Performance Characteristics of a Real-Time Polymerase Chain Reaction Assay for the Detection of Epidermal Growth Factor Receptor (EGFR) Mutations in Plasma Samples of Non-Small Cell Lung Cancer (NSCLC) Patients.
Author(s): O'Donnell P, May T, DeMartin K, Ferguson J, Halait H, Wei W, Yu K, Scudder S
Publication: Mol Diagn Ther, 2020, Vol. 24, Page 451-460
PubMed ID: 32406048 PubMed Review Paper? No
Purpose of Paper
This paper compared the detection of epidermal growth factor receptor (EGFR) mutations in the plasma of non-small cell lung cancer (NSCLC) patients using the real-time PCR-based Cobas assay and next-generation sequencing (NGS). The Cobas limit of detection, linearity, and potential interference by endogenous substances was also investigated by dilution of known amounts of mutation-containing DNA in plasma.
Conclusion of Paper
The positive percent agreement, negative percent agreement, and overall percent agreement for mutation identification between the Cobas and NGS assays were 80%, 94.9%, and 87.8%, respectively. Eight of the nine discordant cases arose from mutations that were detected below the validation specification by NGS but detected by Cobas. Dilution of intact and sheared cell line DNA revealed the Cobas assay had a limit of detection of 20-100 copies/mL with linear ranges for predominant and non-predominant mutations of 10-1000 copies/mL and 100-5000 copies/mL, respectively. Although interference with the Cobas assay was observed when 0.20 g/dL hemoglobin was added to plasma spiked with a known number of copies of DNA harboring EGFR mutations, no interference was observed from 0.15 g/dL hemoglobin or any of the other substances investigated.
Studies
-
Study Purpose
This study compared the detection of EGFR mutations using the real-time PCR-based Cobas assay and NGS in the plasma of NSCLC patients. The Cobas assay limit of detection, linearity, and potential interference by endogenous substances was also investigated by dilution of known amounts of cell line DNA harboring EGFR mutations in plasma. K2EDTA plasma from NSCLC patients with (35 patients) and without (39 patients) EGFR mutations was purchased. DNA was extracted from plasma using the Cobas cfDNA Sample Preparation Kit. Mutation status was detected by real-time PCR using the Cobas z 480 analyzer without prior quantification. Mutation status was confirmed by NGS using a proprietary method on an Illumina MiSeq and a median read depth of 50,000–100,000. The limit of detection of the Cobas assay was investigated by dilution of intact and sheared (220 bp) cell line DNA. The effect of interfering substances on the Cobas assay was investigated by adding hemoglobin (0.15 g/dL and 0.20 g/dL), bilirubin (0.2 g/L conjugated and unconjugated), triglycerides (33 g/L), high albumin (>6 g/dL versus 3.8–4.0 g/L), EDTA (3.8 and 9 mg/mL), neupogen (147 ng/mL), erlotinib (90 mg/L), and Staphylococcus epidermidis (1.0 x106 colony-forming units/mL) to EGFR wildtype plasma spiked with known copies of the cell line DNA harboring EGFR mutations. Clinical reproducibility was investigated by combining sheared cell line DNA containing different mutations to create four double mutation positive combinations and eight single mutations which, along with wildtype specimens, were tested in duplicate daily on 3 non-consecutive days.
Summary of Findings:
Of the 35 patients with EGFR mutations detected by NGS, 28 had the same mutation detected by the Cobas assay, six had a different mutation detected, and one had no mutation detected. Of the 39 patients with no mutation detected by NGS, 37 also had no mutation detected by the Cobas assay but a mutation was detected by Cobas assay in the remaining two patients. Thus the positive percent agreement, negative percent agreement, and overall percent agreement were 80%, 94.9%, and 87.8%, respectively. Further investigation found the six cases with discordant mutations all had a second mutation identified by Cobas that was detected below the validation specification by NGS. In the two patients with a mutation detected by Cobas but not NGS, further analysis showed the mutation was detected by NGS but was also below the validation specification. Dilution of intact and sheared cell line DNA revealed a limit of detection of 20-100 copies/mL for each of the investigated mutations using the Cobas assay. Importantly, the linearity range of the Cobas assay for predominant and non-predominant mutations were 10-1000 copies/mL and 100-5000 copies/mL, respectively. Interference with the Cobas assay was observed when 0.20 g/dL hemoglobin was added to plasma spiked with a known number of copies of EGFR mutant DNA but no interference was observed from 0.15 g/dL hemoglobin or any of the other substances investigated. The Cobas assay was highly reproducible using contrived specimens with 100% agreement for wildtype specimens and agreement of 90.3% for G719X due to missed calls and 98.6-100% for the other mutations/combinations of mutations.
Biospecimens
Preservative Types
- Frozen
Diagnoses:
- Neoplastic - Carcinoma
Platform:
Analyte Technology Platform DNA Next generation sequencing DNA Real-time qPCR Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Real-time qPCR Specific Technology platform Real-time PCR based Cobas EGFR Mutation Test v2
Proprietary NGS assay
Biospecimen Aliquots and Components Biospecimen components Spiked with hemoglobin
Spiked with bilirubin
Spiked with triglycerides
High albumin (>6 g/dL)
Normal albumin (3.8-4.0 g/dL)
Spiked with EDTA
Spiked with neupogen
Spiked with erlotinib
Spiked with Staphylococcus epidermidis