Concordance of two approaches in monitoring of minimal residual disease in B-precursor acute lymphoblastic leukemia: Fusion transcripts and leukemia-associated immunophenotypes

Background/Purpose Real-time quantitative polymerase chain reaction (RQ-PCR) for fusion transcripts and flow cytometry for leukemia-specific markers are widely used for minimal residual disease (MRD) detection in acute lymphoblastic leukemia, but the relation between the results of either method is unclear. Methods Mononucleated cells from 108 bone marrow samples collected from 55 B-precursor acute lymphoblastic leukemia patients (30 with t(12;21)/ETV6-RUNX1, 16 with t(9;22)/BCR-ABL1 and nine with t(1;19)/TCF3-PBX1) were examined in tandem by RQ-PCR and six-color flow cytometry. Results MRD results were concordant in 91 of the 108 paired samples (84.2%; K=0.690); 49 samples were MRD-negative while 42 were MRD-positive by both methods, with < 1 log difference in positive MRD estimates in 39 samples (92.9%). Of the 17 discordant samples, 16 were MRD-positive by RQ-PCR but MRD-negative by flow cytometry; the opposite was true in one sample. Kappa value/concordance was 0.690/85.0% (n = 60) for ETV6-RUNX1, 0.842/93.3% (n = 15) for TCF3-PBX1, and 0.535/78.8% (n = 33) for BCR-ABL1. Specific immunophenotypic abnormalities were more prevalent in each genetic subgroup, such as CD38 underexpression, CD58 overexpression, and CD34 overexpression in ETV6-RUNX1, TCF3-PBX1, and BCR-ABL1, respectively. Conclusion In most follow-up samples, MRD estimates by two methods are in agreement, especially in patients with TCF3-PBX1.