Quantitative Next-Generation Sequencing (NGS) Based Monitoring of Minimal Residual Disease (MRD) Predicts Imminent Relapse in Patients with Myeloid Malignancies after Allogeneic Hematopoietic Stem Cell Transplantation

Introduction

The introduction of novel techniques such as next-generation sequencing (NGS) enabled the discovery of several gene mutations in myeloid malignancies during the last years. This has not only improved our understanding of the pathogenesis but also allows refinement of current risk stratification systems including those predicting outcome after allogeneic hematopoietic stem cell transplantation (allo-HSCT). Still, in patients (pts) with myeloid malignancies relapse represents the main cause of treatment failure after allo-HSCT. The success of therapeutic interventions for relapse after allo-HSCT strongly depends on disease burden implicating that early and sensitive relapse detection is essential. NGS generally allows the identification of candidate mutations prior to transplant which can be used to monitor minimal residual disease (MRD) after transplant. So far, NGS is not broadly employed to monitor MRD in pts with myeloid malignancies after allo-HSCT and evidence from literature is restricted to a few reports (Fu et al., 2013). Here we report our real-life experience regarding the use of NGS to quantitatively monitor MRD after allo-HSCT.

Patients and Methods

For this purpose, pts with myeloid malignancies undergoing allo-HSCT at our center between 2012 and 2016 were screened for mutations in the following genes: ASXL1, CBL, CEBPA, DNMT3A, ETV6, EZH2, IDH1, IDH2, SF3B1, SRSF2, KRAS, MPL, NRAS, RUNX1, TET2 and TP53. For this screening prior to transplantation amplicon deep-sequencing was applied to detect mutations. Patients with at least one mutation identified prior to transplant were quantitatively monitored after allo-HSCT for the individual mutations using the Nextera^® XT DNA Kit (Illumina, San Diego, USA). This quantitative NGS assay has a sensitivity of <1%. Results of this NGS-based monitoring were correlated with the clinical courses of the individual pts and compared with the results obtained by other methods such as STR-based chimerism analyses, cytogenetics, XY-FISH and WT1 mRNA expression.

Results

A total of 26 pts (11 female, 15 male, median age 58 years) who received an allo-HSCT from a related (n = 3) or unrelated donor (n = 23) for the treatment of AML (n = 4), MDS (n = 18) or CMML (n = 4) exhibited at least one mutation detectable by NGS pre-transplant and were therefore included into this retrospective analysis. Among those, in 21 pts (81%) one, in 4 pts (15%) two and in one patient (4%) even three mutations were detectable in the investigated genes. The most frequently mutated genes were ASXL1 in 11 pts (42%), followed by EZH2 in 5 (19%) and TP53, TET2 and RUNX1 in 4 pts (15%), respectively. IDH2 and DNMT3A were mutated in two pts (8%) and SRSF2, SF3B1 and CEBPA in one patient (4%) each. The variant allele frequency (VAF) prior to transplant ranged from 6% to 53%. Eleven pts (42%) suffered from a hematological (73%) or molecular (27%) relapse in a median of 85 days (range, 28 - 383) after allo-HSCT. During a median post-transplant follow-up of 28 months (range, 9 - 70) a total of 86 samples were analyzed by NGS corresponding to a median of 3 samples per patient (range, 1 - 7). MRD positivity was defined as presence of the known mutation in at least 1% of reads. According to this definition, in 12 pts (46%) at least one molecular alteration became detectable again after a median of 3 months after allo-HSCT (range, 1 - 13,7) and was associated with a higher relapse incidence compared to pts without evidence of known molecular alterations (83% vs. 7%, p = 0.0002).

Conclusion

In conclusion, this study shows that molecular monitoring by NGS pre-and post-transplantation can identify biomarkers which may help to predict imminent relapse and may be used to guide therapeutic interventions after transplantation. Taking into account the methodical restrictions of this retrospective analysis, this requires confirmation by a well-designed prospective trial.