Paper Details

ABSTRACT

Background
The optimal treatment for the large group of patients with AML without a class-defining molecular lesion or adverse karyotype remains uncertain. These patients, here called cytogenetically standard risk are currently classified as AML not otherwise specified (AML-NOS) or AML with myelodysplasia related gene mutations (AML-MRG) and stratified in the intermediate or adverse risk groups respectively. How induction therapy – including gemtuzumab-ozogamicin (GO) – and allogeneic stem-cell transplantation (allo-SCT) modify relapse and survival in this population remains unclear.

Methods
NCRI AML17 (2009-2014) and AML19 (2015-2020) were large, randomised trials for younger adults with newly diagnosed AML or high-risk MDS; induction included DA +/- etoposide or FLAG-Ida, with/without GO. We defined the study population by excluding patients with AML defining genetic lesions, TP53 mutation and adverse cytogenetics according to ELN 2022.

Both trials used a risk-directed transplant policy for these patients: allog-SCT in first complete response (CR1) was recommended for patients defined as high-risk using a validated risk score based on clinical parameters after course 1 – age, sex, secondary disease, cytogenetics, response and white blood cell (WBC) count – or because of primary refractory disease after course 1. Additionally, in AML19, patients with FLT3 ITD (without NPM1 co-mutation) were recommended for CR1 allo-SCT. Primary endpoint was OS, and secondary CIR and NRM. Treatment and biological abnormalities were studied in univariate log-rank analysis (UVA) and multivariable Cox proportion hazards regression (MVA); allo-SCT was studied as a time-dependent variable.

Results
868 adults were included (AML17 49%, AML19 51%); median age was 54 (IQR 44–60). 477 (55%) had AML-MRG and 391 (45%) AML-NOS. Patients with AML-MRG were older (median 56 vs 51 years) and more often had a prior haematological disorder (15% vs 6%). Median follow-up was 50 months. Three-year OS and CIR were 49% and 47%. MVA showed improved OS with GO (3y OS 55% vs 44%; HR 0.77, 95% CI 0.62–0.94, p=0.01) but the effect was confined to AML19 (HR 0.69, 0.51-0.96, p=0.02). 

This benefit was caused by an increase in the CR rate after induction (90.8% vs 81.4%) with no effect on relapse. FLAG-Ida reduced relapse (3y CIR 35% vs 51%; CIR HR 0.57, 0.43–0.78, p<0.001) without an OS effect. Adverse factors for OS in MVA were older age (HR 1.19 per decade, p<0.001), higher WBC (HR 1.20, p<0.001), KRAS (HR 1.93, p<0.001) and WT1 mutations (HR 1.60, p=0.002); Patients with mutated IDH2 had reduced CIR (CIR HR 0.68, p=0.01).

OS did not differ between patients with AML-MRG (ELN 2022 Adverse Risk) and AML-NOS (ELN 2022 Intermediate Risk) (3y OS 47% vs 51%, p=0.144). Patients with one or two MRG mutations had similar OS to ELN-2022 Intermediate (OS HR 1.01 and 1.15, both non-significant), while those with ≥3 mutations had worse outcomes (OS HR 1.53, p=0.003). 499 patients (57%) underwent allo-SCT. Allo-SCT improved OS overall (HR 0.74, p=0.005), and patients who underwent it in CR1 had better OS than those transplanted beyond CR1 (HR 0.58, p<0.001).

Importantly, the Allo-SCT benefit was confined to the AML-MRG subgroup (HR 0.67, p=0.005). We could not observe OS benefit for patients allocated in the AML-NOS / ELN 2022 Intermediate risk, although allo-SCT reduced relapse in both groups (6-month landmark CIR p<0.001), indicating that NRM may offset relapse for this subgroup. The strongest per-gene transplant effects in AML-MRG were observed for SF3B1 (HR 0.17, p = 0.001) and U2AF1 (HR 0.24, p < 0.001). Most individual MRG mutations had similar OS than AML-NOS; exceptions were SRSF2 (HR 1.39, p=0.007) and ASXL1 (HR 1.32, p=0.04)

Conclusions
For cytogenetically standard-risk AML, we observed improved outcomes with GO (for response and survival) and FLAG-Ida (for relapse), with these effects seen mainly in AML19. In contrast, the survival benefit of allo-SCT was consistent across both trials and restricted to patients with AML-MRG.

These findings support directing patients with AML-MRG towards allo-SCT in 1st complete remission, where their outcomes are comparable to other intermediate-risk patients. Conversely, the benefit of CR1 allo-SCT for patients with AML-NOS is less clear, and other factors including MRD may better inform this decision. Therefore, the results of this study provide further support for the current ELN2022 risk stratification scheme.