Table 10. Reproducibility of companion diagnostic alterations.
| Gene or biomarker | Tumor type | Number of unique samples | Alteration | Reproducibility (95% two-sided CIsa) |
|---|---|---|---|---|
| EGFR | NSCLC | 3 | Exon 19 deletion | 100.00% (96.50%, 100.00%) |
| 2 | Exon 21 L858R | 100.00% (94.73%, 100.00%) | ||
| 2 | Exon 20 T790M | 100.00% (94.93%, 100.00%) | ||
| KRAS | CRC | 3 | Codons 12/13 substitution | 100.00% (96.57%, 100.00%) |
| ALK | NSCLC | 3 | Fusion | 100.00% (96.57%, 100.00%) |
| BRAF | Melanoma | 2 | V600E/V600K | 100.00% (94.87%, 100.00%) |
| ERBB2 | Breast cancer | 3 | Amplification | 100.00% (96.44%, 100.00%) |
| BRCA1/2 | Ovarian cancer | 34 | SUBs, INDELs, rearrangements, homozygous deletions | 96.02% (94.49%, 97.14%) |
| PIK3CA | Breast cancer | 3c | E545K/ H1047L/C420R | 100.00% (95.58%, 100.00%) |
| FGFR2 | Cholangiocarcinoma (CCA) | 5 | FGFR2 Fusions and rearrangementsd | 99.28% (96.01%, 99.87%) |
| MET | NSCLC | 8 | SNVs and INDELs that lead to exon 14 skipping | 98.95% (96.25%, 99.87%) |
| HRR genes | Prostate | 46 | SUBs, INDELs, rearrangements, homozygous deletions | 94.2% (92.9%, 95.4%) |
| TMB | Solid tumors | 46 | TMB ≥10 mutations per megabase | 99.72% (99.18%, 99.94%) |
| NTRK1 | Solid tumors | 7f | Fusions | NTRK1: 100.0% (92.5%, 100.0%) |
| NTRK2 | NTRK2: 100.0% (85.2%, 100.0%) | |||
| NTRK3 | NTRK3: 99.2% (95.4%, 100.0%) |
CI = confidence interval; HRR = homologous recombination repair; LoD = limit of detection; PMA = premarket approval; sPMA = supplemental premarket approval; SUB = base substitution; TMB = tumor mutation burden.
a For EGFR, KRAS, ALK, BRAF, ERBB2, BRCA1/2, PIK3CA, and FGFR2, Wilson score method was used for 95% two-sided CIs calculation; for MET, HRR genes, TMB, and NTRK1/2/3, exact method was used for 95% two-sided CIs calculation.
b Six samples were evaluated, in which two samples had low reproducibility for calling BRCA1/2 partial loss. This is expected as samples with small 1–2 exon partial losses were more challenging to detect.
c One sample is from the 46 samples originally included in the PMA precision study. One sample was analyzed in a subsequent precision study for the sPMA. An additional sample was analyzed in the associated PMC precision study. A PIK3CA H1047R alteration was also assessed, but excluded here because the alteration was present at only 0.15x the LoD; the reproducibility of that alteration event was 52.78%.
d The precision study included FGFR2-BICC1, FGFR2-CCDC6 fusion; FGFR2-TFCP2 fusion, and an intron 17 rearrangement (no partner).
e For HRR gene rearrangement, 15 variants were included; five variants exhibited reproducibility <90% due to read counts below reported LoD, which resulted in low reproducibility across all rearrangements of HRR genes.
f The precision study included seven samples with companion diagnostic NTRK1/2/3 fusion positive status: four NTRK3-EVT6 fusions, one NTRK1-TPM3 fusion, one NTRK1-LMNA fusion, and one NTRK2-DSTYK fusion.