The Laip-Based-Dfn Approach Is Superior in Terms of Useful MRD Results As Compared to the Laip Approach after Cycle II in Acute Myeloid Leukemia

Clinical decision-making based on measurable residual disease (MRD) in acute myeloid leukemia (AML) is steadily increasing; therefore, accurate MRD results are imperative. Multi-parameter flow cytometry (MFC) is often used due to wide availability, particularly when no suitable mutation is present. Two approaches are most common in MFC-MRD: 1. Leukemia-associated immunophenotype (LAIP) approach, in which the aberrant phenotype is identified at diagnosis and tracked during follow-up. 2. Different-from-Normal (DfN) approach, in which aberrant surface antigen expression patterns not present in normal bone marrow (empty spaces in 2D dot plots) are identified at follow-up. The LAIP approach is sensitive, but does not consider immunophenotypic shifts and cannot be used when no LAIP is identified at diagnosis or when a diagnosis sample is unavailable. The DfN approach can detect these shifts, but ample expertise is needed to determine a malignant population. In the ELN-DAVID AML-MRD guidelines 2022, it is recommended to use the combination of the LAIP approach and the DfN approach to use the advantages of both approaches. In the HOVON-SAKK132 (HO132), the MRD was based on the combination of MFC-MRD and NPM1 RQ-PCR. Initially, the MFC-MRD in this study was based on the LAIP approach. However, the complete MFC panel (four 8-color tubes) published by Cloos et al. (J.Vis.Exp, 2018) was used for measuring MRD during follow-up. Therefore in this study, we had the unique opportunity to retrospectively analyze the samples after induction cycle 2 (C2) of patients in complete remission (CR/CRi) and determine additional MRD results with the LAIP-based-DfN approach.

The original LAIP approach encompasses determining the percentages of the LAIPs in the follow-up sample, which were also identified at diagnosis. For the LAIP-based-DfN approach, we followed the same approach, but determined the percentages of all possible LAIPs in the MFC-MRD panel at diagnosis and after C2. An emerging LAIP is defined as a LAIP that is present in the follow-up but was not found at diagnosis. MRD positivity was defined as ≥ 0.1% LAIP cells of white blood cells. Patients with an emerging LAIP < 0.1% or no LAIPs in follow-up would result in MRD negativity for LAIP-based-DfN, but was only possible when at least 500,000 cells were acquired in all four tubes.

At diagnosis, we received 884 samples of 905 patients (98%), in which we found a LAIP ≥ 10% of CD45dim blasts in 83% of the samples (Figure 1A). After C2, we received 615 samples, for which we had MFC-MRD results for 79% (487/615) by the LAIP approach. When we applied the LAIP-based-DfN approach, we scored 69 of 615 extra MRD results: an emerging LAIP ≥ 0.1% (MRDpos, N = 14), no LAIPs (MRDneg, N = 25) or an emerging LAIP < 0.1% (MRDneg, N = 30). When including these results in the data of the LAIP approach, the percentage of patients with an MRD result increased from 79% (487/615) to 90% (556/615), and the MRD positive percentage increased from 13% (81/615) to 19% (118/615)(Figure 1B). The LAIPs of emerging clones most often included CD56, CD15, and CD22. To assess the prognostic value in the HO132, we combined the reanalyzed LAIP-based-DfN MFC-MRD with the NPM1-MRD (N = 567). In these analyses, overall survival (OS) hazard ratio (HR) 95% confidence interval (CI) 1.92 (1.44-2.56) and relapse-free survival (RFS) HR 95% CI 1.7 (1.31-2.19) were similar to the original LAIP based/ NPM1 approach: OS HR: 95%CI 2.02 (1.48-2.76) and RFS HR: 95%CI 1.78 (1.34-2.35), respectively.

In conclusion, the LAIP-based-DfN approach adds additional MRD results while retaining the prognostic value. Since HOVON-SAKK centers are highly compliant with the protocol and 98% of diagnosis samples were available, the LAIP-based-DfN approach can offer additional value for studies with limited access to diagnosis samples.