Clinical relevance of an objective – limit of detection – limit of quantification - based flow cytometry approach for measurable residual disease assessment in acute myeloid leukemia. A post-hoc analysis of the GIMEMA AML1310 trial.

Abstract

Using a multiparametric flow cytometry (MFC) assay, we assessed the predictive power of a threshold calculated applying the criteria of limit of detection (LOD) and limit of quantitation (LOQ) in adult patients affected with Acute Myeloid Leukemia (AML). This was a post-hoc analysis of 261 patients enrolled in the GIMEMA AML1310 prospective trial. According to the protocol design, using the predefined MRD threshold of 0.035% bone marrow residual leukemic cell (RLC) calculated on mononuclear cells, 154 (59%) were negative (MRD<0.035%) and 107 (41%) were positive (MRD≥0.035%). Using LOD and LOQ, we selected the following categories of patients: 1) LODneg if RLC were below LOD (74; 28.4%); 2) LODpos-LOQneg if RLC were between LOD and LOQ (43; 16.5%); and 3) LOQpos if RLC were above LOQ (144; 54.4%). Two-year overall survival (OS) of these 3 categories was 75.4% vs. 79.8% vs. 66.4%, respectively (p=0.1197). Due to superimposable outcome, LODneg and LODpos-LOQneg categories were combined. Two-year OS of LODneg/LODpos- LOQneg patients was 77.0% versus 66.4% of LOQpos individuals (P=0.043). Such a figure was challenged in multivariate analysis (p=0.048, HR 0.628, 95% CI 0.396-0.997) that confirmed the independent role of LOD-LOQ approach in influencing OS. In the AML1310 protocol, using the threshold of 0.035%, 2-year OS of MRD<0.035% and MRD≥0.035% patients was 74.5% vs. 66.4%, respectively (p=0.3521). In conclusion, the use of LOD-LOQ method results in a more sensitive detection of MRD that, in turn, translates in a more accurate recognition of patients with different outcome.