Molecular Predictors and Current Management of Minimal Residual Disease (MRD) Following Induction Chemotherapy for Acute Myeloid Leukemia (AML)

Background: MRD is a powerful prognostic factor in AML. Emerging data indicate that allogeneic stem cell transplant (alloSCT) with MRD results in outcomes equivalently poor to alloSCT with morphologic AML (Araki et al., JCO 2016). Genomic predictors of MRD are unclear, and relative efficacy of therapies for MRD remains elusive.

Objectives: Here we provide an integrated analysis of responses for 163 patients (pts) who underwent induction chemotherapy with baseline next-generation sequencing (NGS) followed by serial immunophenotypic monitoring for MRD.

Methods:163 patients starting in April 2014 who underwent induction chemotherapy at Memorial Sloan Kettering Cancer Center were retrospectively studied. All received anthracycline + cytarabine, with or without investigational agents. Immunophenotypic MRD was identified in bone marrow aspirates (BMA) by multiparameter flow cytometry. Any level of residual disease was considered MRD+. Molecular analysis was obtained from pre-induction BMA by NGS using 28 or 49 gene panels. Cytogenetics/FISH were performed using standard techniques.

Results: Patient characteristics are in Table 1. 7/163 (4.9%) died within 30 days of induction.153 pts had BM biopsy after induction prior to further therapy. 124/153 underwent flow after induction. 65/124 (52.4%) achieved CR/CRi after induction alone, 31/124 (25%) MRD+CR/CRi, and 34/124 (27.4%) MRD-CR/CRi. Pre-induction molecular analysis from 126 suggests that certain cytogenetic and molecular abnormalities correlate with achievement of MRD-CR. (Figure 1) Only 2/25 (8%) with RUNX1, 0/13 with SF3B1, and 0/11 with TP53 mutations achieved MRD-CR/CRi as best response after 1 cycle of induction. Only 3 additional RUNX1, 2 SF3B1, and 0/11 TP53 achieved MRD-CR/CRi as best response after a second cycle of therapy. In contrast, 7/8 with CBF AML (inv16 and no KIT mutation, n=4) or (t(8;21), n=3) achieved MRD-CR/CRi (n=5) or CR without flow (n=2) after 1 cycle of induction. 91/163 (55.8%) underwent alloSCT following induction or additional therapy. Post-alloSCT follow-up indicates potential value in converting MRD+ to MRD-. 84/91 were evaluable for MRD with flow cytometry prior to alloSCT. 41/84 (48.8%) were MRD-, 30/84 (35.7%) MRD+, and 13/84 (15.4%) persistent AML. 13/41 (31.7%) MRD-pre-alloSCT were MRD- post-induction. 28/41 (68.2%) MRD+ or persistent AML converted to MRD- prior to alloSCT following additional therapy. 23/29 MRD+CR/CRi pts after induction were intermediate/unfavorable and therefore transplant candidates. 19/23 MRD+CR/CRi intermediate/unfavorable underwent transplant (9 without post-induction therapy, 10 after consolidation), while 4 did not proceed to transplant due to relapse after induction (n=1), relapse after consolidation (n=2), and patient preference. There was no significant difference in post-transplant OS between early MRD-CR immediately following induction and later conversion to MRD-CR prior to alloSCT (Figure 1B). Post-transplant analysis reveals that most pts who enter transplant with persistent AML (n=13) or MRD+ (n=30) clear MRD (30/43, 69.7%) by the first post-transplant BM (median 32 days, Figure 1C). Despite initial post-transplant MRD clearance, pts who entered alloSCT with persistent AML or MRD+ had poorer post-transplant OS compared to pts who entered alloSCT with MRD- (p=0.02, Figure 1D).

Conclusion: Our data show that AML pts with specific molecular mutations (RUNX1, SF3B1, and TP53) are unlikely to achieve MRD-CR/CRi after induction chemotherapy. We further show that additional therapy such as consolidation may be advantageous for some MRD+ pts to achieve MRD-CR prior to alloSCT, although others remain resistant to MRD clearance. Post-transplant OS is improved in pts who are MRD- at time of transplant, regardless of whether they required additional therapy beyond induction to achieve this state. Our results suggest that development of MRD-eradicating therapies after AML induction has the potential to improve post-transplant outcomes.

View largeDownload slide

View largeDownload slide

Close modal