDA Summary of Safety and Effectiveness Data {18} | Insertions 1-5bp | 26065 | 54 | 21 | 0 | 11833435 | 100.00% (54/54) [93.40-100.00] | 100.00% (11833435/11833456) [100.00-100.00] | | --- | --- | --- | --- | --- | --- | --- | --- | | Insertions 6-10bp | 2044 | 2 | 0 | 0 | 927974 | 100.00% (2/2) [15.81-100.00] | 100.00% (927974/927974) [100.00-100.00] | | Insertions 11-15bp | 669 | 2 | 0 | 0 | 303724 | 100.00% (2/2) [15.81-100.00] | 100.00% (303724/303724) [100.00-100.00] | | Insertions 16-20bp | 643 | 1 | 1 | 0 | 291920 | 100.00% (1/1) [2.50-100.00] | 100.00% (291920/291921) [100.00-100.00] | | Insertions 21-25bp | 457 | 0 | 0 | 0 | 207478 | NA (0/0) [0-0] | 100.00% (207478/207478) [100.00-100.00] | | Insertions >25bp | 1212 | 0 | 1 | 0 | 550247 | NA (0/0) [0-0] | 100.00% (550247/550248) [100.00-100.00] | | All Deletions | 84558 | 192 | 94 | 8 | 38389038 | 96.00% (192/200) [92.27-98.26] | 100.00% (38389038/38389132) [100.00-100.00] | | Level 1 Deletions | 166 | 0 | 0 | 1 | 75363 | 0.00% (0/1) [0.00-97.50] | 100.00% (75363/75363) [100.00-100.00] | | Level 2 Deletions | 48564 | 100 | 37 | 3 | 22047916 | 97.09% (100/103) [91.72-99.40] | 100.00% (22047916/22047953) [100.00-100.00] | | Level 3 Deletions | 35828 | 92 | 57 | 4 | 16265759 | 95.83% (92/96) [89.67-98.85] | 100.00% (16265759/16265816) [100.00-100.00] | | Deletions 1-5bp | 51675 | 179 | 84 | 8 | 23460179 | 95.72% (179/187) [91.74-98.14] | 100.00% (23460179/23460263) [100.00-100.00] | | Deletions 6-10bp | 8958 | 7 | 4 | 0 | 4066921 | 100.00% (7/7) [59.04-100.00] | 100.00% (4066921/4066925) [100.00-100.00] | | Deletions 11-15bp | 7013 | 1 | 1 | 0 | 3183900 | 100.00% (1/1) [2.50-100.00] | 100.00% (3183900/3183901) [100.00-100.00] | | Deletions 16-20bp | 5553 | 0 | 0 | 0 | 2521062 | NA (0/0) [0-0] | 100.00% (2521062/2521062) [100.00-100.00] | | Deletions 21-25bp | 3198 | 2 | 1 | 0 | 1451889 | 100.00% (2/2) [15.81-100.00] | 100.00% (1451889/1451890) [100.00-100.00] | | Deletions >25bp | 8161 | 3 | 4 | 0 | 3705087 | 100.00% (3/3) [29.24-100.00] | 100.00% (3705087/3705091) [100.00-100.00] | The discordances observed were due to design differences, including variant calling thresholds and variant calling rules, between the orthogonal Targeted Gene Panel Assay and MI Cancer Seek. Twenty-seven (27) false negative variants were observed because the orthogonal Targeted Gene Panel threshold for positive results is lower than the MI Cancer Seek threshold of $5\%$ variant frequency. An additional 19 false PMA P240010: FDA Summary of Safety and Effectiveness Data {19} negatives were the result of differences in merging of multiple variant types between the two assays. For the majority of false positives, variant alignments were present for the Targeted Gene Panel assay, but the calls were not reported. # TMB Concordance A total of 500 samples were tested on both MI Cancer Seek and the externally validated whole exome sequencing assay, of which 497 samples were analyzed to determine TMB accuracy across 38 different tumor types. Three (3) samples were excluded from analysis due to sample swaps (there is an apparent sample swap for 2 cases involved with the whole exome sequencing assay, therefore these samples were excluded because definitive evidence for the sample swap outside of the data analysis could not be obtained) or QC failure during orthogonal method testing. A linear regression analysis was completed using Deming regression to estimate slope and intercept along with $95\%$ confidence intervals (CIs) for the measured TMB values between the two assays, as shown in Figure 1. This analysis reflects low-coverage regions $(<50\mathrm{x})$ that underperformed in the orthogonal method and artificially underestimates the orthogonal method TMB value, which contributed to the observed discordances. Variants were not detected in regions of low coverage by the orthogonal method and therefore did not contribute to the TMB value but were detected in these regions by MI Cancer Seek and contributed to the TMB value for a given sample. This was further evaluated by calculating TMB for a subset of exome regions where both assays (MI Cancer Seek and orthogonal assay) had greater or equal to $50\mathrm{x}$ depth of coverage. The slope and intercept of the Deming regression both improved from 0.856 to 0.927 and -0.859 to -0.68, respectively. These data indicate TMB calls between MI Cancer Seek and the orthogonal method are concordant when coverage regions $\geq50\mathrm{x}$ are evaluated. ![img-0.jpeg](img-0.jpeg) Figure 1: TMB Accuracy Using Deming Regression Analysis PMA P240010: FDA Summary of Safety and Effectiveness Data {20} b. Comparison to an Orthogonal Fluorescence In Situ Hybridization (FISH) Assay for ERBB2 CNAs The accuracy of MI Cancer Seek was evaluated for the detection of ERBB2 CNAs in patient samples with breast cancer. A total of 288 unique FFPE tissue samples from patients histologically confirmed as breast cancer were enrolled and tested with FDA approved fluorescence in situ hybridization (FISH), PathVysion HER2 DNA Probe Kit, and MI Cancer Seek. Eight samples did not meet MI Cancer Seek testing requirements; therefore, 280 samples were tested on MI Cancer Seek. In the 280 samples tested on both MI Cancer Seek and FISH assay, five (5) samples did not have valid FISH assay results. Out of the 275 samples that had valid FISH results, one sample did not have valid MI Cancer Seek result. One (1) sample had intermediate MI Cancer Seek result. Concordance between MI Cancer Seek and the FISH assay is shown in Table 8. Table 8. Agreements between the MI Cancer Seek and Orthogonal FISH assay for ERBB2 CNAs in samples from patients with breast cancer | | FISH+ | FISH- | Total* | | --- | --- | --- | --- | | MI Cancer Seek + | 99 | 1 | 100 | | MI Cancer Seek Intermediate | 1 | 0 | 1 | | MI Cancer Seek - | 18 | 155 | 173 | | MI Cancer Seek Invalid | 1 | 0 | 1 | | Total | 119 | 156 | 275 | *There were another 5 FISH invalid samples. Concordance results from Table 8, treating the MI Cancer Seek “Intermediate” result as positive or negative result were used to determine agreements shown in Table 9. Using the FISH assay results as reference and treating the MI Cancer Seek “Intermediate” results as “Positive” results, the PPA is 84.7% with two-sided 95% confidence interval (77.2%, 90.1%), and NPA is 99.4% with two-sided 95% confidence interval (96.5%, 99.9%). When treating MI Cancer Seek “Intermediate” results as “Negative” results, the PPA is 83.9% with two-sided 95% confidence interval (76.2%, 89.4%) and NPA is 99.4% with two-sided 95% confidence interval 96.5%, 99.9%). Table 9. Agreement Summary for ERBB2 CNAs in breast cancer for Tumor Profiling | | Agreement (95% CI*) Intermediate as Positive | Agreement (95% CI*) Intermediate as Negative | | --- | --- | --- | | PPA | 84.7% (77.2%, 90.1%) | 83.9% (76.2%, 89.4%) | | NPA | 99.4% (96.5%, 99.9%) | 99.4% (96.5%, 99.9%) | *95% confidence intervals were calculated by Wilson score method. PMA P240010: FDA Summary of Safety and Effectiveness Data 21 of 85 {21} Based on the PPA observed for the detection of ERBB2 CNAs, this alteration may not be detected in patients with breast cancer. Additional clinical investigation to confirm the presence of ERBB2 CNAs in the breast cancer patient's tumor with another FDA approved or cleared test is strongly recommended. For additional concordance data for the CDx-associated variants, refer to the Summary of Primary Clinical Studies in Section X. ## 2. Analytical Sensitivity ### a. Limit of Blank This study was designed to estimate the Limit of Blank (LoB) or false positive call rate of MI Cancer Seek assay using adjacent non-tumor tissue across multiple tissue types. A total of 168 data points were generated across 28 FFPE non-tumor samples and 15 tissue types using two lots of reagents. Each sample was extracted in triplicate using one lot of extraction reagent and further utilized to prepare libraries at maximum DNA input (220 ng). A total of six library replicates were prepared and sequenced per sample using two lots of reagents. No false positives were detected for Level 1 CDx biomarkers, Level 2 SNV/indels, or Level 3 indels; for Level 3 SNVs, <0.01% false positives were detected, refer to Table 10. Table 10. LoB Study Summary Results | Category | Per Position False Positive Rate | Per Sample False Positive Rate* | | --- | --- | --- | | Level 1: BRAF V600E | 0% | 0% | | Level 1: BRAF V600K | 0% | 0% | | Level 1: EGFR Exon 19 deletions | 0% | 0% | | Level 1: EGFR Exon 21 substitution mutations | 0% | 0% | | Level 1: PIK3CA C420R | 0% | 0% | | Level 1: PIK3CA E542K, E545A, E545D [1635G>T only], E545G, E545K, Q546E, Q546R | 0% | 0% | | Level 1: PIK3CA H1047L, H1047R, H1047Y | 0% | 0% | | Level 1: KRAS SNVs | 0% | 0% | | Level 1: NRAS SNVs | 0% | 0% | | Level 1: MSI-H | N/A | 0% | | Level 2: ERBB2 CNA | N/A | 0% | | Level 2: Panel-wide SNVs (924) | 0% | 0% | PMA P240010: FDA Summary of Safety and Effectiveness Data 22 of 85 {22} The 3 false positives in Level 3: Panel wide SNVs were each from TERT c.-124C>T and had variant frequencies near the threshold of 5% at 5.3%, 5.2% and 5.8%. An additional study was performed using 10 WT FFPE samples, including 4 breast tissue and 6 lung tissue samples, across two reagent lots. A total of 240 additional data points were generated across two tissue types using two reagent lots at maximum DNA input for the assay. There were no false positive calls for ERBB2 amplification in both breast tissue and lung tissue. The data also demonstrated 0% false positive rates for each SNV, indel, and MSI in both breast tissue and lung tissue, as shown in Table 11. Table 11. Supplemental LoB Study Results Per Position and Sample | Level & Variant Class | Per Position False Positive Rate | Per Sample False Positive Rate* | | --- | --- | --- | | Level 1: BRAF V600E | 0% | 0% | | Level 1: BRAF V600K | 0% | 0% | | Level 1: EGFR Exon 19 Dels | 0% | 0% | | Level 1: EGFR L858R | 0% | 0% | | Level 1: PIK3CA C420R | 0% | 0% | | Level 1: PIK3CA E542K, E545A, E545D [1635G>T only], E545G, E545K, Q546E, Q546R | 0% | 0% | | Level 1: PIK3CA H1047L, H1047R, H1047Y | 0% | 0% | | Level 2: Panel-wide SNVs (N=924) | 0% | 0% | | Level 2: Panel-wide Indels (N=950) | 0% | 0% | | Level 3: Panel-wide SNVs (N=50710) | 0% | 0% | | Level 3: Panel-wide Indels (N=115407) | 0% | 0% | | Level 1: MSI | N/A | 0% | | Level 1: ERBB2 Amplifications | N/A | 0% | | *Calculated from 2 reagent lots (n=240 data points) | | | b. Limit of Detection for SNVs and Indels Due to the large number of variants detected by the MI Cancer Seek across multiple tissue types and the limitation of testing them all to support tumor profiling claims, Limit of Detection (LoD) of SNVs and indels were determined using a representative approach. A total of 29 samples were tested at minimal DNA input (50 ng), including PMA P240010: FDA Summary of Safety and Effectiveness Data {23} 12 samples to support CDx biomarker claims and 17 samples to support tumor profiling claims. At least one variant was tested per CDx biomarker for its respective tumor type. A total of 100 data points (ten replicates x five dilution levels x two reagent lots) were generated per panel member for the 12 CDx biomarkers and 25 data points (five replicates x five dilution levels x one reagent lot) were generated per sample for the 17 tumor profiling markers. Additional testing was performed for tumor profiling marker PIK3CAR1 N453 Y463del, and total of 35 data points were generated across seven dilution levels as repeat testing using one lot of reagents. The LoD for each panel member was determined as the lowest VAF level with $\geq 95\%$ detection (Hit rate $= 0.95$ ) based on the positivity threshold of $5\%$ . The LoD for SNVs and insertions ranged from $6 - 11\%$ VAF, and LoD for deletions ranged from $6 - 8\%$ VAF. The LoD for CDx biomarkers and tumor profiling markers are shown in Table 12 and Table 13, respectively. Table 12. LoD for CDx Biomarkers | Variant Type | Mutation | Tumor Type | LoD (%VAF) | | --- | --- | --- | --- | | SNV | BRAF V600E | Colorectal Adenocarcinoma | 11% | | SNV | BRAF V600E | Melanoma | 7% | | SNV | BRAF V600K | Melanoma | 8% | | Indel | EGFR L747 T751del | NSCLC) | 7% | | SNV | EGFR L858R | NSCLC | 7% | | SNV | KRAS A146T | Colorectal Adenocarcinoma | 8% | | SNV | KRAS G12C | Colorectal Adenocarcinoma | 7% | | SNV | KRAS Q61H | Colorectal Adenocarcinoma | 10% | | SNV | NRAS A146T | Colorectal Adenocarcinoma | 11% | | SNV | NRAS G12A | Colorectal Adenocarcinoma | 7% | | SNV | NRAS Q61R | Colorectal Adenocarcinoma | 10% | | SNV | PIK3CA E542K | Breast Carcinoma | 11% | Table 13. LoD for Tumor Profiling Alterations | Variant Type | Mutation | Tumor Type | LoD (%VAF) | | --- | --- | --- | --- | | Insertion (Long homopolymer) | ARID1A A330fs | Uterine Neoplasms - Endometrial carcinoma | 8% | | SNV | ARID1A Q575* | Uterine Neoplasms - Endometrial carcinoma | 6% | | SNV | ARID1A Q575* | Uterine Neoplasms - Endometrial carcinoma | 10% | | SNV | ARID1A Q575* | Uterine Neoplasms - Endometrial carcinoma | 11% | PMA P240010: FDA Summary of Safety and Effectiveness Data {24} | Variant Type | Mutation | Tumor Type | LoD (%VAF) | | --- | --- | --- | --- | | Deletion (Long non-homopolymer) | ASXL1 E635fs | Ovarian Surface Epithelial Carcinomas | 7% | | SNV | BRCA1 M1775R | Breast Carcinoma | 7% | | SNV | BRCA2 S1882* | Prostatic Adenocarcinoma | 8% | | Insertion | BRCA2 T2125fs | Prostatic Adenocarcinoma | 11% | | Insertion | EGFR A767_V769dup | NSCLC | 9% | | Insertion | ERBB2 G776delinsVC | NSCLC | 9% | | Insertion (Long non-homopolymer) | FANCA D944fs | Colorectal Adenocarcinoma | 8% | | Insertion | NF1 I679fs | Uterine Neoplasms - Endometrial carcinoma | 6% | | SNV | NOTCH1 Y550fs | Low Grade Glioma | 7% | | Deletion (Long homopolymer) | PIK3R1 N453_Y463del | Colorectal Adenocarcinoma | 8% | | Insertion (Short homopolymer) | SMAD2 L175fs | Breast Carcinoma | 9% | | SNV | SMARCA4 R1005* | Colorectal Adenocarcinoma | 7% | | Deletion (Short homopolymer) | SUFU T13fs | Melanoma | 6% | | SNV | TERT c.-146C>T | Glioblastoma | 11% | | SNV | TP53 G266V | Ovarian Surface Epithelial Carcinomas | 6% | c. Limit of Detection for ERBB2 CNA The LoD for ERBB2 CNA was also established by testing two ERBB2 breast cancer samples at five different dilution levels, using two reagent lots and 10 replicates per lot per dilution level for a total of 100 observations per sample. The samples were run at minimum DNA input (50ng). The LoD for each panel member was determined to be the lowest level of copies (CNA) with $\geq 95\%$ detection (Hit rate =0.95). and these data established 8.3 copies as the LoD for ERBB2 CNA. PMA P240010: FDA Summary of Safety and Effectiveness Data {25} d. Tumor Purity The lowest level of tumor content that supports robust performance of MI Cancer Seek was evaluated. A total of 12 unique samples from nine tumor types were prepared and tested at minimal DNA input (50ng) across different variant classes including CNA (2), MSI (9), and TMB (10). At least five dilution levels were prepared per sample resulting in a total of 100 data points (ten replicates x five dilution levels x two reagent lots) per sample for 14 CDx biomarkers using two reagent lots. The results demonstrated LoD for tumor content was between 5 -20% across the various samples and variant classes, which confirmed performance of MI Cancer Seek at the minimum requirement of 20% tumor content. e. DNA Input Study The study evaluated performance of MI Cancer Seek at six DNA input levels to not only verify performance at the assay minimum (50 ng) and maximum/optimum DNA input (220 ng) but to also guard band around these required levels. The DNA input levels tested included 25, 37.5, 50, 220, 275 and 330 ng, which correspond to -25% and -50% of the minimum requirement and +25% and +50% of the maximum requirement. The study tested 16 samples near LoD (1-3x). A minimum of three replicates per sample were processed through library preparation and sequencing using one lot of reagents for each input level. A total of 18 data points were generated per panel member for SNVs, indels, MSI and TMB, and 36 data points were generated for ERBB2 CNA. The PPA and NPA for SNVs, indels, ERBB2 CNA, and MSI were 100% between minimum and maximum DNA input. In addition, for TMB the upper bound of the 95% CI for absolute percent difference in mean TMB values between minimal and optimal inputs ranged 0.0% - 10.5%. 3. Analytical Specificity a. Interfering Substances – Exogenous This study evaluated tolerance to exogenous substances that are introduced during sample processing and are possibly carried over to the next steps in the assay workflow when using MI Cancer Seek assay. Substances tested included paraffin, xylene, Proteinase K, and 80% ethanol, spiked at the appropriate step to mimic carryover of the substance being tested. A total of 17 FFPE clinical samples from eight different tissue types containing genetic alterations at 2-3x LoD were tested across all capabilities including SNVs, indel, ERBB2 CNA, MSI, and TMB using a single reagent lot. A total of 425 data points (17 FFPE clinical samples x five replicates x five conditions) were generated across all interferent conditions including test and nominal/control condition. For ERBB2 CNA, PPA was reported at 100% for all conditions and NPA was reported at 100% for all conditions except Control vs Proteinase K, where NPA was reported at 98.7%. For indels, PPA and NPA were both reported at 100% for all PMA P240010: FDA Summary of Safety and Effectiveness Data 26 of 85 {26} conditions. For SNVs, PPA was reported at 100% for all conditions except Control vs Proteinase K, where PPA was reported at 97.4% and NPA was reported at 100% for all conditions. For MSI, PPA was reported at 100% for all conditions except Control vs Proteinase K, where PPA was reported at 93.3% and NPA was reported at 100% for all conditions. There was no significant impact to TMB detection ability in the presence of the tested exogenous substances. TMB included 0 for the 95% 1-sided CI around the mean difference between all TMB test and control condition. ## b. Interfering Substances – Endogenous To determine the potential impact of endogenous interfering substances on the performance of the MI Cancer Seek assay, in their respective tumor tissue samples for both CDx claims and tumor profiling claims, this study evaluated 23 clinical FFPE samples from ten tumor types (breast carcinoma, CRC, melanoma, NSCLC, small intestine malignancies, pancreatic adenocarcinoma, uterine neoplasms-endometrial carcinoma, prostatic adenocarcinoma, thyroid, cholangiocarcinoma), which were mutation positive samples selected at 2-3x LoD from the tissue types associated with the endogenous interfering substances. For each sample, five replicates each were processed in parallel through the complete assay workflow using one reagent lot at minimum DNA input (50 ng) for both a sample spiked with the potential endogenous interfering substance (refer to Table 14 for the interferents evaluated) and for the corresponding control condition (without the introduction of potentially interfering substance). A total of 130 data points were obtained for both endogenous interferent positive and endogenous interferent negative (control) replicates yielding a total of 260 data points. Table 14. Interferents Evaluated | Substance | Tumor Type | | --- | --- | | Control Condition | All tumor types | | Hemoglobin | All tumor types | | Colloid | Thyroid | | Calcium Phosphate | Thyroid | | Mucin | CRC/Prostate | | Bile acids, conjugated | Cholangiocarcinoma | For ERBB2 CNA, indel, and SNVs, both NPA and PPA were 100% and no deleterious effect between control and test conditions were observed when tested with endogenous potential interfering substances in intended tumor types. For TMB, the mean of each sample at each condition was calculated. The mean difference between test and control conditions across all samples was determined and the distribution of TMB differences were evaluated to obtain the one-sided 95% CI for the mean difference. The results demonstrated that comparison of conditions with hemoglobin, calcium, colloid, mucin, and bile acids (Test) to without interfering PMA P240010: FDA Summary of Safety and Effectiveness Data 27 of 85 {27} substance (Control) included the 0 for the one-sided 95% CI of the mean difference thus passing the study acceptance criteria for all conditions tested. For MSI, NPA was reported at 97.6% (Control vs Hemoglobin), 93.8% (Control vs Calcium), 100% (Control vs Mucin), 100% (Control vs Bile Acids) and 100% (Control vs Colloid). PPA was reported at 91.3% (Control vs Hemoglobin), 100% (Control vs Calcium), 100% (Control vs Mucin), 100% (Control vs Bile Acids) and 50% (Control vs Colloid). NPA for MSI was impacted by the presence of Colloid. However, the only positive sample selected had replicate values that were near the assay cut-off in the control replicates and test replicates. Thus, a lower PPA is expected for this sample. Further no significant difference between the test and control sample using the same analysis method described above for TMB, this condition passes, showing that it is the sample itself, rather than colloid, that is the root cause of lower NPA. Performance of MI Cancer Seek in samples with high level presence of necrotic tissue, melanin and fatty acids has not been demonstrated yet. A post market study will be conducted to assess the impact of these potential interferents. c. Carryover and Cross-Contamination This study was designed to assess the potential of erroneous results due to carryover (run to run) and cross-contamination (within run) throughout the complete MI Cancer Seek assay workflow, at the highest DNA input of 220 ng. A total of 172 data points (85 positive and 87 negative sample replicates) were generated for cross-contamination and 172 data points (86 positive and 86 negative samples replicates) were obtained for carryover using 30 unique FFPE samples (14 high positive and 16 negative) from five tumor types across one lot of reagents and one set of instruments. For the purpose of this study, TMB values were divided into two levels, 1-9 mut/Mb, and ≥10 mut/Mb. The results demonstrated that the NPA for SNV, indels, MSI, TMB, and ERBB2 CNA was 100% for cross-contamination and carryover. d. Index Cross-Contamination The study evaluated index cross-contamination due to potential index hopping and was designed to demonstrate that the rates of index cross contamination do not significantly affect the MI Cancer Seek assay results. The study utilized previously obtained results from 300 randomly selected samples across 20 different flowcells each having a unique index per plate, from previously performed MI Cancer Seek analytical validation studies. Results demonstrated that there was <0.00% cross contamination in all 20 flowcells evaluated. Additionally, all 300 samples had <0.00% cross contamination, which shows that the degree of index hopping does not impact MI Cancer Seek assay performance. PMA P240010: FDA Summary of Safety and Effectiveness Data 28 of 85 {28} e. Hybrid Capture Bait Specificity- in silico Analysis Specificity of the Hybrid Capture Baits (MI Cancer Seek Baits) was evaluated for reportable regions in the MI Cancer Seek assay, including the ability to differentiate between target analyte sequences and sequences generated from other sources (i.e. off-target sequences), using results from random selection of ten flowcells and 300 samples previously tested across MI Cancer Seek analytical validation studies. This study demonstrated that capture of off-targeted sequences does not significantly affect MI Cancer Seek performance. The study confirmed bait specificity, as the mean average depths of coverage were 757x in reportable regions and 1.5x in off-targeted regions. The mean percent off-target coverage was 0.19% of the on-target coverage in reportable regions (ranged from 0.08% to 0.33%). CDx-specific regions represent a small portion of the overall reportable regions, which was confirmed by the percentage of reads mapping to CDx regions (mean of 0.22%), and this study indicates high quality reads are also obtained in the CDx regions given the mean average depth of coverage was 958x. 4. Precision This study was performed to evaluate MI Cancer Seek repeatability (within-run) and reproducibility (total within-laboratory) under varied conditions including different reagent lots, instruments, operators, and non-consecutive run days. A total of 49 panel members representing 14 different tumor types were selected to evaluate CDx markers, challenging tumor profiling variant types including short/long indels in non-homopolymer/homopolymer regions, and the other biomarkers reported by MI Cancer Seek (MSI, ERBB2 CNA, and TMB). Of the 49 panel members, 37 were evaluated at low and high dilution level relative to LoD at 1-1.5x and 2-3x for targeted SNVs and indels yielding 74 samples. Two panel members consisted of ERBB2 CNA positive samples with 1-1.5x and 2-3x LoD levels. The remaining panel members consisted of samples selected to have MSI-H or Not MSI-H status, TMB with various mutations/Mb levels, or no specific mutation (2 panel members). In total there were 86 samples evaluated in the precision study. The study included three operator teams, three instrument sets, three reagent lots and was minimally executed across three non-consecutive days over a 20-day span. This study design resulted in 36 data points per sample. There were no invalid data points for the expected positive results across of replicates per all the samples tested in this study. A call was considered no call if the corresponding exon has a depth equal to or higher than 100x and the target variant was not detected. Agreements for targeted CDx variants and tumor profiling biomarkers evaluated are presented in Table 15. Targeted variants refer to the variants that were the basis for selection of the sample for evaluation in the study. All the CDx biomarkers tested for SNVs and indels had 100% positive agreements for both the levels. ERBB2 CNA at 2-3x LoD and MSI-H with medium positives (level well above the cut-off resulted in 100% positive agreement as well, ERBB2 CNA at 1-1.5x LoD had positive agreement of 97.2% and MSI-H with low positives (close to the cut-off) had agreements of 91.7% and 97.2% in the 2 Panel Members tested. The two samples with not MSI-H had agreements of PMA P240010: FDA Summary of Safety and Effectiveness Data 29 of 85 {29} $100\%$ and $94.4\%$ ; lower agreement corresponding to the not MSI-H sample that had a level close to the cut-off. Table 15. Agreements for Targeted CDx and Tumor Profiling Variants reported by MI Cancer Seek | Sample | Lineage | Variant | Bio-marker Type | Fold LoD Level*** | Average VAF, MSI Score, or ERBB2 Copies | Number Positive / Number Expected | Agreement (95% CI) | | --- | --- | --- | --- | --- | --- | --- | --- | | 1 | Breast Carcinoma | PIK3CA C420R | CDx | 1-1.5x | 0.139 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.289 | 36/36 | 100% (90.3%, 100.0%) | | 2 | Breast Carcinoma | PIK3CA E542K | CDx | 1-1.5x | 0.159 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.187 | 36/36 | 100% (90.3%, 100.0%) | | 3 | Breast Carcinoma | PIK3CA H1047R | CDx | 1-1.5x | 0.135 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.236 | 36/36 | 100% (90.3%, 100.0%) | | 4 | Colorectal Adenocarcinoma | BRAF V600E | CDx | 1-1.5x | 0.124 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.21 | 36/36 | 100% (90.3%, 100.0%) | | 5 | Colorectal Adenocarcinoma | KRAS A146T | CDx | 1-1.5x | 0.112 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.169 | 36/36 | 100% (90.3%, 100.0%) | | 6 | Colorectal Adenocarcinoma | KRAS G12C | CDx | 1-1.5x | 0.302 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.518 | 36/36 | 100% (90.3%, 100.0%) | | 7 | Colorectal Adenocarcinoma | KRAS Q61K | CDx | 1-1.5x | 0.093 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.3 | 36/36 | 100% (90.3%, 100.0%) | | 8 | Colorectal Adenocarcinoma | NRAS A146V | CDx | 1-1.5x | 0.143 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.223 | 36/36 | 100% (90.3%, 100.0%) | | 9 | Colorectal Adenocarcinoma | NRAS G12A | CDx | 1-1.5x | 0.094 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.189 | 36/36 | 100% (90.3%, 100.0%) | | 10 | Colorectal Adenocarcinoma | NRAS Q61R | CDx | 1-1.5x | 0.153 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.214 | 36/36 | 100% (90.3%, 100.0%) | | 11 | Female Genital Tract Malignancy | MSI-H | CDx | Medium Positive | 58.417 | 36/36 | 100% (90.3%, 100.0%) | | 12 | Lung Non-small cell lung cancer (NSCLC) | EGFR L747_T751del | CDx | 1-1.5x | 0.108 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.195 | 36/36 | 100% (90.3%, 100.0%) | | 13 | Lung Non-small cell lung cancer (NSCLC) | EGFR L858R | CDx | 1-1.5x | 0.089 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.187 | 36/36 | 100% (90.3%, 100.0%) | | 14** | Melanoma | BRAF V600E | CDx | 1-1.5x | 0.078 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.18 | 36/36 | 100% (90.3%, 100.0%) | | 15 | Ovarian Surface Epithelial Carcinomas | MSI-H | CDx | Medium Positive | 106.722 | 36/36 | 100% (90.3%, 100.0%) | PMA P240010: FDA Summary of Safety and Effectiveness Data {30} | Sample | Lineage | Variant | Bio-marker Type | Fold LoD Level*** | Average VAF, MSI Score, or ERBB2 Copies | Number Positive / Number Expected | Agreement (95% CI) | | --- | --- | --- | --- | --- | --- | --- | --- | | 16 | Uterine Neoplasms - Endometrial carcinoma | MSI-H | CDx | Low positive | 47.5 | 35/36 | 97.2% (85.5%, 99.9%) | | 17 | Uterine Serous Carcinoma | MSI-H | CDx | Low Positive | 45.556 | 33/36 | 91.7% (77.5%, 98.2%) | | 18 | Breast Carcinoma | ERBB2 CNA | Tumor Profiling | 2-3x | 11.627 | 36/36 | 100% (90.3%, 100.0% | | 19 | Breast Carcinoma | ERBB2 CNA | Tumor Profiling | 1-1.5x | 7.877 | 35/36 | 97.2% (85.5%, 99.9%) | | 20 | Breast Carcinoma | PALB2 M723fs | Tumor Profiling | 1-1.5x | 0.088 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.183 | 36/36 | 100% (90.3%, 100.0% | | 21 | Breast Carcinoma | SMAD2 L175fs | Tumor Profiling | 1-1.5x | 0.101 | 35/36 | 97.2% (85.5%, 99.9%) | | | | | | 2-3x | 0.196 | 36/36 | 100% (90.3%, 100.0% | | 22 | Colorectal Adenocarcinoma | MLH1 K618del | Tumor Profiling | 1-1.5x | 0.088 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.192 | 36/36 | 100% (90.3%, 100.0% | | 23 | Colorectal Adenocarcinoma | MSH2 E12fs | Tumor Profiling | 1-1.5x | 0.079 | 35/36 | 97.2% (85.5%, 99.9%) | | | | | | 2-3x | 0.201 | 36/36 | 100% (90.3%, 100.0% | | 24 | Colorectal Adenocarcinoma | MSH2 V705fs | Tumor Profiling | 1-1.5x | 0.088 | 34/36 | 94.4% (81.3%, 99.3%) | | | | | | 2-3x | 0.164 | 36/36 | 100% (90.3%, 100.0% | | 25 | Colorectal Adenocarcinoma | MSH6 R1068* | Tumor Profiling | 1-1.5x | 0.105 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.189 | 36/36 | 100% (90.3%, 100.0% | | 26 | Colorectal Adenocarcinoma | PIK3R1 N453_Y463del | Tumor Profiling | 1-1.5x | 0.107 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.198 | 35/35 | 100% (90.0%, 100.0% | | 27 | Colorectal Adenocarcinoma | PMS2 R211* | Tumor Profiling | 1-1.5x | 0.113 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.177 | 36/36 | 100% (90.3%, 100.0% | | 28 | Gastrointestinal Stromal Tumors (GIST) | KIT K550_K558del | Tumor Profiling | 1-1.5x | 0.103 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.158 | 36/36 | 100% (90.3%, 100.0% | | 29 | Gastrointestinal Stromal Tumors (GIST) | NF1 I1679_Y1680del | Tumor Profiling | 1-1.5x | 0.095 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.162 | 36/36 | 100% (90.3%, 100.0% | | 30 | Glioblastoma | TERT c.-146C>T | Tumor Profiling | 1-1.5x | 0.1 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.159 | 36/36 | 100% (90.3%, 100.0% | | 31 | Lung Non-small cell lung cancer (NSCLC) | ARID1A Q1409fs | Tumor Profiling | 1-1.5x | 0.097 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.189 | 36/36 | 100% (90.3%, 100.0% | | 32 | Lung Non-small cell lung cancer (NSCLC) | EGFR A767_V769dup | Tumor Profiling | 1-1.5x | 0.108 | 36/36 | 100% (90.3%, 100.0% | | | | | | 2-3x | 0.211 | 36/36 | 100% (90.3%, 100.0% | PMA P240010: FDA Summary of Safety and Effectiveness Data {31} | Sample | Lineage | Variant | Bio-marker Type | Fold LoD Level*** | Average VAF, MSI Score, or ERBB2 Copies | Number Positive / Number Expected | Agreement (95% CI) | | --- | --- | --- | --- | --- | --- | --- | --- | | 33 | Lung Non-small cell lung cancer (NSCLC) | EGFR T790M | Tumor Profiling | 1-1.5x | 0.115 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.189 | 36/36 | 100% (90.3%, 100.0%) | | 34 | Lung Non-small cell lung cancer (NSCLC) | ERBB2 G776delinsVC | Tumor Profiling | 1-1.5x | 0.118 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.18 | 36/36 | 100% (90.3%, 100.0%) | | 14** | Melanoma | TERT c.-146C>T | Tumor Profiling | 1-1.5x | 0.113 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.275 | 36/36 | 100% (90.3%, 100.0%) | | 35 | Ovarian Surface Epithelial Carcinomas | BRCA1 E23fs | Tumor Profiling | 1-1.5x | 0.103 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.25 | 36/36 | 100% (90.3%, 100.0%) | | 36 | Prostatic Adenocarcinoma | BRCA2 S1882* | Tumor Profiling | 1-1.5x | 0.113 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.197 | 36/36 | 100% (90.3%, 100.0%) | | 37** | Prostatic Adenocarcinoma | BRCA2 T2125fs | Tumor Profiling | 1-1.5x | 0.127 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.177 | 36/36 | 100% (90.3%, 100.0%) | | 37** | Prostatic Adenocarcinoma | CTNNB1 T41A | Tumor Profiling | 1-1.5x | 0.097 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.146 | 36/36 | 100% (90.3%, 100.0%) | | 38 | Small Intestinal Malignancies | APC T1556fs | Tumor Profiling | 1-1.5x | 0.108 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.208 | 36/36 | 100% (90.3%, 100.0%) | | 39 | Small Intestinal Malignancies | KRAS G13D | Tumor Profiling | 1-1.5x | 0.259 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.428 | 36/36 | 100% (90.3%, 100.0%) | | 40 | Soft Tissue Tumors | TP53 Q167* | Tumor Profiling | 1-1.5x | 0.117 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.176 | 36/36 | 100% (90.3%, 100.0%) | | 41 | Uterine Neoplasms - Endometrial carcinoma | ARID1A A330fs | Tumor Profiling | 1-1.5x | 0.163 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.097 | 34/36 | 94.4% (81.3%, 99.3%) | | 42** | Uterine Serous Carcinoma | TP53 F113fs c.332dupT | Tumor Profiling | 1-1.5x | 0.117 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.197 | 36/36 | 100% (90.3%, 100.0%) | | 42** | Uterine Serous Carcinoma | TP53 F113fs c.336_337delC T | Tumor Profiling | 1-1.5x | 0.117 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.198 | 36/36 | 100% (90.3%, 100.0%) | | 43 | Uterine Serous Carcinoma | TP53 N263fs | Tumor Profiling | 1-1.5x | 0.123 | 36/36 | 100% (90.3%, 100.0%) | | | | | | 2-3x | 0.209 | 36/36 | 100% (90.3%, 100.0%) | | 44 | Uterine Neoplasms - Endometrial carcinoma | Not MSI-H | CDx | Low Negative | 21.5 | 36/36 | 100% (90.3%, 100.0%) | PMA P240010: FDA Summary of Safety and Effectiveness Data {32} | Sample | Lineage | Variant | Bio-marker Type | Fold LoD Level*** | Average VAF, MSI Score, or ERBB2 Copies | Number Positive / Number Expected | Agreement (95% CI) | | --- | --- | --- | --- | --- | --- | --- | --- | | 45… --- **Source:** [https://fda.innolitics.com/device/P240010](https://fda.innolitics.com/device/P240010) **Published by [Innolitics](https://innolitics.com)** — a medical-device software consultancy. We help companies design, build, and clear FDA-regulated software and AI/ML devices. If you're preparing [a PMA](https://innolitics.com/services/regulatory/), [a 510(k)](https://innolitics.com/services/510ks/), [a SaMD](https://innolitics.com/services/end-to-end-samd/), [an AI/ML medical device](https://innolitics.com/services/medical-imaging-ai-development/), or [an FDA regulatory strategy](https://innolitics.com/services/regulatory/), [get in touch](https://innolitics.com/contact). **Cite:** Innolitics at https://innolitics.com