Page 237 - Personalised medicine of fluoropyrimidines using DPYD pharmacogenetics Carin Lunenburg

Confirmation practice in pharmacogenetic testing
proportional to the amount of heterozygotes present. In other words, discrepancies for CYP2D6*4 (allele frequency 23%)30 will be detected much earlier than discrepancies for, e.g. CYP2D6*7 (allele frequency 0.05%).30 In this aspect, the determined discrepancy rate of 0.01% might actually be higher for specific variants. In addition, the tests in this study were mainly executed in patients with a Caucasian ethnic background. As frequencies of genetic variants can vary between different ethnic populations, results could be different in another population.
The large number of genotyping test results is a strength of this study. However, specific allele dropout will depend on the number of samples with a particular variant. The low discrepancy rate shows high concordance and robustness of the methods used. As described before, the consequence of a misclassified genotype can be substantial, resulting in either underdosing or overdosing, sub sequentially leading to inefficacy or, potentially lethal, toxicity (e.g. DPYD genotyping). We expect that next generation sequencing (NGS) might replace some of the current assays in the upcoming years. NGS is also subject to allele dropout as it is PCR based, but possibly less compared to current techniques. This is caused by the fact that NGS has multiple coverage depth of the same variants, thus a failed reaction of one primer will not directly results in a misclassification of the variant.
Differences exist between laboratories in which DPYD variants are genotyped, or they might not genotype for DPYD variants at all. This could have great impact on patient care as DPD phenotypes might be predicted differently between laboratories. The impact could be greater compared to the impact due to differences between laboratories in confirmation practice as quality control of these tests. This also accounts for other variants in other genes, and for the fact that not all associated variants per gene are discovered yet. Besides assay errors, human errors (switch of samples) might also occur. However, this discussion is out of the scope of this paper, were we focus on the dilemma of confirmation practice.
Conclusions
We have shown substantial variability between laboratories in the use of a second confirmatory technique for PGx testing. The risk of a discrepancy may differ between assays and the clinical implications will depend on the gene tested. Therefore we feel that a second, independent technique is useful for genetic tests with a high clinical impact, such as DPYD testing. Guidelines can help to align confirmatory laboratory practices for PGx, however, they may need to be specified per gene and per test.
9
235