S134: flag-ida combined with gemtuzumab ozogamicin (go) reduced mrd levels and improved overall survival in npm1mut aml independent of flt3 and mrd status, results from the aml19 trial

Background: The NCRI AML17 trial established the prognostic impact of measurable residual disease (MRD) by RT-qPCR in patients with NPM1mut AML. Patients testing MRD+ve in the peripheral blood (PB) following the second course of chemotherapy (PC2) had a high risk of relapse (77% at 3 years) and poor overall survival (OS, 25% at 3 years). The subsequent NCRI AML19 trial randomised 1033 patients with newly diagnosed AML or MDS-EB2 between FLAG-Ida and DA3 + 10 and either a single dose of GO (GO1) or a fractionated schedule (GO2). No FLT3 inhibition was used. The outcomes for this trial have been reported with no OS benefit (Russell et al, ASH, 2022), however in NPM1mut patients (n=307) there was an OS benefit with FLAG-Ida-GO (5y OS 82% vs 64%, HR 0.50, CI O31-0.81, p <0.005). Aims: To analyse the impact of FLAG-Ida-GO on MRD and outcome of post-induction therapy in NPM1mut AML with and without a FLT3 mutation Methods:NPM1mut MRD was measured following each chemotherapy course. Patients who were PC2 PB MRD+ve were recommended for transplant in CR1 Post-remission therapy for MRD-ve patients consisted of up to 2 courses of HDAC. Results: The survival benefit for FLAG-Ida-GO was seen in both FLT3mut (HR 0.32, 95CI 0.17-0.62) and FLT3wt patients (HR 0.75, 95CI 0.37-1.51) with no evidence of differential benefit on tests for heterogeneity (Figure 1). FLAG-Ida-GO increased the number of patients who were PC2 PB MRD-ve (88% vs 77% with DA-GO, p=0.02), as well as the number with an MRD-ve bone marrow (BM) both PC2 (56% vs 37%, p=0.004) and at end of treatment (70% vs 58%, p=0.32). In those with undetectable PB MRD, the BM response was deeper in the FLAG-Ida-GO arm, with 60% also BM MRD-ve compared to 47% with DA-GO (p=0.069). The same trend for a deeper BM response was seen in those who were PB PC2 MRD+ve. The improved FLAG-Ida-GO MRD response was seen in both FLT3mut (PC2 PB-ve 83% vs 68%) and FLT3wt (PC2 PB-ve 92% vs 82%). Within the DAGO arm, GO2 resulted in an increase in the number testing PB PC2 MRD-ve (84% vs 69% for GO1, p=0.04) but did not improve survival. There was no effect of GO dose on MRD in the FLAG-Ida-GO arm. For PB PC2 MRD+ve patients, 61% proceeded to transplant in CR and this did not differ by randomisation. For these patients 3y OS was 59%, those randomised to FLAG-Ida-GO had a trend to better survival than those allocated DA-GO (74% vs 51% at 3 years, HR 0.52, 95CI 0.17-1.57). For PC2 PB MRD-ve patients, outcomes were excellent with both therapies, but here again survival was superior in patients treated with FLAG-Ida-GO (3y OS 90% vs 78%, HR 0.43, 95CI 0.22-0.87). There was no heterogeneity in the FLAG-Ida-GO benefit based on MRD response (Figure 1). Patients could receive up to 2 cycles of consolidation with HDAC, for patients who were PB MRD-ve after FLAG-Ida-GO (n=115) there was no evidence that overall or relapse-free survival (RFS) was improved by consolidation (for 0, 1 and 2 cycles of consolidation respectively, 3 year OS 90% vs 83% vs 93%, p=0.53; 3 year RFS 75% vs 65% vs 81%, p=0.14) Summary/Conclusion: FLAG-Ida-GO improved the OS of patients with NPM1mut AML and reduced the number who were PB PC2 MRD+ve compared to DA-GO. This survival benefit was independent of FLT3 mutation status and was seen in both PB PC2 MRD+ve and MRD-ve patients and was supported by the finding that BM MRD levels in both groups were lower with FLAG-Ida-GO including those testing PB post C2 negative. For those patients who were PB MRD-ve after 2 cycles of FLAG-Ida-GO, this treatment appears sufficient.Keywords: AML