Leukemia stem cells for relapse prediction in AML patients receiving allografts: long-term follow-up of a prospective study.

In this study, we explored the ability of leukemia stem cell (LSC)-based method and traditional multiparameter flow cytometry (MFC) assay to predict leukemia relapse after long-term follow-up. 360 AML patients who received allografts between July 2018 and November 2019 were prospectively enrolled. Patients with positive measurable residual disease (MRD) based on CD34CD38cocktail LSCs (≥0.004%) exhibited a greater 5-year cumulative incidence of relapse (CIR) (49.7% vs. 8.5%, P < 0.001), inferior leukemia-free survival (LFS) (48.2% vs. 84.4%, P < 0.001) and inferior overall survival (OS) (59.7% vs. 82.8%, P < 0.001), than did patients without CD34CD38cocktail LSCs (<0.004%). Patients with detectable traditional MFC-MRD exhibited a greater CIR than patients without MRD (45.8% vs. 10.9%, P < 0.001), thereby leading to decreased LFS (54.2% vs. 81.9%, P < 0.001) and decreased OS (56.0% vs. 85.9%, P = 0.001). Compared with traditional MFC-MRD, LSCs-based MRD assay demonstrated high sensitivity (52.4% vs. 33.3%), high C-index (0.72 vs. 0.65) and high Youden index (0.44 vs. 0.27). The median time from LSCs positivity to relapse was longer than the median time from traditional MRD positivity to relapse (144 days vs. 65 days, P = 0.012). Our data confirmed the superiority of the LSC-based MRD assay compared to traditional MFC MRD methods in AML patients who received allografts after long-term follow-up.