Prognostic Implications of Minimal Residual Disease Detection in Acute Myeloid Leukemia (AML) with Cytarabine Based Therapy of Differing Dose Intensities

Background:

Persistence of minimal residual disease (MRD) by multi-parameter flow cytometry (MFC), is an important predictor of outcome in AML treated with cytarabine-based regimens. However, differences in intensity of chemotherapeutic regimens may also influence the rate of achieving MRD negativity and clinical outcome. We sought to evaluate the comparative value of MRD measurement at different time points in two populations of patients (pts) treated, for AML, with induction regimens of widely differing cytarabine-dose intensities.

Methods:

Newly diagnosed pts were prospectively enrolled in two treatment protocols: cohort1 was treated with cladribine (5mg/m2 on D1-5), idarubicin (10 mg/m2 on D1-3), and cytarabine (1 g/m2 on D1-5) (CLIA) for induction and consolidation (Jain ASH 2016). Patients in cohort2 [CLAD/LDAC cohort] were treated with cladribine (5 mg/m2 D1-5) or clofarabine (20 mg/m2) combined with low dose cytarabine (LDAC) at 20 mg BID on days 1-10 (Kadia Cancer 2015, Kadia ASH 2014). This regimen also included up to 18 alternating cycles of cladribine/clofarabine plus LDAC and decitabine consolidation/maintenance. MRD assessment was performed by 8-color flow cytometry on the bone marrow at approximately days 30 and 60 from start of therapy, with an optimum sensitivity threshold of 0.01%. Relapse free (RFS) and overall (OS) survival was analyzed by MRD status at days 30 and 60, among complete responders with (CR) or without (CRp/CRi) count recovery. For time to event analyses (OS), survival curves were established by Kaplan-Meier approach. Comparison of survival curves was done using the log rank test. Variables with p ≤ 0.05 on univariate analysis were chosen for the multivariate cox regression model.

Results:

Study population included 51 pts treated with CLIA and 78 pts with CLAD/LDAC who achieved a CR/CRp/CRi. Baseline pt and disease characteristics by treatment cohort are outlined in table 1. The median age and WBC count for cohort1 was 54 [range, 22-65] years and 4.3 [0.8-58.2] K/µL, respectively. Forty six (90%) of 51 pts had achieved complete response by the day 30 time point. All 46 CR/CRi/CRp-at-day-30 pts were assessed for MRD elimination at 30 days, of whom 28 (61%) were MRD negative (MRD-NEG). Twenty four (47%) pts had MRD data available at 60 days; 19 (79%) were MRD-NEG.

The median age and WBC count for cohort2 was 69 [34-88] years and 2.2 [0.5-27.8] K/µL, respectively. 62 (79%) of the 78 pts had achieved complete response by the day 30 time point. Sixty one of the 62 (98%) CR/CRi/CRp-at-day-30 pts were assessed for MRD at 30 days, of whom, 40 (65%) were MRD-NEG [vs. cohort1; p = 0.34]. Twenty five (32%) of all pts had MRD data available at 60 days; 14 (56%) were MRD-NEG [vs. cohort1; p = 0.13]. Analyzing by day 30 MRD status, OS [Fig 1A] and RFS [Fig 1B] was superior among MRD-NEG pts in cohort1. Day 30 MRD status predicted for superior RFS and OS on multivariate analysis. Likewise, day 60 MRD negativity was also associated with a superior OS [Fig 1C], with RFS showing a trend toward statistical significance, probably due to small sample estimates [Fig 1D].

Unlike in the case with cohort 1, day 30 MRD negativity did not impact either OS [Fig 2A] or RFS [Fig2B] among pts in cohort2. However, day 60 MRD negativity was associated with a significantly superior OS [Fig 2C] and a trend toward improved RFS [Fig 2D] (likely due to low sample size estimates). While day 30 MRD status did not predict for either RFS or OS, a multivariate analysis analyzing the impact of karyotype, response type (CR vs CRi/p), MRD status, and transplant status on OS demonstrated only day 60 MRD negativity and transplant status to be predictive for OS.

Conclusion:

In contrast to CLIA, where post induction MRD negativity status at both day 30 and 60 time points impact RFS and OS, a delayed post-induction time point (day 60) for MRD determination is more informative with cladribine-LDAC low intensity induction therapy. Our study supports the general hypothesis that delayed time points are more informative than earlier ones, particularly with low intensity therapies.

Table 1

View largeDownload slide

Table 1

View largeDownload slide

Close modal