No product issue found.This is a sample-specific issue, not related to roche reagents.It is likely the sample¿s poor dna quality that makes the l858r amplification signal so late by the cobas® egfr assay, therefore missing the cutoff for being called positive despite the presence of the mutation.Sequencing results of the exon 20 region show that the s768i mutation is not present; the idylla result may be a false positive.The customer used 2, 5-micron thick slides for dna extraction, but only one 5-micron ffpet slide should be used for testing with the cobas® egfr assay.The dna concentration measured with a spectrophotometer (nanodrop) for the samples was 179.9 ng/l for sample id: (b)(6)(2020) and 154.4 ng/l for sample id: (b)(6) (2022).Roche does not have claims for samples that are stored for 2 years.The cobas® egfr mutation test v2 method sheet states: the stability of ffpet samples (ffpet block) has been verified for up to 12 months after the date of collection when stored at 15°c to 30°c.Five-m ffpet sections mounted on slides may be stored at 15°c to 30°c for up to 60 days.According to the cobas® egfr mutation test v2 method sheet, detection of a mutation is dependent on the number of copies present in the specimen and may be affected by sample integrity, amount of isolated dna, and the presence of interfering substances.To check for any mutations that would block the primer/probe binding site, the samples were requested and obtained for further in-house testing to determine the possible cause of the discrepancy.The eluate from both cobas® egfr extractions was tested with: 1.Cobas® egfr assay to replicate the results 2.Miseq ngs on cobas assay amplicons 3.Miseq ngs on egfr exon regions of eluate dna - internal testing showed that the cobas® egfr internal control (ic) cts were very late (>28) showing that the sample is low-quality dna.- cobas® egfr assay and sequencing results showed l858r amplicon is present but with a late ct, and it missed the cutoff for calling the mutation.No signal was detected in the s768i target.- sequencing results for cobas® amplicons will always show mutation.Sequencing results of the exon 20 and 21 regions show that the mutation is present at: -(b)(6) has (b)(4) l858r mutation and (b)(4) mutation -(b)(6) has (b)(4)l858r mutation and (b)(4) mutation the internal control ct values in both cobas® testing (2020 and 2022) are late across all mmx in comparison to the positive control and other tissue samples in the runs, which suggests that the quality of the sample is not optimal.Although the measured dna concentration was high, the results for l858r are outside of the predefined range and not detectable for the s768i mutation.The same pattern was observed during in-house investigative testing.For testing with idylla, the customer extracted a new ffpet slide from the same block.Two different slides can contain different levels of mutations due to heterogeneity of the biopsy sample (block) therefore discrepancy can be observed during testing.In addition, cobas® and idylla tests are two different technologies.Their sensitivities and specificities claims can differ.Idylla detected l858r, which was out of range in cobas® testing, and detected low levels of the (b)(4) mutation.The ngs and pyrosequencing were performed on the eluate from cobas® egfr testing.Ngs detected the l858r, but not the (b)(4).Pyrosequencing also detected the l858r mutation; the (b)(4) was not tested.
|