Clinical applications of measurable residual disease monitoring by multi-parametric flow cytometry and next-generation sequencing in B-cell precursor acute lymphoblastic leukemia.

Abstract

121 Background: Minimal residual disease (MRD) is accepted as the strongest independent prognostic factor in B-cell precursor acute lymphoblastic leukemia (BCP-ALL). The most frequently used methods for MRD monitoring include multi-parametric flow cytometry (MFC) and next-generation sequencing (NGS)-based IGH/IGK gene rearrangement. Both methods have been described to reach a low sensitivity threshold and be broadly applicable to most patients with BCP-ALL, although the clinical MRD cut-off of the two methods has not been fully defined. In this study, we compared MFC-MRD and NGS-MRD and established their clinical cut-off in BCP-ALL. Methods: Bone marrow aspirates with less than 5% blasts were obtained from 117 treated BCP-ALL patients. MFC-MRD assessment was performed using DuraClone RE ALB tubes (Beckman Coulter, Marseille, France) modified by adding 5 additional antibodies. In total, up to 1×105 cells were acquired using DxFLEX flow cytometer (Beckman Coulter), and all acquired data were analyzed using the Kaluza software v2.1. NGS-MRD was performed using the LymphoTrack IGH and IGK assay panels (InVivoScribe, San Diego, CA). We amplified 240 ng of high-quality DNA with a single multiplex master mix containing primers, and the library DNA was denatured and loaded onto a Miseq reagent kit v2. The FASTQ files were analyzed using the LymphoTrack MRD software v2.0.2. MRD evaluated using both of MFC and NGS in each sample, and the virtual sensitivity of both methods was 10-5. For 84 patients available, we analyzed the relapse-free survival (RFS) and overall survival (OS) rate for each MRD status and established a clinical cutoff. Results: We found a high correlation between methods (r=0.640) and a concordance rate of 81.2% for detecting MRD at 0.01% level. An MRD value of 0.01% proved to be a significant threshold value for 1-year RFS and OS. There was a difference in prognosis in both RFS and OS, but statistical significance was found only in RFS (MFC-MRD: 93.4±3.2% for MRD <0.01% vs. 37.5±27.3% for MRD ≥0.01%, P=0.007; NGS-MRD: 93.7±3.6% for MRD <0.01% vs. 68.8±14.2% for MRD ≥0.01%, P=0.016). In a Kaplan-Meier estimate of RFS at a threshold of 0.01%, patients who were negative for MRD by MFC and positive by NGS (n=12) had an intermediate prognosis between patients who were negative by both MFC and NGS (n=55) and patients who were positive by both MFC and NGS (n=11) (98.14±3.2%, 83.34±10.8%, 40.94±29.5%, respectively, P=0.017). Conclusions: An MRD threshold of 0.01% for both MFC and NGS is reasonable for clinical identification of poor-risk patients. NGS-MRD detection improves risk stratification for BCP-ALL.