Impact of KMT2A-PTD Mutational Subgroups on Outcome of AML Patients after Induction Therapy and Allogeneic Hematopoietic Cell Transplantation

Introduction

Genetic profiling of acute myeloid leukemia (AML) has increasingly revealed the possibility of adjusted therapeutic options. Recently, updates in genetic risk classifications introduced novel molecular features, but also excluded aberrations like partial tandem duplications in the lysine (K)-specific methyltransferase 2A (KMT2A-PTD), formerly associated with adverse prognosis. Furthermore, there is growing evidence that molecular subtypes confer differential prognosis. Therefore, we analyzed a large cohort of AML pts to decipher the prognostic impact of KMT2A-PTD mutations (KMT2A-PTD^mut) including variants e9e3, e10e3 and e11e3 and their role in allogeneic hematopoietic cell transplantation (alloHCT).

Methods

Genomic DNA and total RNA of pts enrolled within the SAL registry (NCT03188874) was extracted at diagnosis. Samples were screened for KMT2A-PTD^mut using the Mentype AMLplexQS Kit (Biotype, Dresden, Germany). Statistical as-treated analyses included standard statistical methods for categorial and continuous variables. Effects of alloHCT modeled as a time dependent covariate were estimated using Cox regression. Complete remission (CR) and survival rates were defined according to the current European Leukemia Net criteria. Logistic regression was used to evaluate CR rates and Cox regression to estimate hazard ratios (HR). The significance level was set at 0.05. All patients gave written informed consent. The study was conducted in accordance to the Declaration of Helsinki and was approved by the responsible ethics committees.

Results

Among 1409 pts, 206 pts harbored KMT2A-PTD^mut, while 1203 pts showed wildtype (WT) alleles. KMT2A-PTD^mut status was more often associated with secondary AML (p = .001) and less likely with concomitant NPM1 mutations (p < .001). Variants e9e3 (40.8%) and e10e3 (36.6%) were more frequently detected than e11e3 (22.6%).

A total of 136 KMT2A-PTD^mut and 859 WT pts received intensive induction therapy (IT). Rates of first complete remission (CR1) did not differ (p = .14). However, 5-year RFS was significantly lower in KMT2A-PTD^mut pts compared to WT pts (15% vs. 25%), with median RFS of 23 vs. 29 months (HR = 1.4, p = .008). Accordingly, median OS was shorter (29 vs. 61 months; HR = 1.6, p = .002) and 5-year OS was lower for KMT2A-PTD^mut pts compared to the WT counterparts (28% vs. 51%).

In CR1, 81.6% of KMT2A-PTD^mut pts proceeded to alloHCT and 18.4% of KMT2A-PTD^mut pts received chemotherapy-based consolidation only. Median RFS improved from 15 to 37 months (HR = 0.5, p = .13) and chance of 5-year RFS enhanced from 17% to 21% when KMT2A-PTD^mut pts underwent alloHCT in CR1 compared to KMT2A-PTD^mut pts receiving chemo-consolidation. Median OS increased from 30 to 43 months when being allografted (HR = 0.5, p = .2), paralleled by increased rates of 5-year OS (33% chemo-consolidation vs. 47% alloHCT).

Concerning KMT2A-PTD variants, there was no difference in likelihood to achieve CR1 between the three variants (p = .4). However, a trend towards shorter median OS (e9e3: 41 months, e10e3: 27 months, e11e3: 27 months) and worse median RFS (e9e3: 24 months, e10e3: 23 months, e11e3: 23 months) compared to WT pts (median OS: 61 months, median RFS: 29 months) could be revealed. Variant e10e3 was associated with worst prognosis (OS: HR = 2.1, p = .006; RFS: HR = 1.6, p = .06) and e9e3 with best prognosis (OS: HR = 1, p = .9; RFS: HR = 1.1, p = .7).

e9e3 demonstrated best prognosis for both consolidation strategies. Median OS was not reached until last follow-up. Median RFS was higher in alloHCT pts (15 vs. 38 months, HR = 0.5, p = .3). e10e3 and e11e3 also had a higher median OS (12 vs. 43 months, HR = 0.3, p = .3 and 12 months vs. median survival not reached, HR = 0.2, p = .3, respectively) and RFS (11 vs. 41 months, HR = 0.3, p = .3 and 11 months vs. median survival not reached, HR = 0.2, p = .3, respectively) when allografted.

Conclusions

Based on our retrospective analysis, KMT2A-PTD^mut are associated with reduced prognosis and survival in AML patients compared to WT pts. However, poor prognosis was effectively mitigated by alloHCT. Concerning KMT2A-PTD variants, we hypothesize alloHCT to be beneficial for improved survival. This effect seemed to be less pronounced for variant e9e3and more distinct for variant e11e3. Therefore, some KMT2A-PTD variants could be potentially implemented in AML risk stratification. However, prospective studies and larger patient numbers are needed.