Drivers of Deep Molecular Response and Outcomes in Patients with Core Binding Factor Acute Myeloid Leukemia

Introduction: Core-binding factor acute myeloid leukemia (CBF-AML), including t (8;21) and Inv (16), constitutes 15% of all AML cases and generally carries a favorable prognosis. However, relapse rate is ≈ 30% and predictors are not well understood. Although CBF-AML responds to induction chemotherapy, post-remission treatment can be further optimized, particularly with current next generation sequencing (NGS) and measurable residual disease (MRD) detection techniques. Impact of baseline somatic mutations on outcomes remains unclear, as does role of allogeneic stem cell transplant (allo-SCT). In this study, we investigated the potential drivers of treatment response and overall outcomes in CBF-AML.

Methods: In this single center retrospective study, clinical, demographic, and molecular data were collected from adult pts with CBF-AML who were treated at Moffitt Cancer Center. We included pts with MRD by polymerase chain reaction (PCR) for CBF abnormality (CBF-MRD) results available from time of induction and consolidation. All statistical analyses were performed in SPSS.

Results: Our study included 68 patients (pts), including 26 (38%) t (8;21) AML. The median age at diagnosis was 48 years, and 54% pts were male. Pts with inv (16) AML had significantly more leukocytosis than t (8;21) (10x10⁹/L vs. 23x10⁹/L; p=0.001). Only 5% of pts (n=3) in our cohort had extramedullary disease on presentation. Common mutations identified on baseline NGS (n=60) were KIT (35%), NRAS (27%), KRAS (15%), FLT3-ITD and FLT3-TKD (12% each; 2 pts had both types). Mutation in KIT was more common in t (8;21) AML (50% vs. 25%; p=0.04), whereas inv (16) cases were enriched in NRAS mutations (36% vs. 13%; p=0.04). Overall, 80% of CBF-AML pts had at least one kinase mutation (KIT/RAS/FLT3).

Majority of pts were treated with “7+3” (44%) and “7+3+ gemtuzumab ozagamicin (GO)” (40%) as induction regimens (Table 1). Daunorubicin was the predominant anthracycline, used in 80% of pts treated with “7+3” and 48% received dose of 90 mg/m². Overall, 31 (46%) pts received GO as part of induction treatment, with equal distribution of “day 1 only” and “day 1,4,7” schedules at 3mg/m². Median number of consolidation cycles was 3, and GO was incorporated in consolidation (commonly with high dose cytarabine) for 29 (43%) pts. After induction, 16% pts experienced CBF-MRD negativity which increased to 73% following consolidation, with no significant difference between t (8;21) and inv (16) AML. Use of GO during induction or consolidation was associated with significantly higher CBF-MRD negativity rate at completion of induction and consolidation (83% vs. 50%; p= 0.021). However, we did not find any significant correlation between groups based on presence of baseline kinase mutation/s and CBF-MRD negativity at end of induction or consolidation. In FLT3-positive cases, 100% of pts (5/5 evaluated;4 ITD and 1 TKD) achieved FLT3-MRD negativity by PCR (sensitivity of 5X10^-5) following consolidation.

Overall, 35% of pts (n=24) underwent allo-SCT in our cohort, including 15 in first remission (CR1). Kinase mutations were present at baseline in majority of these pts (75%; 87% in CR1 group). Notably, 9 pts (7 in CR1 group) in the allo-SCT cohort had positive CBF-MRD at the end of consolidation. Seven pre-transplant CBF-MRD positive cases turned MRD negative following allo-SCT. Relapse rate was low (n=10; 15%) in our CBF-AML pt population. Among pts with relapse, 6 were CBF-MRD positive post-induction and 3 had MRD even after consolidation. Use of GO in induction was associated with significantly lower relapse rate (0% vs. 28%; p= 0.001). Presence of kinase (KIT/RAS/FLT3) mutation at diagnosis was associated with shorter relapse free survival (RFS) in these pts (median RFS- 7.7 vs. 12.9 months (mo); p= 0.014; Figure 1). After a median follow up of 33.4 mo, median OS was 94.1 mo in our pt cohort. There was no significant OS difference between subgroups based on CBF-AML subtype, baseline mutations, or receipt of GO during treatment.

Conclusions: CBF-AML continues to be a favorable risk disease, with a median OS >7 years in our study population. Post-treatment CBF-MRD negativity rate was 73% and was predicted by the incorporation of GO in induction or consolidation. Rate of relapse rate was low at 15%, and all occurred in pts who did not receive GO during induction or consolidation. Baseline kinase mutations (KIT/RAS/FLT3) on NGS was associated with a shorter RFS in our cohort.