Best Practices for Validation of Measurable Residual Disease Assessments By Multiparameter Flow Cytometry in Emerging Clinical Trials of Acute Myeloid Leukemia

Background: Measurable Residual Disease (MRD) assessments are gaining increasing acceptance as a prognostic factor for tailoring treatment in hematological malignancies. Acute Myeloid Leukemia (AML) is a heterogeneous disease with high relapse rates and presents a high unmet need for effective treatment options. Measurement of residual disease after therapy reflects a combination of all resistance mechanisms and is currently used for guiding treatment options.

Study Design: In this study, we aimed to validate an AML-MRD assay by multiparameter flow cytometry (MFC) methodology. This is a 4-tube, 8-parameter assay designed to incorporate cell differentiation (CD) markers for identification of a diverse group (covering roughly 90% of patients, Cloos et al, 2018) of Leukemia Associated Immunophenotypes (LAIPs) to accurately identify both native phenotypes and phenotype shifts after drug treatment. These CD markers were selected based on extensive investigation of many markers and in line with the consensus recommendations from European Leukemia Network AML working party (Schuurhuis et al, 2018), while specimen testing and interpretation principles were performed in accordance with Cloos et al, 2018. The assay validation focused on evaluation of sensitivity (MRD cut point and LOD), precision and accuracy as key criteria for evaluating assay performance utilizing primary patient specimens and AML cell lines representing different LAIPs. The results were orthogonally verified in a blinded manner by morphologic assessment at Navigate and by the MRD-team at VUMC Amsterdam.

Results: Two experimental approaches were adopted to evaluate analytical and functional sensitivity (clinical applicability) of the assay. Results indicated analytical sensitivity (LOD) as low as 0.01% LAIPs of total WBC and functional sensitivity (LOQ) of 0.1% (MRD cut point). Excellent repeatability and reproducibility (less than 20% CV) was observed across instruments, operators and independent measurements (n = 75). The frequencies of AML blasts detected by MFC and morphological examination were highly concordant (Spearman r = 0.95, P value < 0.001, n = 24). LAIPs deduced across nine patient specimens by the Navigate laboratory were independently confirmed by the MRD-team at VUMC Amsterdam.

Conclusion: In summary, based on the use of consensus markers recommended by ELN for reliable capture of a broad group of LAIPs in AML patients and verification of key assay performance characteristics, we believe this comprehensive MFC based AML MRD assay is fit-for-purpose for accurately assessing measurable residual disease. Following clinical trial validation, MRD might be used as a surrogate endpoint for approval of emerging agents.