Robustness of amplicon deep sequencing underlines its utility in clinical applications

We investigated the robustness of amplicon deep sequencing to study its utility in routine clinical applications offering patient-specific individualized assays for molecular disease characterization and monitoring. Amplicons were designed targeting RUNX1, CEBPA, CBL, NRAS, KRAS, DNMT3A, EZH2, and TP53 using different PCR amplification strategies and Roche GS FLX Titanium and Illumina MiSeq sequencing platforms. Thirty-three patients with leukemia were selected as an exemplary cohort representing heterogeneous cancer specimens. Both standard two-primer amplification and four-primer microfluidics PCRs yielded highly linear characteristics in detecting molecular alterations in series of dilution experiments. By fitting a linear mixed-effects model to the logarithmized data, a slope β of −1.000 (95% CI, ±0.046) was obtained for two-primer assays and of −0.998 (95% CI, ±0.105) was obtained for four-primer assays, which represented a near-perfect decrease of the mutation load. Furthermore, data are presented on technical precision, limit of detection, and occurrence of small subclones in TP53- and RUNX1-mutated patients to identify clonal disease progression and residual disease. We demonstrate that, depending on the local sequence context for each amplicon, the limit of detection of the assay cannot be lower than a range of 0.25% to 3.5%. In conclusion, amplicon deep sequencing enabled the assessment of mutations in a highly robust manner and across a broad range of relative frequencies of mutations. This assay detects residual disease or identifies mutations with predictive relevance to direct treatment strategies.