Impact of Cytogenetics, Induction Chemotherapy and MRD on Outcomes of Intensively Treated AML with KMT2A Rearrangements – Experience from UK NCRI AML17 and AML19

KMT2A rearrangements (KMT2Ar) occur in ~5% of adult AML and their outcomes are generally poor, with high rates of early relapse. Menin inhibitors have shown encouraging activity in relapsed/refractory disease, and trials combining these agents with intensive chemotherapy or venetoclax-azacitidine are ongoing. However, the optimal backbone chemotherapy regimen, and the role of MRD during induction, are not well established. We sought to address these issues by examining outcomes of KMT2Ar patients in two large prospective RCTs.

Methods

NCRI AML17 (2009-2014) and AML19 (2015-2020) enrolled younger adults with newly diagnosed AML. Results from karyotype, FISH, RT-qPCR and RNA sequencing were reviewed to identify all patients with KMT2Ar. MRD was generally assessed by flow cytometry in AML17, while RT-qPCR was used in AML19.

Results

218 patients (AML17 n=135, AML19 n=83) were identified with a median age of 41y (range 16 – 68). 8 (4%) had a prior myeloid disorder and 22 (10%) were therapy related. 75 (34%) patients had ELN 2022 intermediate risk disease with t(9;11)/MLLT3::KMT2A and the remaining 66% were adverse risk. 131 patients (60%) had no other cytogenetic abnormalities. Complex cytogenetics (defined here as 3 structural abnormalities in addition to KMT2Ar) were seen in 12 patients (6%) and 7 (3%) had other adverse abnormalities (del5q or –7). FLT3 ITD mutations were present in 9 (4%) and TKD in 15 (7%). Treatment was with DA in 127 (58%), FLAG-Ida in 48 (22%), ADE in 35 (16%) and CPX-351 in 8 (3.7%). 82 (38%) received GO.

196 patients (90%) achieved CR or CRi by the end of the second induction, with 81% in CR/CRi after course 1. Higher rates of refractory disease were seen in patients with a complex karyotype (17%) or other adverse abnormalities (29%).

Median follow-up was 4.8y. Day 30 and 60 mortality were 4% and 5% respectively. Median overall survival (OS) was 2.1y (95CI 1.7 – 3.1) and was not significantly different between ELN intermediate (2.7y) and adverse risk (1.9y) disease. On multivariable analyses, higher WCC (HR 1.23, 95CI 1.08-1.41) and complex karyotype (HR 3.46, 95CI 1.72-6.98) were associated with worse survival while FLAG-Ida produced better survival (HR 0.42, 95CI 0.24-0.75). 2-yr OS was 68% (median not reached) for FLAG-Ida compared to 48% (median 1.8y) for other regimens.

Of 196 patients achieving remission, 99 (51%) were transplanted in CR1 (60% of ELN adverse and 31% of ELN intermediate patients). CR1 transplant was associated with improved RFS (0.66, 95CI 0.45 – 0.97) and a trend to better OS (HR 0.75, 95CI 0.50 - 1.11), with no heterogeneity based on ELN risk group.

69 patients were included in a direct randomisation of FLAG-Ida (n=42) vs DA (n=27) in AML19. FLAG-Ida was associated with a numerically higher remission after C1 (93% vs 81%, p=0.2) but not after second induction (95% vs 96%). Despite a lower rate of CR1 transplant after FLAG-Ida (40% vs 59%), relapse was significantly reduced (2y CIR 23% vs 61%, p<0.001). Median OS was not reached with FLAG-Ida vs 2.2y for DA, with 2y OS 67% vs 57% (HR 0.54, 95CI 0.26-1.13).

No survival difference was seen in the AML17 randomisations of DA vs ADE or DA with 60mg/m² vs 90mg/m² of daunorubicin, nor in the AML19 randomisations of GO1 vs GO2 or CPX-351 vs FLAG-Ida.

MRD was performed by flow cytometry in 76 and RT-qPCR in 61 patients. 50 patients had a post C2 qPCR result in either PB or BM, with results concordant in all patients with both samples (n=38). 18 patients (36%) remained MRD positive (52% for adverse risk, 14% for intermediate risk). Detectable MRD was strongly associated with relapse (CIR 74% vs 23%; HR 5.58, 95CI 2.4-13.2, p<0.001) and poor OS (2y OS 31% vs 77%, HR 4.42, 95CI 1.9-10.2, p<0.001). Only 3 MRD positive patients remain alive and relapse-free at last follow-up.

Flow MRD results were analysed at both post C1 and C2, examining thresholds of either 0.1% or any positivity. The only threshold predictive of outcomes was detection of any positivity post C1, with 60% of patients MRD positive and OS HR 1.88 (95CI 1.00-3.52).

Conclusion

Here we demonstrate a survival advantage for intensified induction treatment with FLAG-Ida in younger adults with KMT2Ar AML. We also show that molecular MRD after course 2 is a strong predictor of long-term survival. These data are important to inform the design of future trials incorporating menin inhibitors in front-line therapy, and suggest that intensified induction and RT-qPCR MRD should be incorporated into these studies.