xT CDx
TECHNICAL INFORMATION
Tempus AI, Inc.
600 W Chicago Ave Ste #510, Chicago, IL 60654
Phone: (833) 514-4187
Indications For Use
xT CDx is a qualitative Next Generation Sequencing (NGS)-based in vitro diagnostic device intended for use in the detection of substitutions (single nucleotide variants (SNVs) and multi-nucleotide variants (MNVs)) and insertion and deletion alterations (INDELs) in 648 genes, as well as microsatellite instability (MSI) status, using DNA isolated from Formalin-Fixed Paraffin Embedded (FFPE) tumor tissue specimens, and DNA isolated from matched normal blood or saliva specimens, from previously diagnosed cancer patients with solid malignant neoplasms.
The test is intended as a companion diagnostic (CDx) to identify patients who may benefit from treatment with the targeted therapies listed in the Companion Diagnostic Indications table in accordance with the approved therapeutic product labeling.
Additionally, xT CDx is intended to provide tumor mutation profiling to be used by qualified health care professionals in accordance with professional guidelines in oncology for patients with previously diagnosed solid malignant neoplasms. Genomic findings other than those listed in the Companion Diagnostic Indications table are not prescriptive or conclusive for labeled use of any specific therapeutic product.
xT CDx is a single-site assay performed at Tempus AI, Inc., Chicago, IL.
Companion Diagnostic Indications
Tumor Type Biomarker(s) Detected Therapy
Colorectal cancer (CRC) KRAS wild type (absence of mutations in codons 12 or 13) Erbitux
(cetuximab)
(cetuximab)Colorectal cancer (CRC) KRAS wild type (absence of mutations in exons 2, 3, or 4) and NRAS wild type (absence of mutations in exons 2, 3, or 4) Vectibix (panitumumab)
(panitumumab)Contraindications
There are no known contraindications.
PAGE 1 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Limitations
● For in vitro diagnostic use.
● For prescription use only. This test must be ordered by a qualified medical professional in accordance with clinical laboratory regulations.
● The acceptable preparation method for xT CDx tumor specimens is formalin-fixation and paraffin-embedding (FFPE). Other preparations have not been evaluated.
● The test is designed to report out somatic variants and is not intended to report germline variants. xT CDx sequences tumor and patient-matched normal samples to allow personalized subtraction of germline variants from tumor sequencing results.
● xT CDx requires a minimum tumor percentage of 20% for detection of variants, with tumor content enrichment
recommended for specimens with tumor percentage lower than 20%. This assay may not detect variants if the proportion of tumor cells in the sample is less than 20%. xT CDx requires a minimum tumor percentage of 30% in order to determine MSI status.
● Genomic findings other than those listed in the Companion Diagnostic Indications table are not prescriptive or conclusive for labeled use of any specific therapeutic product.
● A negative result does not rule out the presence of a mutation below the limits of detection of the assay.● The clinical validity of the device to guide MSI-related treatment decisions has not been established. MSI status is based on genome-wide analysis of 239 microsatellite loci and is not based on the 5 or 7 MSI loci described in current clinical practice guidelines. The threshold for MSI-H/MSS was determined by analytical concordance to comparator assays (IHC and PCR) using multiple cancer types. An MSI result of Equivocal indicates that microsatellite instability status of MSI-H or MSS could not be determined.
● Performance of xT CDx has not been established for detection of insertions or deletions larger than 25 base pairs.● xT CDx is only approved for use with Tempus pre-qualified Illumina NovaSeq 6000 instruments.
● The test is intended to be performed on specific serial number-controlled instruments by Tempus AI, Inc.
● Decisions on patient care and treatment must be based on the independent medical judgment of the treating physician, taking into consideration all applicable information concerning the patient’s condition, such as patient and family history, physical examinations, information from other diagnostic tests, and patient preferences, in accordance with the standard of care in a given community.
Test Principle
The CDx Assay (xT CDx) is a single site next generation sequencing (NGS) assay. The assay includes reagents, software, instruments, and procedures for testing DNA extracted from formalin-fixed, paraffin embedded (FFPE) tumor specimens and matched normal saliva or blood specimens. The assay employs DNA extraction methods from routinely
obtained FFPE tissue samples and matched normal saliva or blood samples. Extracted DNA undergoes whole-genome shotgun library construction and hybridization-based capture of specified regions from 648 cancer-related genes (including intronic overhangs and selected promoter regions), and 239 loci for MSI. Refer to Table 1 for a complete list of genes included in xT CDx. Using the IlluminaNovaSeq 6000 platform, hybrid-capture-selected libraries are sequenced to highly uniform depth (targeting >500x median coverage of tumor samples, with >95% of exons at >150x coverage and ≥98% of exons at ≥100x coverage). Sequence data is processed using a customized analysis pipeline designed to
detect substitutions (SNVs and MNVs), insertions, and deletions in coding and noncoding genomic regions targeted by the assay. Additionally, MSI status is reported based on a genomic signature.
PAGE 2 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Table 1. xT CDx Gene List
ABCB1 CBFB DNMT3A FGFR2 HLA-G MAGI2 PAK1 RANBP2 SUZ12
ABCC3 CBL DOT1L FGFR3 HNF1A MALT1 PALB2 RARA SYK
ABL1 CBLB DPYD FGFR4 HNF1B MAP2K1 PALLD RASA1 SYNE1
ABL2 CBLC DYNC2H1 FH HOXA11 MAP2K2 PAX3 RB1 TAF1
ABRAXAS1 CBR3 EBF1 FHIT HOXB13 MAP2K4 PAX5 RBM10 TANC1
ACTA2 CCDC6 ECT2L FLCN HRAS MAP3K1 PAX7 RECQL4 TAP1
ACVR1(ALK2) CCND1 EGF FLT1 HSD11B2 MAP3K7 PAX8 RET TAP2
ACVR1B CCND2 EGFR FLT3 HSD3B1 MAPK1 PBRM1 RHEB TARBP2
AGO1 CCND3 EGLN1 FLT4 HSD3B2 MAX PCBP1 RHOA TBC1D12
AJUBA CCNE1 EIF1AX FNTB HSP90AA1 MC1R PDCD1 RICTOR TBL1XR1
AKT1 CD19 ELF3 FOXA1 HSPH1 MCL1 PDCD1LG2 RINT1 TBX3
AKT2 CD22 ELOC(TCEB1) FOXL2 IDH1 MDM2 PDGFRA RIT1 TCF3
AKT3 CD274(PDL1) EMSY FOXO1 IDH2 MDM4 PDGFRB RNF139 TCF7L2
ALK CD40 ENG FOXO3 IDO1 MED12 PDK1 RNF43 TCL1A
AMER1 CD70 EP300 FOXP1 IFIT1 MEF2B PHF6 ROS1 TERT*
APC CD79A EPCAM FOXQ1 IFIT2 MEN1 PHGDH RPL5 TET2
APLNR CD79B EPHA2 FRS2 IFIT3 MET PHLPP1 RPS15 TFE3
APOB CDC73 EPHA7 FUBP1 IFNAR1 MGMT PHLPP2 RPS6KB1 TFEB
AR CDH1 EPHB1 FUS IFNAR2 MIB1 PHOX2B RPTOR TFEC
ARAF CDK12 EPHB2 G6PD IFNGR1 MITF PIAS4 RRM1 TGFBR1
ARHGAP26 CDK4 EPOR GABRA6 IFNGR2 MKI67 PIK3C2B RSF1 TGFBR2
ARHGAP35 CDK6 ERBB2(HER2) GALNT12 IFNL3 MLH1 PIK3CA RUNX1 TIGIT
ARID1A CDK8 ERBB3 GATA1 IKBKE MLH3 PIK3CB RUNX1T1 TMEM127
ARID1B CDKN1A ERBB4 GATA2 IKZF1 MLLT3 PIK3CD RXRA TMEM173
ARID2 CDKN1B ERCC1 GATA3 IL10RA MN1 PIK3CG SCG5 TMPRSS2
ARID5B CDKN1C ERCC2 GATA4 IL15 MPL PIK3R1 SDHA TNF
ASNS CDKN2A ERCC3 GATA6 IL2RA MRE11 PIK3R2 SDHAF2 TNFAIP3
ASPSCR1 CDKN2B ERCC4 GEN1 IL6R MS4A1 PIM1 SDHB TNFRSF14
ASXL1 CDKN2C ERCC5 GLI1 IL7R MSH2 PLCG1 SDHC TNFRSF17
ATIC CEBPA ERCC6 GLI2 ING1 MSH3 PLCG2 SDHD TNFRSF9
ATM CEP57 ERG GNA11 INPP4B MSH6 PML SEC23B TOP1
ATP7B CFTR ERRFI1 GNA13 IRF1 MTAP PMS1 SEMA3C TOP2A
PAGE 3 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
ATR CHD2 ESR1 GNAQ IRF2 MTHFD2 PMS2 SETBP1 TP53
ATRX CHD4 ETS1 GNAS IRF4 MTHFR POLD1 SETD2 TP63
AURKA CHD7 ETS2 GPC3 IRS2 MTOR POLE SF3B1 TPM1
AURKB CHEK1 ETV1 GPS2 ITPKB MTRR POLH SGK1 TPMT
AXIN1 CHEK2 ETV4 GREM1 JAK1 MUTYH POLQ SH2B3 TRAF3
AXIN2 CIC ETV5 GRIN2A JAK2 MYB POT1 SHH TRAF7
AXL CIITA ETV6 GRM3 JAK3 MYC POU2F2 SLC26A3 TSC1
B2M CKS1B EWSR1 GSTP1 JUN MYCL PPARA SLC47A2 TSC2
BAP1 CREBBP EZH2 H19 KAT6A MYCN PPARD SLC9A3R1 TSHR
BARD1 CRKL FAM46C H3F3A KDM5A MYD88 PPARG SLIT2 TUSC3
BCL10 CRLF2 FANCA HAS3 KDM5C MYH11 PPM1D SLX4 TYMS
BCL11B CSF1R FANCB HAVCR2 KDM5D NBN PPP1R15A SMAD2 U2AF1
BCL2 CSF3R FANCC HDAC1 KDM6A NCOR1 PPP2R1A SMAD3 UBE2T
BCL2L1 CTC1 FANCD2 HDAC2 KDR NCOR2 PPP2R2A SMAD4 UGT1A1
BCL2L11 CTCF FANCE HDAC4 KEAP1 NF1 PPP6C SMARCA1 UGT1A9
BCL6 CTLA4 FANCF HGF KEL NF2 PRCC SMARCA4 UMPS
BCL7A CTNNA1 FANCG HIF1A KIF1B NFE2L2 PRDM1 SMARCB1 VEGFA
BCLAF1 CTNNB1 FANCI HIST1H1E KIT NFKBIA PREX2 SMARCE1 VEGFB
BCOR CTRC FANCL HIST1H3B KLF4 NHP2 PRKAR1A SMC1A VHL
BCORL1 CUL1 FANCM HIST1H4E KLHL6 NKX2-1 PRKDC SMC3 VSIR
BCR CUL3 FAS HLA-A KLLN NOP10 PRKN SMO WEE1
BIRC3 CUL4A FAT1 HLA-B KMT2A NOTCH1 PRSS1 SOCS1 WNK1
BLM CUL4B FBXO11 HLA-C KMT2B NOTCH2 PTCH1 SOD2 WNK2
BMPR1A CUX1 FBXW7 HLA-DMA KMT2C NOTCH3 PTCH2 SOX10 WRN
BRAF CXCR4 FCGR2A HLA-DMB KMT2D NOTCH4 PTEN SOX2 WT1
BRCA1 CYLD FCGR3A HLA-DOA KRAS NPM1 PTPN11 SOX9 XPA
BRCA2 CYP1B1 FDPS HLA-DOB L2HGDH NQO1 PTPN13 SPEN XPC
BRD4 CYP2D6 FGF1 HLA-DPA1 LAG3 NRAS PTPN22 SPINK1 XPO1
BRIP1 CYP3A5 FGF10 HLA-DPB1 LATS1 NRG1 PTPRD SPOP XRCC1
BTG1 CYSLTR2 FGF14 HLA-DPB2 LCK NSD1 PTPRT SPRED1 XRCC2
BTK DAXX FGF2 HLA-DQA1 LDLR NSD2 QKI SRC XRCC3
BUB1B DDB2 FGF23 HLA-DQA2 LEF1 NT5C2 RAC1 SRSF2 YEATS4
C11orf65 DDR2 FGF3 HLA-DQB1 LMNA NTHL1 RAD21 STAG2 ZFHX3
C3orf70 DDX3X FGF4 HLA-DQB2 LMO1 NTRK1 RAD50 STAT3 ZMYM3
C8orf34 DICER1 FGF5 HLA-DRA LRP1B NTRK2 RAD51 STAT4 ZNF217
PAGE 4 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
CALR DIRC2 FGF6 HLA-DRB1 LYN NTRK3 RAD51B STAT5A ZNF471
CARD11 DIS3 FGF7 HLA-DRB5 LZTR1 NUDT15 RAD51C STAT5B ZNF620
CARM1 DIS3L2 FGF8 HLA-DRB6 MAD2L2 NUP98 RAD51D STAT6 ZNF750
CASP8 DKC1 FGF9 HLA-E MAF OLIG2 RAD54L STK11 ZNRF3
CASR DNM2 FGFR1 HLA-F MAFB P2RY8 RAF1 SUFU ZRSR2
•
promoter region also sequenced
Summary and Explanation
xT CDx is a companion diagnostic (CDx) test for two therapeutic indications. Information generated by this test is an aid in
the identification of patients who are most likely to benefit from the specific therapeutic products identified in the indications
for use. In addition to use as a CDx, xT CDx identifies cancer-relevant alterations in genes identified in Table 1 that may
inform patient management in accordance with professional guidelines.
xT CDx uses DNA extracted from FFPE tumor tissue, and from patient-matched normal blood or saliva tissue, to perform
whole-genome shotgun library construction and hybridization-based capture followed by uniform and deep sequencing on
Illumina NovaSeq 6000 sequencers qualified by Tempus. Following the sequencing of both the tumor specimen and the
patient-matched normal sample, custom software is used to accurately identify somatic variants in the tumor by filtering out
germline variants identified from a patient’s normal DNA.
This allows identification of tumor-specific genomic biomarkers, including substitutions (single nucleotide variants, SNVs
and multi-nucleotide variants, MNVs), insertion and deletion variants (INDELs); and microsatellite instability (MSI). The
output of xT CDx includes information derived from the FDA-recognized content of OncoKB, Memorial Sloan Kettering
, Memorial Sloan KetteringCancer Center's precision oncology knowledge base (). xT CDx results are presented in three
categories:
Level 1: CDx claims for KRAS and NRAS as noted in the Indications for Use
Level 2: Cancer Mutations with Evidence of Clinical Significance
Level 3: Cancer Mutations with Potential Clinical Significance
The xT CDx Assay includes four critical checks conducted across the assay workflow to closely monitor assay performance
and ensure that only high-quality data are generated and used for biomarker detection. These checks operate at each step
of the assay as follows:
1. DNA Extraction (QC1)
2. Library Preparation (QC2)
3. Hybridization Capture (QC3)
4. Sequencing (QC4)
PAGE 5 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Test Kit Contents
The xT CDx Assay includes specimen collection and shipping kits for each specimen type used with the assay. These kits include specimen preparation instructions, shipping instructions, and a return shipping label.
All other reagents, materials and equipment needed to perform the assay are used exclusively in the Tempus AI Laboratory.
Sample Collection and Test Ordering
To order the xT CDx Assay, a test requisition form must be fully completed and signed by an ordering physician or authorized medical professional. Specimen preparation and mailing instructions are provided in the Specimen Kit.
For more detailed information, including Performance Characteristics, please find the FDA Summary of Safety and Effectiveness Data at: .
Instruments
xT CDx uses Illumina NovaSeq 6000 Sequencers qualified by Tempus, high throughput sequencing systems employing sequencing-by-synthesis chemistry.
Performance Characteristics
Performance characteristics were established using DNA derived from a wide range of FFPE tissue types along with patient-matched normal (blood or saliva) specimens. Studies included CDx variants and cancer types as well as a broad range of representative alteration types, including substitutions (SNVs, MNVs) and INDELs (insertions, deletions) in various genomic contexts across a number of genes. Analysis of the genomic signature for MSI was also performed.
1. Sample Coverage
The sequencing read depth of the device was evaluated by sequencing duplicate libraries from 10 normal diploid samples using worst-case run conditions for detection of somatic alterations. The interlibrary mean coverage (read depth) for all targeted regions across all samples ranged from 508x to 1218x (with an overall mean of 905x). All sequenced libraries had >98% of exons sequenced with a read depth ≥150x. The interlibrary mean coverage for all targeted hotspots ranged from 564x to 1557x (mean of 1042x). The coverage of target regions supports calling of variants by xT CDx at a VAF as low as 3% for substitutions and 5% for INDELs at hotspots, and 5% for substitutions and 10% for INDELs at non-hotspots.
2. Accuracy
The detection of alterations by xT CDx was compared to results of an externally validated orthogonal method (OM). Overall, there were 114 overlapping genes between the two assays. The comparison between SNVs, MNVs, insertions, and deletions detected by xT CDx and the OM included 416 samples representing 31 different tumor types. The distribution of tumor types is provided in Table 2, below.
PAGE 6 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Table 2. Distribution of Cancer Types for Characterization of Tumor Profiling Accuracy
Number of
Cancer Type Samples
Colorectal Cancer 69
Breast Cancer 44
Ovarian Cancer
38 Glioblastoma 34
Non-Small Cell Lung Cancer 29
Endometrial Cancer 26
Clear Cell Renal Cell Carcinoma
22 Bladder Cancer 18
Melanoma 17
Pancreatic Cancer 14
Thyroid Cancer
12 Low Grade Glioma 12
Sarcoma 10
Tumor of Unknown Origin 8
Meningioma
7 Prostate Cancer 7
Gastrointestinal Stromal Tumor 7
Endocrine Tumor 6
Gastric Cancer
5 Head and Neck Squamous Cell Carcinoma 4
Kidney Cancer 3
Brain Cancer 3
Small Cell Lung Cancer
3 Biliary Cancer 3
Cervical Cancer 3
Esophageal Cancer 3
Oropharyngeal Cancer
2 Liver Cancer 2
Head and Neck Cancer 2
Mesothelioma 2
PAGE 7 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Number of
Cancer Type Samples
Adrenal Cancer 1
Concordance was evaluated in both hotspot and non-hotspot regions. PPA and NPA were determined for each variant type
to assess the accuracy of xT CDx tumor profiling. Differences in the number of reportable variants between the two assays
were expected as a result of pipeline-specific variant filtering or germline variant classifications. In particular, the OM only
evaluates tumor samples, whereas xT CDx sequences tumor and patient-matched normal samples to allow personalized
subtraction of germline variants from tumor sequencing results.
Across all samples evaluated, a total of 148 variants reported as somatic by the OM were identified as germline variants by
xT CDx (Table 3). However, because the OM is unable to distinguish germline from somatic variants these were included as
an output of xT CDx for the purposes of this analytical concordance study. A summary of Positive Percent Agreement (PPA)
and Negative Percent Agreement (NPA) is provided in Table 4, below, for substitutions and INDELs.
Table 3. Germline Variants that would be Subtracted by xT CDx but were Classified as Somatic by the Orthogonal Method
Type Number of Variants
Substitutions 139
INDELs 9
All Short Variants 148
Table 4. Concordance for Short Variants (Substitutions and INDELs) Relative to the Orthogonal Method (OM)
Variant Total Unique True False False True PPA NPA
Type Variants Positives Positives Negatives Negatives [Exact 95% CI] [Exact 95% CI]
All Variants 1028 1221 80 11 414920 [98.4%, 99.6%] 99.1%
[100.0%, 100.0%]
100.0%
All SNVs 736 971 19 8 297042 [98.4%, 99.6%] 99.2%
[100.0%, 100.0%]
100.0%
All MNVs 22 18 3 1 8881 [74.0%, 99.9%] 94.7%
[99.9%, 100.0%]
100.0%
All Insertions 71 58 17 2 28656 [88.5%, 99.6%] 96.7%
[100.0%, 100.0%]
100.0%
All Deletions 199 174 41 0 80341 [97.9%, 100.0%] 100.0%
[100.0%, 100.0%]
100.0%
For hotspot concordance analysis with the OM, reported variants in hotspot regions overlapping with OM targeted regions
were analyzed. From the 416 analyzed study samples, 164 samples had at least 1 reported variant in an overlapping hotspot
region. The intersection of the defined hotspot regions of xT CDx and OM targeted regions included 214 total Base Pairs. In
hotspots, a total of 192 reported variants from both assays were evaluated, including 187 substitutions (50 unique SNVs, 3
PAGE 8 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
unique MNVs) across 10 genes, and 5 INDELs (2 unique insertions and 3 unique deletions) across 4 genes. The total variant
counts of each classification across all study samples were used to calculate the PPA and NPA for Substitutions and
INDELS within hotspot regions as metrics to evaluate the accuracy of the device (Table 5).
Table 5. Concordance Summary for Short Variants (Substitutions and INDELs) within Hotspot Regions Relative to the Orthogonal
Method
Variant Total Unique True False False True PPA NPA
Type Variants Positives Positives Negatives Negatives [Exact 95% CI] [Exact 95% CI]
98.9% 100.0%
All Variants 58 188 2 2 23298
[96.2%, 99.9%] [100.0%, 100.0%]
98.9% 100.0%
All SNVs 50 180 2 2 20066
[96.1%, 99.9%] [100.0%, 100.0%]
100.0% 100.0%
All MNVs 3 3 0 0 1212
[29.2%, 100.0%] [99.7%, 100.0%]
100.0% 100.0%
All Insertions 2 2 0 0 808
[15.8%, 100.0%] [99.5%, 100.0%]
100.0% 100.0%
All Deletions 3 3 0 0 1212
[29.2%, 100.0%] [99.7%, 100.0%]
The detection of specific KRAS and NRAS CDx variants in the 69 colorectal cancer samples tested with the OM was
evaluated. Of the 31 CDx variants identified by the OM, 31 were identified by xT CDx, yielding a PPA of 100% (95% CI:
88.8-100.0%). Of the 649 CDx variants identified as negative by the OM, 648 were identified as negative by xT CDx, yielding a
NPA of 99.8% (95% CI: 99.1-100.0%).
The detection of MSI status by xT CDx was assessed by comparison with results obtained using a validated orthogonal
method (IHC staining of MLH1, MSH2, MSH6 and PMS2). A total set of 316 patient-matched tumor and normal samples
representing 30 cancer types were sequenced with xT CDx. The distribution of tumor types is provided in Table 6, below.
Table 6. Distribution of Cancer Types for Characterization of MSI Accuracy
Abnormal IHC Number Normal IHCNumber of
Cancer Type Number of samples
of MSI-H (by IHC) MSS (by IHC)
CRC/EC* 108 75 33
non-CRC/non-EC** 208 42 166
Total 316 117 199
•
colorectal or endometrial cancer
•
* non-colorectal, non-endometrial cancer
The reported MSI status from xT CDx was compared with results of IHC staining and used to calculate the PPA and NPA for
MSI. Of the 117 samples identified as positive by IHC testing, 110 were identified as MSI-H by xT CDx, yielding a PPA of
PAGE 9 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
94.0% (95% CI: 88-98%). Of the 199 samples identified as negative by IHC testing, 195 were identified as MSS by xT CDx,
yielding a NPA of 98% (95% CI: 95-99%) Results of MSI concordance testing are provided in Tables 7 and 8, below.
Table 7. MSI Concordance Between xT CDx and IHC
Type Normal IHC (IHC-) Abnormal IHC (IHC+)
xT CDx MSI Stable (MSS) 195 7
xT CDx MSI High (MSI-H) 4 110
Table 8. Agreement for MSI Status Overall and by Cancer Type
Cancer Type OPA [Exact 95% CI] PPA [Exact 95% CI] NPA [Exact 95% CI]
All 96.5% [94%, 98%] 94.0% [88%, 98%] 98.0% [95%, 99%]
CRC/EC* 96.3% [91%, 99%] 96.0% [89%, 99%] 97.0% [84%, 100%]
non-CRC/non-EC** 96.6% [93%, 99%] 90.5.8% [77%, 97%] 98.2% [95%, 100%]
•
colorectal or endometrial cancer
•
* non-colorectal, non-endometrial cancer
3. Precision
3.1 PRECISION IN WELL-CHARACTERIZED MATERIAL
The panel-wide precision/reproducibility of xT CDx was assessed for detecting SNVs and INDELs in well-characterized
reference material by repeated measurement of NA12878, a nucleic acid (NA) extracted from the GM12878 cell
line. Precision was evaluated across 22 replicates which were processed over multiple library preparation days (n=17),
hybridization capture batches (n=8), and sequencing flow cells (n=8).
A total of 2673 variants were called across all 22 replicates, and 2624 of these variants were in the Genome in a Bottle
(GIAB) high confidence dataset. Table 9 shows the Coefficient of Variation (CV) distribution for all 2673 variants analyzed.
95.5% of samples had a CV below 10%. Across all samples, the mean CV was 3.7% +/- 3.9%. Table 10 shows Mean %CV by
zygosity of the variant, as declared in the GIAB variant call file (VCF) and type variant.
Table 9. Distribution of Variants by %CV in Well-Characterized Reference Material
CV < 10% 10% ≤CV < 15% 15% ≤CV < 20% 20% < CV
Number of Variants 2552 73 24 24
Percent of Variants 95.5% 2.7% 0.9% 0.9%
1 Zook, J. M. et al. Extensive sequencing of seven human genomes to characterize benchmark reference materials. Sci.
Data 3:160025 doi: 10.1038/sdata.2016.25 (2016)
PAGE 10 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Table 10. Mean Percent Coefficient of Variation (%CV) by Zygosity Declared in the GIAB VCF and Type of Variant for
Well-Characterized Reference Material
SNVs and INDELs SNVs Only INDELs Only
Zygosity
(%CV) (%CV) (%CV)
All1 3.7% +/- 3.8% 3.5% +/- 3.5% 7.3% +/- 6.5%
Homozygous Only 0.23% +/- 0.72% 0.14% +/- 0.39% 1.8% +/- 2.1%
Heterozygous Only 5.3% +/- 3.2% 5.3% +/- 3.1% 7.9% +/- 5.6%
1 Homozygous, Heterozygous, and missing (from GIAB VCF)
3.2 PANEL-WIDE PRECISION IN CLINICAL SPECIMENS
Panel-wide precision in clinical specimens was based on repeated measurement of 49 patient specimens representing 23
different tumor types (including melanoma, CRC, glioblastoma, and lung cancer). Replicates (n=5-10) of each specimen
were measured across 3 non-consecutive days, with multiple operators, reagent lots, and instruments. A total of 317
replicates contributed to the evaluation of precision. The distribution of tumor types is provided in Table 11, below.
Table 11. Distribution of Cancer Types for Characterization of Panel-Wide Precision
Cancer Type | Number of Samples |
Basal Cell Carcinoma 1
Bladder Cancer 6
Breast Cancer 4
Colorectal Cancer 5
Endocrine Tumor 2
Endometrial Cancer 4
Esophageal Cancer 1
Gastric Cancer 1
Head and Neck Cancer 2
Liver Cancer 1
Melanoma 2
Meningioma 1
Non-Small Cell Lung Cancer 4
Ovarian Cancer 1
Prostate Cancer 1
Skin Cancer 2
PAGE 11 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Cancer Type Number of Samples
Tumor of Unknown Origin 4
Adrenal Cancer 1
Cervical Cancer 1
Head and Neck Squamous Cell Carcinoma 1
Pancreatic Cancer 1
Sarcoma 2
Small Cell Lung Cancer 1
All 49
Among the specimens evaluated, there were 289 total variants represented by 151 SNVs, 9 MNVs, 26 insertions, and 103
deletions. The overall positive call rate across all precision conditions (days, operators, reagent lots, and instruments) for all
specimens and replicates was 94.5%, and 97.0% for variants with a VAF ≥15%. Results are provided in Table 12.
Table 12. Precision by Variant Type and Variant Allele Fraction (VAF)
Positive Call Rate
Variant Type VAF Threshold (%) Total Variants Mean VAF Range Positive/Total Calls
(2-sided 95% CI)
≥0 151 3.8-84.343 911/944 96.5% (95.1,97.6)
≥5 150 5.388-84.343 907/939 96.6% (95.2,97.7)
SNV
≥10 132 10.418-84.343 841/849 99.1% (98.2,99.6)
≥15 110 15.067-84.343 718/726 98.9% (97.8,99.5)
≥0 9 12.657-58.597 61/61 100.0% (94.1,100)
≥5 9 12.657-58.597 61/61 100.0% (94.1,100)
MNV
≥10 9 12.657-58.597 61/61 100.0% (94.1,100)
≥15 6 15.124-58.597 35/35 100.0% (90.0,100)
≥0 26 11.25-61.114 153/165 92.7% (87.6,96.2)
≥5 26 11.25-61.114 153/165 92.7% (87.6,96.2)
Insertion
≥10 26 11.25-61.114 153/165 92.7% (87.6,96.2)
≥15 23 15.187-61.114 139/145 95.9% (91.2,98.5)
≥0 103 10.054-94.976 683/744 91.8% (89.6,93.7)
≥5 103 10.054-94.976 683/744 91.8% (89.6,93.7)
Deletion
≥10 103 10.054-94.976 683/744 91.8% (89.6,93.7)
≥15 91 15.123-94.976 646/679 95.1% (93.2,96.6)
PAGE 12 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Positive Call Rate
Variant Type VAF Threshold (%) Total Variants Mean VAF Range Positive/Total Calls
(2-sided 95% CI)
≥0 289 3.8-94.976 1808/1914 94.5% (93.3,95.4)
≥5 288 5.388-94.976 1804/1909 94.5% (93.4,95.5)
All
≥10 270 10.054-94.976 1738/1819 95.5% (94.5,96.4)
≥15 230 15.067-94.976 1538/1585 97.0% (96.1,97.8)
3.3 PRECISION FOR DETERMINATION OF MSI STATUS
All 49 unique specimens and 317 replicates were evaluated for MSI precision. Of these, 46/49 (94%) showed a positive call
rate for MSI of 100% across all replicates. The other 3 specimens each had 80% concordance across 5 replicates due to 4
MSS and 1 MSI-H call in each case.
3.4 PRECISION FOR KRAS AND NRAS DETECTION
Precision of detection of alterations associated with CDx claims was evaluated independently of panel-wide precision.
Intra-run (run on same plate under same conditions) and inter-run (run on different plates under different conditions)
conditions were assessed and compared across multiple instruments, reagent lots, days, and operators. 18 different CDx
variants across all relevant exons of each CDx gene were included in the study. Included variants are provided in Table 13.
Table 13. Variants evaluated for Precision of KRAS and NRAS Detection
Gene | Variant | Number of Specimens |
KRAS p.Gly12Ser 1
KRAS p.Gly12Arg 1
KRAS p.Gly12Ala 1
KRAS p.Gly12Cys 2
KRAS p.Gly12Asp 5
KRAS p.Gly12Val 1
KRAS p.Gly13Asp 1
KRAS p.Gly13Cys 1
KRAS p.Ala59Thr 1
KRAS p.GlyGln60GlyLys 1
KRAS p.Gln61Arg 1
KRAS p.Ala146Pro 1
KRAS p.Ala146Thr 1
NRAS p.Gly12Val 2
NRAS p.Gly13Arg 1
PAGE 13 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Gene Variant Number of Specimens
NRAS p.Gln61Leu 1
NRAS p.Gln61His 1
NRAS p.Ala146Val* 1
N/A wild type 4
•
evaluated using a cell line, all other variants were evaluated in clinical specimens
522 total replicates across 26 unique CRC samples, and 24 replicates from one cell line, were evaluated; one clinical sample
included two variants. The overall positive call rate was 99.8% and 25 of the 26 samples had a positive call rate of 100%. No
false positive results were observed across all potential CDx biomarker positions and all replicates (>28,000 positions).
Precision results by variant are shown in Table 14, a summary of results by gene is shown in Table 15.
Table 14. Precision for KRAS and NRAS Detection by Exon and Variant
Gene Exon Variant n True Positive False Negative % Correct Call 95% CI
2 All KRAS Exon 2 242 241 1 99.6 (97.7, 100.0)
2 p.Gly12Ala 18 18 0 100 (81.5, 100)
2 p.Gly12Arg 19 19 0 100 (82.4, 100)
2 p.Gly12Asp 102 101 1 99 (94.7, 100.0)
2 p.Gly12Cys 43 43 0 100 (91.8, 100)
2 p.Gly12Ser 23 23 0 100 (85.2, 100)
2 p.Gly12Val 22 22 0 100 (84.6, 100)
KRAS 2 p.Gly13Asp 15 15 0 100 (78.2, 100)
3 All KRAS Exon 3 60 60 0 100 (94.0, 100)
3 p.Ala59Thr 19 19 0 100 (82.4, 100)
3 p.Gln61Arg 19 19 0 100 (82.4, 100)
3 p.GlyGln60GlyLys 22 22 0 100 (84.6, 100)
4 All KRAS Exon 4 39 39 0 100 (91.0, 100)
4 p.Ala146Pro 20 20 0 100 (83.2, 100)
4 p.Ala146Thr 19 19 0 100 (82.4, 100)
2 All NRAS Exon 2 56 56 0 100 (93.6, 100)
2 p.Gly12Val 39 39 0 100 (91.0, 100)
2 p.Gly13Arg 17 17 0 100 (80.5, 100)
3 All NRAS Exon 3 37 37 0 100 (90.5, 100)
NRAS
3 p.Gln61His 17 17 0 100 (80.5, 100)
3 p.Gln61Leu 20 20 0 100 (83.2, 100)
4 All NRAS Exon 4 24 24 0 100 (85.8, 100)
PAGE 14 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Gene Exon Variant n True Positive False Negative % Correct Call 95% CI
4 p.Ala146Val 24 24 0 100 (85.8, 100)
Table 15. Positive and Negative Percent Agreement for CDx Biomarkers by Gene and Overall
Gene TP FP TN FN Total PPA (95% CI) NPA (95% CI)
KRAS 340 0 14275 1 14616 99.7 (98.4, 100.0) 100.0 (100.0, 100)
NRAS 117 0 14499 0 14616 100.0 (96.9, 100) 100.0 (100.0, 100)
Total 457 0 28774 1 29232 99.8 (98.8, 100.0) 100.0 (100.0, 100)
4. Analytical Sensitivity
4.1 TUMOR PURITY
The minimum tumor purity for detection of CDx variants was determined by evaluating 31 CRC FFPE specimens (and
patient-matched normal tissue) with known CDx biomarkers, ranging in tumor purity from 5% to 50%. All CDx biomarkers
were concordant between xT CDx and results of orthogonal testing for all tumor purities at or above 10%. Macrodissection
(enrichment for tumor content) of specimens below 10% tumor purity enabled successful detection of the CDx biomarkers
in all samples. The minimum recommended tumor purity for detection of CDx variants is 20%, with macrodissection
required for specimens with tumor purity lower than 20%.
4.2 DNA INPUT AND LIMITS OF DETECTION (LOD)
The minimum DNA input needed to detect CDx biomarkers was determined by testing 2 CRC FFPE tumor specimens (with
patient-matched normal specimens) with a previously detected KRAS variant (p.G12D) at six different DNA mass inputs
(37.5 ng, 50 ng, 62.5 ng, 75 ng, 100 ng, 125 ng), with each input level tested in duplicate, for a total of 12 replicates per
specimen. The LOD for CDx biomarker VAF was then assessed by testing minimal acceptable DNA inputs of 50 ng and 100
ng. DNA from 2 CRC FFPE specimens with previously detected CDx biomarkers were serially diluted with DNA isolated from
a known wild-type FFPE specimen to achieve expected VAF as follows: undiluted, 15%, 5%, 2.5%, 1.25%, and 0.63%. For each
specimen, at each DNA input level, 2 replicates of each undiluted sample were processed and analyzed, and 20 replicates
were processed and analyzed at each subsequent dilution level. A total of 198 tumor-normal paired replicates passed all QC
metrics and were used for determination of LOD, with results provided in Table 16.
Table 16. Summary of LOD for CDx Variants
DNA Input LOD VAF % (Hit Rate)* LOD VAF % (Probit)**
50 ng 2.41% 2.25%
100 ng 3.61% 2.30%
PAGE 15 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
•
LOD calculations for CDx variants were based on the hit rate approach, as there were less than three dilution levels between 10-90%. LOD from the hit rate approach was defined as the lowest level with 95% hit rate
•
*LOD calculations for the CDx variants based on the probit approach with 95% probability of detection
Additional samples were evaluated for the assay gene panel to determine the minimum DNA input and LOD for short variants (substitutions and INDELs) and for determination of MSI status. The minimum DNA inputs of 50 ng and 100 ng for short variants were established using 3 tumor-normal paired specimens at five dilution levels per specimen, with each replicate measured in duplicate.
The LOD for short variants was then assessed using minimal acceptable DNA inputs for processing 12 tumor-normal paired samples, representing 8 tumor types, each containing at least one known variant. Tumor DNA including known variants was serially diluted with tumor DNA known to be wild-type for those variants to generate a range of expected mutation allele frequencies. This dilution series was used to establish a preliminary LOD, which was subsequently confirmed by testing replicates of 17 tumor-normal paired samples diluted to achieve expected VAFs for the tested variants at or around the target LOD for each variant type (5% for substitutions and 10% for INDELs; 3% for hotspot substitutions and 5% for hotspot INDELs). The results of the gene panel LOD confirmation for short variants is summarized in Table 17.
Table 17. Summary of Variant Detection Near LoD Allele Fraction
Variant Type Tested VAF Positive Call Rate
Substitution 5% 97.5% (79/81)
Substitution (hotspot) 3% 100% (10/10)
INDEL 10% 100% (87/87)
INDEL (hotspot) 5% 100% (23/23)
Preliminary MSI LOD determination was evaluated in 22 CRC FFPE specimens known to be MSI-H based
on orthogonal method testing. Each tumor specimen was diluted using its matched normal specimen to generate 3 dilution levels simulating tumor purities ranging from 10% to 40%. Specimens were evaluated with minimum DNA mass input into library preparation to identify the minimum tumor purity at which MSI status could be detected. This dilution series was used to establish a preliminary LOD, which was subsequently confirmed in an independent study by testing 5 additional replicates of each specimen at or around the expected tumor purity LOD (30%). Positive agreement of xT CDx MSI-H status was 94.6% (142/150 replicates identified as MSI-H) for samples diluted to achieve a tumor purity at or around 30%.
4.3 LIMIT OF BLANK
The LOB of was established by assessing the frequency of false-positive identification of CDx and tumor profiling biomarkers in 23 FFPE tumors (with patient-matched normal specimens) known to be wild-type for KRAS and NRAS.
Specimens were evaluated with 4 or 5 replicate measures per specimen based on tissue availability. No false-positive variants were detected at a VAF threshold of 3% in 102 replicates of these samples, confirming the LOB. 22 replicates of well-characterized material were evaluated for false positive results at any reportable position; no false positives were detected.
PAGE 16 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
5. Reagent Lot Interchangeability
Reagent lot interchangeability was assessed for CDx variants by testing 4 CRC samples containing alterations in the KRAS
or NRAS gene over 63 replicates using multiple reagent lots in 3, 5, and 8 combinations for library preparation, hybridization
capture, and sequencing reagents, respectively, across all tested specimens. No effect of interchanging reagents lots was
observed for variant detection for KRAS and NRAS CDx biomarkers. In addition, variant detection across the entire gene
panel was assessed in 375 replicates across 52 specimens representing a broad diversity of tumor types sequenced with
multiple reagent lots. Results showed 97.8% positive agreement (2294/2345) and 100% negative agreement for
substitutions and INDELs, and 96.9% positive agreement and 96.2% negative agreement for MSI.
6. Stability
6.1 REAGENT STABILITY
The stability of reagents used in the library preparation, hybridization capture, and sequencing steps for xT CDx were
evaluated using 3 lots of reagents for each assay step, tested at defined time points. Results support the stability of library
preparation and hybridization capture reagents up to 7 months and sequencing reagents up to 5 months.
6.2 SAMPLE STABILITY
6.2.1 EXTRACTED DNA
Stability of DNA was evaluated using specimens extracted with the Tempus xT LDT assay. Samples from 468 unique
clinical tumor specimens and 454 unique clinical normal specimens from 33 different tissues of origin were evaluated. DNA
specimens evaluated were stored at -80°C for either 91-180 days or >210 days. More than 99% of the specimens that had
been stored for longer than 9 months were successfully used to generate libraries with xT CDx. Based on this data, DNA
stored in accordance with internal procedures can be considered stable for up to 9 months.
6.2.2 FFPE SLIDES
FFPE slide stability study was assessed prospectively and by analysis of previously prepared aged slides. For
prospective analysis, results were analyzed from 5 tumor specimens across 4 cancer types with slides stored at room
temperature for 0 days, 15 days, or 30 days, and then processed with xT CDx. 15 variants were detected at all 3 timepoints
tested, as summarized in Table 18.
Table 18. Variants Detected in Tumor Specimens at Each Timepoint
Tumor Type T=0 Variants T=15 Days Concordance T=30 Days Concordance
Ovarian 3 3/3 3/3
Prostate 2 2/2 2/2
Lung 4 4*/4 4/4
Ovarian 2 2/2 2/2
Colorectal 4 4/4 4/4
PAGE 17 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Tumor Type T=0 Variants T=15 Days Concordance T=30 Days Concordance
Total 15 100.0% (15/15) 100.0% (15/15)
•
A variant existed in the T= 15 time point which was below LOD in the T=0 timepoint. The T=15 sample had a VAF of 3.5% and the T=0
sample had a VAF of 2.9%
Analysis of previously prepared aged slides involved analysis of slides from 124 tumor specimens representing 23 tumor
types. Slides were stored for varying durations at room temperature prior to DNA extraction. Stability was assessed by the
number of specimens meeting minimum DNA yield criteria for xT CDx; results are summarized in Table 19.
Table 19. Evaluation of FFPE Slides at QC1 Based on Length of Storage
Number of Specimens with
Months since Slide Preparation Number of Specimens Evaluated
≥50 ng DNA Yield at Extraction
0-3 50 47 (94.0%)
3-6 60 58 (96.7%)
6-18 11 11 (100.0%)
18-82 3 3 (100.0%)
Total 124 119 (96.0%)
6.2.3 FFPE BLOCKS
The stability of FFPE blocks was established by studying 349 FFPE blocks of tumor specimens stored at
room temperature for 1-7 years by evaluating DNA extraction yield. The blocks were grouped into 5 age groups based on
duration of storage since block preparation. More than 95% of the blocks in each age group produced 3x the minimum DNA
yield of 50 ng needed for the device when processed under standard conditions. Results are summarized in table 20.
Table 20. DNA Yield from aged FFPE Blocks
Year of Block Number of Mean % Samples ≥150 ng
Age Group
Preparation Specimens DNA Yield DNA Yield
1 2019 40 4000.5 100.0%
2 2018 22 2792.7 95.5%
3 2016-2017 117 2683.0 99.2%
4 2014-2015 125 2564.5 96.8%
5 2012-2013 45 3646.2 100.0%
PAGE 18 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
6.2.4 BLOOD AND BUFFY COAT STABILITY
Stability of blood and buffy coat samples used as the source of matched normal specimens in xT CDx was established by
collecting blood samples from 6 healthy volunteers. Buffy coat stability was determined by separation of buffy coat from
blood upon receipt of a specimen, with storage of the buffy coat fraction at –20°C for 0, 15, 30, and 60 days, followed by
DNA extraction and processing through xT CDx. Blood stability was determined by storage of whole blood specimens at
room temperature for 0, 5, 10, 15, and 20 days followed by separation of the buffy coat fraction, DNA extraction, and
processing through xT CDx. Concordance was evaluated by comparing results at each time point to results from the day 0
time point. For both blood and buffy coat, somatic variant concordance by matching with a randomly selected tumor
specimen was 100% and germline concordance was >99% at each time point evaluated. These results establish storage of
whole blood at room temperature for up to 20 days, and storage of the buffy coat fraction at –20°C for up to 60 days.
7. Tissue Comparability
A large-scale retrospective analysis was conducted using 6,373 unique tumor specimens across 34 cancer types in order to
establish the comparability of assay performance across tumor tissue types. The dataset for analysis consisted of routine
clinical samples analyzed using the Tempus xT LDT assay, from 06/06/2020 to 10/05/2020. Approximately 89% of samples
were matched to blood and 11% of samples were matched to saliva. xT CDx includes four QC checks conducted across the
assay workflow to closely monitor performance at each step and ensure that only high-quality data are generated and used
for variant detection. The QC checks are as follows: DNA Extraction (QC1), Library Preparation (QC2), Hybridization Capture
(QC3), and Sequencing (QC4). The pass rate for each of these QC steps for each cancer type is summarized in Table 21.
More than 91% of specimens passed the check at each assay step regardless of cancer type, demonstrating that assay
performance of xT CDx is independent of tissue type.
Table 21. Pass Rate at Each Assay Step Across Cancer Types
DNA Extraction Library Preparation Hybridization Sequencing Total
Cancer Type
Pass Rate Pass Rate Capture Pass Rate Pass Rate Samples
Adrenal Cancer 100.0% 100.0% 93.3% 100.0% 15
Biliary Cancer 99.5% 99.5% 96.7% 99.5% 184
Bladder Cancer 99.6% 100.0% 97.7% 99.6% 259
Brain Cancer 100.0% 100.0% 100.0% 100.0% 22
Breast Cancer 99.8% 99.7% 97.3% 99.1% 639
Cervical Cancer 100.0% 100.0% 95.9% 100.0% 49
CRC 100.0% 99.8% 97.8% 98.6% 808
Endocrine Tumor 100.0% 100.0% 94.7% 100.0% 95
Endometrial Cancer 100.0% 100.0% 97.8% 98.9% 184
Esophageal Cancer 99.3% 100.0% 95.9% 99.3% 148
Gastric Cancer 100.0% 100.0% 98.2% 99.1% 109
PAGE 19 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
DNA Extraction Library Preparation Hybridization Sequencing Total
Cancer Type
Pass Rate Pass Rate Capture Pass Rate Pass Rate Samples
Gastrointestinal Stromal Tumor 100.0% 100.0% 96.4% 96.4% 28
Glioblastoma 100.0% 100.0% 99.4% 100.0% 163
Head and Neck Cancer 100.0% 100.0% 97.5% 100.0% 40
Head and Neck Squamous Cell Carcinoma 100.0% 100.0% 96.4% 98.2% 111
Kidney Cancer 99.3% 100.0% 95.9% 100.0% 58
Liver Cancer 100.0% 100.0% 95.0% 100.0% 40
Low Grade Glioma 100.0% 100.0% 100.0% 100.0% 34
Melanoma 99.4% 100.0% 98.8% 98.2% 164
Meningioma 100.0% 100.0% 93.3% 100.0% 45
Mesothelioma 100.0% 100.0% 95.2% 100.0% 21
Non-Small Cell Lung Cancer 99.6% 99.6% 97.3% 98.9% 851
Oropharyngeal Cancer 100.0% 100.0% 100.0% 98.0% 49
Ovarian Cancer 100.0% 100.0% 98.2% 100.0% 326
Pancreatic Cancer 99.3% 99.8% 97.7% 99.1% 432
Peritoneal Cancer 100.0% 100.0% 100.0% 100.0% 10
Prostate Cancer 99.2% 99.4% 96.4% 98.0% 511
Sarcoma 99.7% 99.7% 97.5% 98.1% 317
Skin Cancer 100.0% 100.0% 96.0% 100.0% 50
Small Cell Lung Cancer 100.0% 100.0% 100.0% 100.0% 64
Testicular cancer 100.0% 100.0% 100.0% 100.0% 18
Thyroid Cancer 100.0% 100.0% 98.8% 97.6% 85
Tumor of Unknown Origin 100.0% 99.4% 97.9% 99.1% 332
8. Interference
The robustness of the Tempus xT CDx Assay process was assessed while evaluating human FFPE samples in the presence
of exogenous and endogenous interfering samples. 22 FFPE specimens representing 13 different tumor types and their
matched normal specimens were evaluated. The addition of interfering substances including xylene, ethanol, melanin, and
proteinase K, each at two concentrations, was evaluated to determine if they were impactful to xT CDx and the results were
compared to the control (no interference) condition. 274 data points were analyzed across the four interfering substances,
which were considered non-interfering if the positive agreement for variant detection in the presence and absence of that
substance was >90%. Results are presented in Table 22.
PAGE 20 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Table 22. Interference Study Summary
PPA NPA
Substance Concentration Replicates TP FN FP TN PPA Confidence NPA Confidence
Intervals Intervals
Ethanol 5% 46 412 7 2 9355657 98.30% [96.6, 99.3] 100.00% [100.0, 100.0]
Ethanol 10% 32 277 5 3 6508291 98.20% [95.9, 99.4] 100.00% [100.0, 100.0]
Melanin 0.05 ug/mL 48 360 12 3 9762489 96.80% [94.4, 98.3] 100.00% [100.0, 100.0]
Melanin 0.1 ug/mL 32 239 9 3 6508325 96.40% [93.2, 98.3] 100.00% [100.0, 100.0]
ProK 0.03 mg/mL 32 239 9 8 6508320 96.40% [93.2, 98.3] 100.00% [100.0, 100.0]
ProK 0.05 mg/mL 19 114 6 1 3864346 95.00% [89.4, 98.1] 100.00% [100.0, 100.0]
Xylene 0.000025% 39 314 7 4 7932002 97.80% [95.6, 99.1] 100.00% [100.0, 100.0]
Xylene 0.000050% 26 209 5 3 5288001 97.70% [94.6, 99.2] 100.00% [100.0, 100.0]
Analysis of all four substances on MSI determination showed 100% concordance for MSI calling under all conditions except
for 93.3% concordance for MSS samples tested at 0.05 mg/mL of Proteinase K. Interference of necrotic tissue was
evaluated across 348 CRC specimens with necrotic tissue percentage ranging from <5% to >50%. Equivalent invalid rates
were observed at all necrotic tissue levels evaluated, and only a single clinically discordant result was observed in the
dataset, in a sample with <5% necrotic tissue.
9. Guardbanding
Guardbanding studies were performed to evaluate the performance of xT CDx and the impact of process variation with
regard to the measurement of DNA input at various stages of the workflow. Guardbands were evaluated relative to observed
and measured process variability for Library Construction (LC), Hybrid Capture (HC), and Sequencing (Seq).
For each process, at least 12 unique FFPE specimens were evaluated in duplicate at 6-8 input levels representing inputs
below the minimum and above the maximum recommended input at each assay step. Each of the three guardbanding
experiments demonstrated reliable and robust performance at DNA input levels above and below the range. Results are
summarized in Table 23.
Table 23. Summary of the Success Rate per Process and per Input Level
Process Input Level # of Samples Passing QC
LC 12.5 ng – 0.25x minimum 6/26
LC 25 ng – 0.5x minimum 20/26
PAGE 21 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Process Input Level # of Samples Passing QC
LC 50 ng – 1x minimum 26/26
LC 300 ng – 1x maximum 26/26
LC 375 ng – 1.25x maximum 26/26
LC 450 ng – 1.5 maximum 26/26
HC 43.75 ng - 0.25x minimum 24/24
HC 87.5 ng – 0.5x minimum 24/24
HC 175 ng – 1x minimum 24/24
HC 250 ng – 1x maximum 24/24
HC 312 ng – 1.25x maximum 24/24
HC 375 ng – 1.5x maximum 24/24
Seq 0.25x minimum 15/15
Seq 0.5x minimum 26/26
Seq 0.8x minimum 26/26
Seq 0.9x minimum 32/32
Seq 1x minimum 31/31
Seq 1x maximum 26/26
Seq 1.25x maximum 26/26
Seq 1.5x maximum 32/32
10. Cross-Contamination
10.1 CARRYOVER / CROSS-CONTAMINATION
DNA sample carryover (between plates) and cross-contamination (within plates) during the library preparation and
hybridization capture steps of the xT CDx Assay were assessed. DNA from two FFPE specimens with unique KRAS
genotypes, one with a KRAS alteration and one wild-type for KRAS, were plated in a checkerboard matrix pattern as
alternating positive and negative samples run with 9 total replicates per specimen. Carryover and cross-contamination were
assessed as evidence of germline mutations unique to one specimen being found in the other specimen or as evidence of
the KRAS variant in the wild-type specimen. Across all replicates, the overall percent agreement of germline mutations was
100% indicating no sample carryover or cross-contamination. In addition, the KRAS variant was only detected in the
specimen that was known to have a KRAS variant based on previous LDT results and was not detected in the known KRAS
wild-type specimen. No carryover or cross-contamination was observed.
10.2 INDEX CROSS-CONTAMINATION
xT CDx uses unique dual index adaptors to generate libraries; captured libraries are pooled for sequencing. Index
cross-contamination based on incorrect assignment of reads between samples in a pool, as a result of read misassignment
PAGE 22 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
from index hopping, was assessed across >138 billion reads obtained on 22 flowcells used during xT CDx performance characterization. The probability of read misassignment from dual index hopping ranged from 5.85x10-5to 6.42x10-9, with an average probability across all analyzed flowcells of 1.35x10-5.
11. Hybrid Capture Bait Specificity
Bait specificity was addressed through an assessment of coverage at the base level for targeted regions included in xT CDx in 20 samples. Lack of bait specificity and/or insufficient bait inclusion would result in regions of diminished high quality mapped reads due to the capture of off-target content. The mean coverage for CDx genes (KRAS and NRAS) was >500x, with >95% of reads mapping to these genes having high base quality scores of >30. When assessing panel-wide coverage, within-sample mean coverage for all targeted regions ranged from 508x-1218x (mean of 904.8x), with >98% of exons with a depth of ≥150x and >99% of exons with a depth of ≥100x.
12. DNA Extraction
DNA extraction was assessed by duplicate extraction of 124 tumor specimens representing 22 different tumor types (including melanoma, prostate, lung, GBM, breast, and bladder), using 2 extraction instruments and 3 extraction reagent lots. The average DNA yield and concordance of variant calling across all samples was evaluated. The mean yield across all 248 extractions was 5076.4 ng, significantly higher than the minimum DNA input of 50ng needed for library preparation. Variant concordance was assessed in 68 tumor specimens across 11 tumor types extracted in duplicate. Variant concordance in the duplicate samples with sufficient DNA was 97.0%, shown in Table 24.
Table 24. Somatic Variant Concordance Observed in Duplicate DNA Extractions
Overall
Level 1 Variants Level 2 Variants Level 3 Variants # Concordant # Total 95% CI Concordance
1/1 29/30 193/199 223 230 97.0% (93.8, 98.8)
13. Invalid Rates
A large-scale retrospective analysis was conducted using 4628 unique tumor-normal matched specimens across 41 cancer types in order to establish the invalid rates at each step of the xT CDx workflow for a variety of cancer and specimen types. The dataset for analysis consisted of routine clinical samples analyzed using the Tempus xT LDT assay from 06/01/2020 to 12/08/2020. The samples were subjected to pre-specified retrospective analysis based on thresholds for success at each assay step. Results are presented in Table 25. Of the 4628 tumor-normal paired samples evaluated, 4122 (89.1%) were successfully processed across all steps of the assay.
Table 25. Summary of Invalid Rates at Each QC Step by Specimen Type
Invalids | |
Assay Step | FFPE Blood (n=4054) Saliva (n=574) |
DNA Extraction | 9/4628 (0.19%) 0/4054 (0.00%) 0/574 (0.00%) |
PAGE 23 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Library Preparation | 7/4619 (0.15%) 2/4504 (0.05%) 0/574 (0.00%) |
Hybridization Capture | 116/4612 (2.52%) 104/4052 (2.57%) 14/574 (2.44%) |
Sequencing | 223/4392 (5.08%) 48/3847 (1.25%) 5/545 (0.92%) |
14. Clinical Concordance for KRAS and NRAS
Clinical validity of xT CDx as a CDx used for identifying patients with CRC who may not be eligible for treatment with cetuximab when mutations are detected in KRAS codons 12 or 13 or panitumumab when mutations are detected in exons 2, 3, or 4 of KRAS or NRAS was established by evaluating 412 samples from CRC patients. Samples were not pre-screened to enrich for positive samples. All specimens were assessed for a minimum tumor percentage of 20% based on pathology review and availability of matched-normal tissue. Based on this evaluation, samples from 348 patients were included in the study. All 348 samples were sent for orthogonal testing with two FDA-approved CDx assays used as comparators: (1) the Illumina Praxis Extended RAS Panel (P160038); and (2) the Qiagen therascreen KRAS RGQ PCR Kit (P110027).
Orthogonal testing was conducted in duplicate for each sample, for each comparator method. Concordance of xT CDx with the Illumina Praxis Extended RAS Panel (Praxis comparator device, PCD) was evaluated using a total of 190 samples; those that passed all xT CDx quality control metrics and with two successful measurements with the comparator (PCD1 and PCD2 denote the replicate measurements). Concordance of xT CDx with the Qiagen therascreen KRAS RGQ PCR Kit (therascreen comparator device, TCD) was evaluated using a total of 250 samples; those that passed all xT CDx quality control metrics and with two successful measurements with the comparator (TCD1 and TCD2 denote the replicate measurements). Samples used in the study were not obtained from a clinical trial, and not all samples had demographic data available. Based on samples evaluated for concordance and with available data, the sex, age, and race were similar between the xT CDx concordance study and the clinical studies of the two comparator methods, with a more even distribution of sexes in the xT CDx concordance study relative to the clinical studies of the comparator methods. Specimen characteristics, including tumor percentage, percent necrosis, and variant allele distribution, were similar for specimens in the xT CDx concordance study and in the clinical studies for both comparator methods,
By defining the reference result as the consensus calls between two replicate measurements from each comparator methods, the overall concordance between xT CDx and the Illumina Praxis Extended RAS Panel was 100.00% (190/190), and overall concordance between xT CDx and the Qiagen Therascreen KRAS RGQ PCR Kit was 99.60% (249/250). Results of concordance testing are summarized in Table 26 below.
Table 26. Concordance of CDx Variant Calling with Comparator Methods
PCD1+ | PCD1- | TCD1+ | TCD1- | |
PCD2+ PCD2- | PCD2+ PCD2- | TCD2+ TCD2- | TCD2+ TCD2- | |
xT CDx+ | 82 0 | 0 0 | 87 0 | 0 0 |
xT CDx- | 0 0 | 0 108 | 1 0 | 0 162 |
Non-inferiority analysis demonstrated that the agreement between xT CDx and the Illumina Praxis Extended RAS Panel is non-inferior to the agreement between two replicates of that assay; and that the agreement between xT CDx and the Qiagen Therascreen KRAS RGQ PCR Kit is non-inferior to the agreement between two replicates of that assay.
PAGE 24 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
xT CDx
PHYSICIAN INSERT
For in vitro Diagnostic Use
Genetic Companion Diagnostic (CDx) Test for Targeted Therapy Selection in Colorectal Cancer (CRC)
For the most current information on the association of the biomarker and therapeutic outcomes, refer to the therapeutic labels available at Drugs@FDA on the FDA website.
Tempus xT CDx Intended Use
xT CDx is a qualitative Next Generation Sequencing (NGS)-based in vitro diagnostic device intended for use in the detection of substitutions (single nucleotide variants (SNVs) and multi-nucleotide variants (MNVs)) and insertion and deletion alterations (INDELs) in 648 genes, as well as microsatellite instability (MSI) status, using DNA isolated from Formalin-Fixed Paraffin Embedded (FFPE) tumor tissue specimens, and DNA isolated from matched normal blood or saliva specimens, from previously diagnosed cancer patients with solid malignant neoplasms.
The test is intended as a companion diagnostic (CDx) to identify patients who may benefit from treatment with the targeted therapies listed in the Companion Diagnostic Indications table in accordance with the approved therapeutic product labeling.
Additionally, xT CDx is intended to provide tumor mutation profiling to be used by qualified health care professionals in accordance with professional guidelines in oncology for patients with previously diagnosed solid malignant neoplasms. Genomic findings other than those listed in the Companion Diagnostic Indications table are not prescriptive or conclusive for labeled use of any specific therapeutic product.
xT CDx is a single-site assay performed at Tempus AI, Inc., Chicago, IL.
Companion Diagnostic Indications
Tumor Type Biomarker(s) Detected Therapy
Colorectal cancer (CRC) KRAS wild type (absence of mutations in codons 12 or 13) Erbitux (cetuximab)
(cetuximab)Colorectal cancer (CRC) KRAS wild type (absence of mutations in exons 2, 3, or 4) and NRAS wild type (absence of mutations in exons 2, 3, or 4) Vectibix (panitumumab)
(panitumumab)PAGE 25 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Warnings and Precautions
Biopsy may pose a risk to the patient when archival tissue is not available for use with the assay. The patient’s physician should determine whether the patient is a candidate for biopsy.
Test Limitations
● For in vitro diagnostic use.
● For prescription use only. This test must be ordered by a qualified medical professional in accordance with clinical laboratory regulations.
● The test is designed to report out somatic variants and is not intended to report germline variants. xT CDx sequences tumor and patient-matched normal samples to allow personalized subtraction of germline variants from tumor sequencing results.
● xT CDx requires a minimum tumor percentage of 20% for detection of variants, with tumor content enrichment recommended for specimens with tumor percentage lower than 20%. This assay may not detect variants if the proportion of tumor cells in the sample is less than 20%. xT CDx requires a minimum tumor percentage of 30% in order to determine MSI status.
● Genomic findings other than those listed in the Companion Diagnostic Indications table are not prescriptive or conclusive for labeled use of any specific therapeutic product.
● A negative result does not rule out the presence of a mutation below the limits of detection of the assay.● The clinical validity of the device to guide MSI-related treatment decisions has not been established. MSI status is based on genome-wide analysis of 239 microsatellite loci and is not based on the 5 or 7 MSI loci described in current clinical practice guidelines. The threshold for MSI-H/MSS was determined by analytical concordance to comparator assays (IHC and PCR) using multiple cancer types. An MSI result of Equivocal indicates that microsatellite instability status of MSI-H or MSS could not be determined.
● Performance of xT CDx has not been established for detection of insertions or deletions larger than 25 base pairs.
● Decisions on patient care and treatment must be based on the independent medical judgment of the treating physician, taking into consideration all applicable information concerning the patient’s condition, such as patient and family history, physical examinations, information from other diagnostic tests, and patient preferences, in accordance with the standard of care in a given community.
Explanation of the Tiered Reporting
Genomic findings other than those listed in the Intended Use are not prescriptive or conclusive for labeled use of any specific therapeutic product. Test results should be interpreted in the context of pathological evaluation of tumors, treatment history, clinical findings, and other laboratory data. The test report includes genomic findings reported in the following levels (Table 1).2
2
PAGE 26 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)
Table 1. FDA Levels of Biomarkers
FDA Level of Biomarkers Description
Level 1: Companion Diagnostics CDx biomarkers that provide information that is essential for the safe and effective use of a corresponding therapeutic product, such as a drug.
Such claims are supported by analytical validity of the test for
each specific biomarker and a clinical study establishing either
the link between the result of that test and patient outcomes or
clinical concordance to a previously approved CDx.
For Tempus xT CDx, Level 1 results are reported for CRC patients
who may benefit from treatment with cetuximab due to the
presence of a KRAS wild-type biomarker (the absence of
mutations in codons 12 or 13) or panitumumab due to the
presence of NRAS and KRAS wild-type biomarkers (the absence
of mutations in exons 2, 3, or 4).
Level 2: Cancer Mutations with Evidence of Clinical Significance Biomarkers described as cancer mutations with evidence of clinical significance enable health care professionals to use information about their patients’ tumors in accordance with clinical evidence, such as clinical evidence presented in professional guidelines, as appropriate.
Such claims are supported by a demonstration of analytical
validity (either on the mutation itself or via a representative
approach, when appropriate) and clinical validity (typically based
on publicly available clinical evidence, such as professional
guidelines and/or peer-reviewed publications).
Level 3: Cancer Mutations with Potential Clinical Significance Biomarkers described as cancer mutations with potential clinical significance. These mutations may be informational or used to direct patients towards clinical trials for which they may be eligible.
Such claims are supported by analytical validation, principally
through a representative approach, when appropriate, and clinical
or mechanistic rationale for inclusion in the panel. Such
rationales would include peer-reviewed publications or in vitro
pre-clinical models.
PAGE 27 OF 27 TEMPUS xT CDx—TECHNICAL INFORMATION (08/2024)