NGS-Assisted DNA-Based Digital qPCR Facilitates Stratification Of CML Patients In Long-Term Molecular Remission Based On The Presence Of Detectable BCR-ABL1 DNA

Recent studies indicate that 40% of CML patients on imatinib who achieve complete molecular remission (CMR; equivalent to MR⁴ or better in current terminology) remain disease-free after drug discontinuation, raising the possibility of an “operational cure”. However, the safe introduction of TKI withdrawal into clinical practice would require a method of evaluating risk of relapse, which is likely to be related to the presence of residual disease undetectable by RT-qPCR. A PCR method based on the use of genomic DNA has been shown to be more sensitive for the detection of residual disease than one that relies on cDNA and may, therefore, help to predict outcome post-withdrawal. However, the former method is arduous requiring a customised patient-specific assay. We describe here a method based on targeted next-generation sequencing (NGS) which allows the rapid identification of BCR-ABL1 breakpoints and the generation of DNA-based qPCR assays. The location of the BCR-ABL1 fusion junction was mapped in disease samples from 32 CML patients using Illumina's MiSeq platform. A custom TruSeq DNA target enrichment system (Illumina) was used to enrich for fragments containing sequences from the BCR and ABL1 genes using probes covering both genes including an additional 50kb of sequence in both 5’ and 3’ directions. All 32 patients’ t(9;22) translocation junctions were successfully mapped using a custom designed bioinformatics algorithm. To date we have designed and validated DNA qPCR assays for 26 of these.

On testing consecutive clinical samples from all 26 patients after achievement of MR⁴ or better (previously defined as CMR; mean 6 samples per patient), the DNA-qPCR assays detected residual disease in 8 out of 26 patients (31%), demonstrating that a DNA-based method is capable of identifying the presence of residual leukaemia in a proportion of patients who would be defined as “disease-free” using conventional RNA-based methodology.

One area in which inaccuracy may be introduced into the DNA qPCR technique is in the use of a standard curve generated using serial dilutions of patient’s diagnostic material. We therefore sought to further enhance the sensitivity of a DNA-based approach by optimising this technique for use on a digital quantitative PCR (dqPCR) platform, which provides absolute molecular quantification without the need for standard curve. Using DNA dqPCR on a Fludigm BioMark HD instrument, we measured the level of BCR-ABL1 in 48 samples from 6 CML patients after achievement of a sustained response of MR⁴ or better. At the time point at which patients were first determined to have achieved at least MR⁴, all samples tested positive by DNA dqPCR. However, in subsequent samples collected two years following achievement of at least an MR⁴ (the point of entry into ongoing withdrawal trials, at which point all six patients had achieved MR^4.5 or better), we found that patients fell into two distinct categories: 1) those whose disease continued to be detectable by DNA dqPCR (n=3), and 2) those whose disease had fallen to a level that was no longer detectable by DNA dqPCR (n=3). In contrast the conventional BCR-ABL1 DNA qPCR method detected disease in only one of these six patients at this time point, suggesting superior sensitivity of a digital PCR-based screening platform.

From this pilot study we conclude that NGS-facilitated DNA dqPCR may be used to stratify patients with RT-qPCR-undetectable disease according to residual disease burden and may therefore prove valuable in the identification of patients for whom TKI therapy could be safely reduced or stopped.