Prognostic Impacts of Next-Generation Sequencing-Based Measurable Residual Disease Monitoring before and after Allogeneic Hematopoietic Cell Transplantation in AML

Only a few studies investigated prognostic values of next-generation sequencing-based measurable residual disease (NGS-MRD) detection in AML before or after allogeneic hematopoietic stem cell transplantation (Allo-HSCT). Moreover, NGS-MRD assessments were mostly performed only at pre-HSCT, and results were discordant.

The current study longitudinally collected samples before and after Allo-HSCT and clinical data from two independent prospective cohorts (n=132) registered at ClinicalTrials.gov (NCT01751997) and CRIS (Clinical Research Information Service, KCT0002261), and investigated the role of NGS-MRD assessment. Donor groups consisted of matched sibling (23%), matched-unrelated (35%), and haploidentical familial donors (42%). Enrolled patients received myeloablative (MAC; 44%) or reduced-intensity conditioning (RIC; 56%). Customized amplicon-based targeted NGS including 67 genes (41 entire coding regions and 26 hot spots) was used and mean coverage was over 2000.

With 33 months of median follow-up of survivors, 24 patients experienced post-transplant relapse. Persistent mutations were detectable at pre-HSCT (57/132, 43%) and at 1 month (23/114, 20%) after HSCT. Mutant allelic burdens at pre-HSCT and at 1 month after HSCT in relapsed patients were higher than in non-relapsed patients. Any persistent mutations at pre-HSCT and at 1 month after HSCT were significantly associated with post-transplant relapse (34.8% vs. 6.7%, p<0.001 at pre-HSCT; 43.5% vs. 13.0%, p<0.0001 at 1 month after HSCT) and worse overall survival (54.4% vs. 78.7%, p=0.010 at pre-HSCT; 44.9% vs. 76.8%, p=0.002 at 1 month after HSCT). NGS-MRD positivity was determined as complete mutational clearance by comparisons with various cutoffs of variant allele frequencies. Multivariate analysis confirmed that MRD positivity was an adverse prognostic factor for relapse and overall survival. Of note, optimal time points of NGS-MRD assay were different according to conditioning intensity. NGS-MRD detection at pre-HSCT was significantly associated with higher relapse in those who received MAC, while NGS-MRD detection at 1 month after HSCT was in those who received RIC. We also found that MRD positivity in genes related with clonal hematopoiesis were also significantly associated with post-transplant relapse. Serial NGS-MRD monitoring after HSCT revealed that most residual clones of false positive patients at pre-HSCT and at 1 month after HSCT were disappeared within 3 months after HSCT. At relapse, NGS showed not only clonal expansion or reappearing but also evolution of new clones.

The current study demonstrated that NGS-MRD assessment both at pre-HSCT and at 1 month after HSCT were useful for predicting post-transplant relapse and there were no differences according to type of mutations. Optimal time points of NGS-MRD assessment depend on conditioning intensity.