Measurable Residual Disease in Adult B Lymphoblastic Leukemia: A Study of Concordance between Multiparametric Flow Cytometry, Next-Generation Sequencing of Immunoglobulin Gene Rearrangements, and Quantitative PCR

Introduction: Measurable residual disease (MRD) is the strongest independent prognostic factor in B lymphoblastic leukemia (B-ALL). The methods used for MRD monitoring include multiparametric flow cytometry (MFC), next-generation sequencing (NGS) of immunoglobulin gene rearrangements, and quantitative PCR (qPCR). In this study, we aim to evaluate the concordance between these three well-defined MRD assessment methods to better characterize the biology of MRD in adult B-ALL patients.

Method: This study involved adult patients aged 19 or older with B-ALL, treated with modified hyper-CVAD (hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone) and allogeneic hematopoietic stem cell transplant (allogeneic-HSCT) at Catholic Hematology Hospital from May 2022 to June 2024. A total of 646 bone marrow aspirates were collected from 182 patients (96 Ph-negative ALL, 86 Ph-positive ALL) after remission induction [Time-point 1 (TP1), 4-6 weeks], after first [Time-point 2 (TP2), 10-12 weeks] and second consolidation [Time-point 3 (TP3), 16-20 weeks], and at 3-month interval after transplantation. Poor MRD response was defined as > 0.1%, and complete MRD response as < 0.001%, and poor MRD responders were treated with MRD-directed therapy using blinatumomab or next-generation tyrosine kinase inhibitors. MFC-MRD utilized DuraClone RE ALB tubes (Beckman Coulter, Marseille, France), complemented with additional antibodies (CD22, CD66c, CD73, CD123, and CD304). NGS-MRD was performed using the LymphoTrack^® IGH and IGK assay panels (InVivoScribe, San Diego, CA). The qPCR-MRD employed the Real-Q BCR-ABL Quantification Kit (BioSewoom, Seoul, Republic of Korea). All three methods demonstrated a virtual sensitivity of 10^-6.

Results: The study showed strong correlation between the MRD assessment methods: MFC-MRD and NGS-MRD (r = 0.8459, P <0.0001), MFC-MRD and qPCR (r = 0.8470, P <0.0001), and NGS-MRD and qPCR (r = 0.8319, P <0.0001). At a 0.1% MRD threshold, the concordance rates were 94.3%, 85.3%, and 87.0%, respectively, while at a 0.01%, they were 89.6%, 78.6%, and 79.6%. A significant discordance was noted with 49 samples from 25 patients showing qPCR-MRD positivity but MFC and NGS negativity. Among these discordant cases, seven patients had more than three consecutive discordant results, two of which involved p210 (b2a2 and b3a2) and the others p190 (e1a2), suggesting the possibility of lymphoblastic phase of chronic myeloid leukemia (CML) or CML-like B-ALL. We also identified qPCR method could more detect poor MRD responders compared to MFC or NGS in Ph-positive ALL, but the detection power was similar between MFC and NGS in Ph-negative ALL. Our MRD-directed therapy suggested prognostically significant MRD time point was TP3 at 16-20 weeks after treatment, and poor MRD responders (> 0.1%) at TP3 by any detection methods showed significantly poor survival outcome even after allogeneic-HSCT in both Ph-negative and positive ALL. Major MRD responders (0.001%-0.1%) showed similar survival outcome compared to complete MRD responders (< 0.001%) at any time points.

Conclusion: Our data suggested all MRD detection methods showed acceptable power and good concordance rates, but the detection power was different between Ph-positive and Ph-negative ALL. We also suggested MRD-directed therapeutic strategies might predict the significant time point of MRD for the prediction of survival outcomes.