Minimal Residual Disease Monitoring in Acute Myeloid Leukemia (AML) with Translocation t(8;21)(q22;q22): Results of the AML Study Group (AMLSG)

Background: Acute myeloid leukemia (AML) with t(8;21)(q22;q22) results in the formation of the RUNX1-RUNX1T1 fusion transcript which can be used to monitor minimal residual disease (MRD) by quantitative reverse transcriptase polymerase chain reaction (qRT-PCR). Early identification of patients (pts) with a high risk of relapse will allow pre-emptive therapy including allogeneic hematopoietic cell transplantation (alloHCT). Recent studies in AML with NPM1 mutation or the CBFB-MYH11 gene fusion revealed that MRD persistence is significantly associated with a high risk of relapse. However, the prognostic impact of MRD assessment in RUNX1-RUNX1T1-positive AML is not well established.

Aims: To assess the prognostic impact of qRT-PCR-based MRD monitoring in bone marrow (BM) of pts with t(8;21)/RUNX1-RUNX1T1-positive AML obtained at defined time-points (diagnosis, first and second cycle of chemotherapy, end of treatment).

Methods: In total, 120 pts were included based on the availability of a diagnostic BM sample and at least two subsequent BM samples obtained during therapy and at the end of treatment; 106 pts were enrolled in one of six AMLSG treatment trials: AML HD93 (n=1), AML HD98A (NCT00146120; n=13), AMLSG 06-04 (NCT00151255; n=4), AMLSG 07-04 (NCT00151242; n=43), AMLSG 11-08 (NCT00850382; n=31), AMLSG 21-13 (NCT02013648; n=14); 14 pts were treated outside clinical trials. All pts received anthracycline- and cytarabine-based intensive induction followed by subsequent high-dose cytarabine consolidation cycles. For MRD assessment, qRT-PCR from BM specimens was performed using TaqMan technology; RUNX1-RUNX1T1 transcript levels (TL) were reported as the normalized value of RUNX1-RUNX1T1 per 10⁶ transcripts of the housekeeping gene beta2-microglobulin. The maximum sensitivity of the assay was 10^-6.

Results: The median age of the pts was 47 years (yrs; range, 18-73 yrs); at the time of diagnosis there was a broad range of RUNX1-RUNX1T1 TL (18490 to 14440000) with a median of 227800. RUNX1-RUNX1T1 TL did not correlate with clinical features (age, WBC, platelets, LDH, BM blasts) or associated gene mutations such as KIT, FLT3-ITD/TKD, NRAS or ASXL2. However, pts with additional FLT3 mutation showed higher TL compared to wild-type pts (median, 412955 vs 219052). Cox regression analysis using RUNX1-RUNX1T1 TL as a log10 transformed continuous variable showed that higher RUNX1-RUNX1T1 TL were significantly associated with a higher cumulative incidence of relapse (CIR), inferior event-free survival (EFS) and shorter overall survival (OS) for the two time points "after first treatment cycle" and "at end of treatment" (CIR: HR, 1.84, p=0.001; HR, 1.60, p=0.03; EFS: HR, 1.59, p=0.01, HR, 1.74, p=0.01; OS: HR, 1.63, p=0.02, HR 2.13, p=0.009, respectively). In univariate analyses achievement of MRD negativity (n=35) at the end of treatment was significantly associated with a superior 4-yr OS (93% vs 67%; p=0.007) and 4-yr EFS (81% vs 61%; p=0.04) whereas achievement of MRD negativity after the first (1/85) and second (20/89) treatment cycle was low not reaching significance for any of the clinical endpoints. Separation of the RUNX1-RUNX1T1 TL according to quartiles of distribution showed significant differences in OS (p=0.04), and remission duration (p=0.006) "after first cycle" whereas "at end of treatment" significant differences were only found for OS (p=0.009). Finally, we evaluated the impact of concurrent KIT mutations on the kinetics of RUNX1-RUNX1T1 TL. Following the first treatment cycle, the median RUNX1-RUNX1T1 TL were significantly lower in the KIT wildtype group compared with the KIT mutated group (p=0.02); the same was true "at the end of treatment" (p=0.02).

Conclusions: In our study, achievement of MRD negativity at the end of treatment was significantly associated with a better outcome in t(8;21)-positive AML. The fact that earlier time points did not allow the identification of pts with a high relapse risk is probably due to the high sensitivity of the qRT-PCR assay which is also reflected by the low number of pts achieving qRT-PCR negativity after first and second treatment cycle, respectively. Further analyses are ongoing including multivariable as well as molecular subgroup analyses.

*These authors contributed equally to the work: MA, AC

MA was supported by the Else-Kröner-Fresenius-Stiftung (EKFS).