Next Generation Sequencing (NGS)-Based Post-Transplant Monitoring in Acute Myeloid Leukemia Patients Receiving Allogeneic Hematopoietic Stem Cell Transplantation (HCT)

Objective

To assess applicability of serial next generation sequencing (NGS) in monitoring acute myeloid leukemia (AML) patients after allogeneic hematopoietic stem cell transplantation (HCT).

Introduction

Clinical application of NGS in allogeneic HCT is a topic of interest. Post-HCT mutation dynamics using longitudinal NGS have not been thoroughly examined. We hypothesized that serial sequencing of pre- and post-HCT using NGS could provide a clinically relevant value in AML patients receiving HCT. The present study aimed to evaluate mutation dynamics in AML using serial samples from pre- and post-HCT with respect to transplant outcomes.

Patients and Methods

104 AML patients were enrolled and sequenced using an Illumina HiSeq 2000 sequencer (84 myeloid custom gene panel) on 529 bone marrow/peripheral blood samples at diagnosis (n=104), pre-HCT (n=104), 21 days post-HCT (n=104), at 6 (n= 38), 12 (n= 20), beyond 12 months (n=4) and at relapse (n=20). Two patients relapsed by day 21. T-cell (n=80) and donor samples (n=57) were also sequenced. All computational and statistical analyses were performed using Python and R.

Results

The average on-target coverage in 529 samples was 1725.6x. In total, we detected 256 mutations at diagnosis in 90/104 patients (86.5%, median 3 mutations per patient). DNMT3A ( n=26, 25.0%), NPM1 ( n=23, 22.1%), CEBPA ( n=15, 14.4%), IDH2 ( n=15, 14.4%), FLT3 (n=14, 13.5%), PTPN11 (n=11, 10.6%) , and NRAS (n=11, 10.6%) were commonly mutated at diagnosis (Figure A). When grouped by 8 defined biological pathways, activated signaling (n=45, 43.2%), DNA methylation (n=44, 42.3%), NPM1 (n=23, 22.1%), and myeloid transcription factors (n=21, 20.2%) were mutated in more than 20% of the cohort (Figure B). Assessment of mutation dynamics in serial samples revealed that mutations from non-relapsed patients show rapid clearance after HCT without acquisition of somatic mutations and that allelic burden is maintained at very low level in longer follow-up period (less than 0.1%) (Figure C). On the other hand, relapsed patients tend to carry higher allelic burden even at day 21 after HCT (2.9% as opposed to 0.04% in non-relapsed patients), which suggests that allelic burden throughout the course of the disease and treatment can possess predictive power for detecting early relapse and predicting long-term overall survival (Figure D). We then assessed whether the mutation status at pre- and post-HCT has any impact on overall survival (OS) and relapse after HCT. With median follow-up of 6.3 years (range 0.7 - 12.3 years), patients with overall variant allele frequency (VAF) ≥ 0.2% at day 21 after HCT showed worse OS (Hazard Ratio (HR) 2.62 [1.28-5.35], p= 0.008) as well as increased risk of relapse (HR 6.16 [2.57-14.73], p < 0.001) (Figure E and F). Multivariate analyses verified that VAF ≥ 0.2% at day 21 was an adverse prognostic factor independent of age and cytogenetics risk group for OS (HR 3.06, [1.46-6.39], p = 0.003). Non-relapse mortality did not show significant difference regardless of VAFs ≥ 0.2% at day 21 (p = 0.21). Overall, 23 mutations from 16 patients were detected at day 21 after HCT (VAFs ≥ 0.2%) where 22 mutations were detected either at diagnosis or relapse, or both. In addition, mutations in DNMT3A (n=10) and NPM1 (n=3) were found in one than one patients at day 21 after HCT (13/23 mutations, 56%).

Conclusion and summary

This study revealed that the allelic burden immediately after HCT, within 21 days, with low VAF (0.2%) can be used to predict relapse after HCT as well as maintenance of nearly cleared allelic burden was observed in samples taken at longer follow-up period (mostly 6 and 12 months) from non-relapsed patients (less than 0.1%). Altogether, our data and analyses illustrated the value of longitudinal NGS-based monitoring strategies for AML patients after allogeneic HCT.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal