Role of Measurable Residual Disease (MRD) in Redefining Complete Response (CR) in Elderly Patients with Acute Myeloid Leukemia (AML): Results from the Pethema-Flugaza Phase III Clinical Trial

Background: Incidence of AML is highest among the elderly and the general outcome is poor as compared to young patients, even those who tolerate intensive induction chemotherapy and achieve morphological CR. However, the role of MRD in redefining CR in elderly AML remains poorly investigated due to the reluctance to treat them with intensive chemotherapy, the renewed interest in low-intensity therapy such as hypomethylating agents (HMA), and the lack of molecular MRD markers in most patients.

Aim: To help defining the role of MRD assessment by multidimensional flow cytometry (MFC) and therapeutic decision making in older AML patients treated with semi-intensive chemotherapy vs HMA.

Methods: A total of 285 AML patients (excluding APL) with a median age of 75 were included in the phase III PETHEMA-FLUGAZA clinical trial and were randomized to receive 3 induction cycles with fludarabine and cytarabine (FLUGA) followed by 6 consolidation cycles with reduced intensity FLUGA, vs 3 induction cycles with 5-azacitidine (AZA) followed by 6 consolidation cycles with AZA. After consolidation, patients continued with the same treatment if MRD ≥0.01% or stopped if MRD <0.01%. MRD was prospectively assessed after induction and consolidation among patients in CR with or without incomplete blood count recovery, in a central laboratory blinded for clinical outcomes. At diagnosis, the EuroFlow panel for MDS/AML (first-five 8-color combinations) was used to identify leukemia-associated immunophenotypes (LAIP). Patients without a LAIP (6%) were excluded from this analysis. At CR, bone marrow samples were immunophenotyped with ≥2 8-color combinations based on previously identified maturation arrest, lineage commitment and LAIPs, maintaining markers' position from diagnosis to MRD to provide a digital fingerprint of leukemic blasts at diagnosis during MRD assessment. Over 1 million events per tube were measured for assessing MRD with an estimated sensitivity of 0.01%. The cumulative incidence of relapse (CIR) was calculated from the date of CR to the date of relapse, considering death without relapse as a competing event.

Results: On intention-to-treat, 38/141 (27%) patients achieved CR after 3 cycles of FLUGA, and 31/144 (21.5%) after 3 cycles of AZA (P =.33). Among patients in CR with previously identified LAIP, 14/69 (20%) achieved a negative MRD status whereas the remaining 55 (80%) had persisting MRD: 56% with ≥0.1% MRD and 24% with 0.01%-0.09% MRD. Of note, negative MRD rates were particularly lower among AML patients with myelodysplasia-related changes as compared to other subtypes (11% vs 27%, respectively; P=.09). Regarding the effect of semi-intensive chemotherapy vs HMA on depth of response, we observed a non-significant trend toward higher MRD-negative rates among patients in CR after FLUGA vs those treated with AZA (26% vs 15%, respectively; P=.28).

The 2-year CIR rates for MRD-positive and MRD-negative patients were 88% and 47%, respectively (HR, 3.3; P =.001). Of note, the CIR of patients in CR but with persistent MRD were similarly poor as compared to those in partial remission (HR, 0.82; P =.48). Furthermore, MRD-positive patients with adverse cytogenetics displayed the poorest outcome with significantly higher 2-year CIR rates than MRD-positive cases with intermediate/favorable cytogenetics (HR, 2.1; P =.008). Interestingly, among patients in CR and persistent MRD, 2-year CIR rates showed a non-significant trend towards slightly more frequent relapses in those treated with FLUGA vs AZA (91% vs 77%, respectively; HR, 1.7; P =.09). On multivariable analysis for CIR including MFC-MRD and cytogenetics, MFC-MRD (HR, 3.6; P =.001) and cytogenetics (HR, 2.0; P =.007) retained significant prognostic value. The median overall survival (OS) for MRD-positive and MRD-negative patients was 17 and 29 months, respectively (P =.04). On multivariable analyses for OS including age, WBC count, cytogenetic grouping and secondary disease, persistent MRD showed a trend for independent prognostic value (HR, 2.4; P =.07).

Conclusions: This study reveals that sensitive MFC-MRD assessment supersedes CR and is an independent prognostic factor in older patients with AML, treated with semi-intensive chemotherapy or HMA. Nevertheless, the risk of relapse among the few patients with no MRD (5%) remains high after stopping treatment, and warrants innovative approaches aimed at maintaining an MRD-negative CR status.