FLAG-Ida Combined with Gemtuzumab Ozogamicin (GO) Improves Event Free Survival in Younger Patients with Newly Diagnosed Acute Myeloid Leukaemia (AML) and Shows an Overall Survival Benefit in NPM1 and FLT3 mutated Subgroups. Results from the UK NCRI AML19 Trial

Background

The optimal induction regimen for younger patients with newly diagnosed AML is uncertain. The MRC AML15 trial suggested a higher response rate and reduced relapse risk with FLAG-Ida compared to a Daunorubicin-araC (DA) +etoposide but did not show an overall survival (OS) benefit (Burnett, JCO,2013,31,3360). GO has been shown to improve survival in patients with favourable and intermediate risk cytogenetics, although it is unclear whether a single or fractionated dose is optimal. (Hills, Lancet Oncology,2014, 15,986)

Methods

The NCRI AML19 trial (ISRCTN78449203) randomised 1475 patients with newly diagnosed AML or MDS-EB2, not known to have adverse cytogenetics, between FLAG-Ida ( n=738) and DA (Daunorubicin, AraC, n=737). Of these, 1031 were also randomised to receive a single dose of GO (GO1, 3mg/m² on D1, n=513) or a fractionated schedule (GO2, max 5mg on D1+4 , n=518). Post-remission therapy consisted a second course of DA or FLAG-Ida without GO followed by up to 2 courses of HDAC . No FLT3 inhibition was used.

Patients were designated as high risk after course 1 based on a validated risk score, or if they had the genotype FLT3-ITD⁺, NPM1^-. After course 2, patients could be designated high risk based on measurable residual disease (MRD) positivity (i.e. detectable NPM1 mutated transcripts in the peripheral blood (PB) or following protocol amendment, flow cytometric MRD >0.1% in the bone marrow). High-risk patients were recommended for transplant and could enter a separate randomisation, but are included in this analysis.

The primary endpoint was OS. Secondary endpoints included event free survival (EFS), relapse free survival (RFS), response defined according to ELN criteria and MRD. Patients aged 18-60 years were eligible but older patients could be enrolled at the investigators discretion. Here we report the outcomes of the patients randomised between FLAG-Ida-GO and DA-GO.

Results

The median age was 51.5 yrs. 88% had de novo AML, 7% secondary AML and 5% MDS-EB2. MRC defined cytogenetic risk was favourable in 12%, intermediate in 75% and adverse in 9%, the remaining had missing data. Of major molecular subgroups, 30% had NPM1 and 26% FLT3 ITD or TKD mutations. Median follow-up was 32 months. Complete remission (CR/CRi) was 91% for DA-GO and 93% for FLAG-Ida-GO and did not differ by GO dose (92.3% for GO1 and 91.2% for GO2). Day 30 and 60 mortality were not different between DA and FLAG-Ida (D30, 2.9% vs 3.1%, p=0.37; D60 4.6% vs 4.3%, p=0.27) or between GO1 and GO2 (D30, 2.15% vs 3.3%, p=0.1; D60 3.3% vs 5.6%, p=0.07).

There was no difference in OS or EFS by GO dose so further outcome analyses were undertaken combined. EFS was significantly better with FLAG-Ida-GO compared to DA-GO (HR 0.73, CI 0.61-0.87, p<0.001) but there was no difference in OS (HR 0.92, CI 0.75-1.13, p=0.41) (Figure 1a and 1b). RFS also favoured FLAG-Ida-GO.

In subgroup analysis OS benefit was observed for FLAG-Ida-GO in patients with NPM1 mutation (OS 82% vs 64% at 3 yrs, HR 0.5, CI 0.31-0.81, p=0.005). Furthermore NPM1 MRD positivity by RT-qPCR in the PB after 2 courses of induction was 23% (30/130) with DA-GO compared to 12% (16/130) (p=0.03) with FLAG-Ida-GO. The survival benefit for FLAG-Ida-GO was also seen in FLT3+ve AML (OS of 64% vs 54% at 3 yrs, HR 0.67, CI 0.45 -0.99, p=0.047) with the greatest benefit seen in NPM1 co-mutated patients (HR 0.32 CI 0.16 –0.61). In patients with favourable risk cytogenetics, there was no EFS or OS benefit for FLAG-Ida-GO over DA-GO.

Fewer transplants were performed in the FLAG-Ida-GO arm overall, 237 (46%) compared to 278 (54%) with DA-GO including fewer in CR1 (177 compared to 194) reflecting the application of MRD negativity to guide transplant decisions and the reduced relapse risk with FLAG-Ida.

Conclusion

In this large, randomized comparison we observed no benefit for a fractionated GO schedule but FLAG-Ida-GO significantly improved EFS compared to DA-GO. Although there was no overall survival benefit for FLAG-Ida-GO, there was evidence of an OS benefit in major sub-groups, including patients with NPM1 and FLT3 mutations. This finding was supported by MRD analysis that showed a reduction in NPM1 MRD positivity after two courses of FLAG-Ida-GO. Furthermore this survival benefit was associated with a reduction in the requirements for transplant in CR1 and overall. Given the benefit observed in FLT3 mutated AML in the absence of a FLT3 inhibitor, studies combining FLAG-Ida-GO with Midostaurin are warranted.