The role of dynamic MRD in treatment decision making for intermediate risk AML

Most younger patients with acute myeloid leukemia (AML) achieve a complete remission (CR) after induction chemotherapy but require post-remission treatment (PRT) to prevent relapse. PRT decisions are mostly based on risk-stratification using cytogenetics and molecular markers but selecting the optimal PRT, particularly for patients with intermediate risk AML, remains a challenge due to contradictory conclusions from previous studies. As such, there is a need for additional predictive factors to optimize PRT selection. Earlier reports have indicated a possible role for measurable residual disease (MRD), although the optimal timing of MRD assessment to guide treatment selection is widely debated. Here we summarize a report by Yu and colleagues which aimed to explore PRT choices based on dynamic MRD for younger patients with intermediate risk AML.¹

Study design

A total of 549 younger patients (median age, 37 years; range, 14−60 years) with de novo intermediate risk AML in first CR (CR1) were included in the study. Exclusion criteria included:

• Acute promyelocytic leukemia
• NPM1 variant with FLT3-ITD
• Failure to achieve a CR after two cycles of induction therapy
• Fewer than three cycles of consolidation in the chemotherapy group
• Absence of MRD parameters

Patients were classified into three groups according to PRT:

• Cytarabine-based consolidation chemotherapy (CMT, n = 154)
• Chemotherapy followed by autologous stem cell transplant (auto-SCT, n = 116)
• Chemotherapy followed by allogeneic stem cell transplant (allo-SCT, n = 279)

MRD was assessed in the bone marrow after induction therapy, after each cycle of PRT, and then at 2, 3, 4, and 6-month intervals during the first, second, third, and fourth/fifth years after PRT, respectively. To explore the association between MRD dynamics and PRT selection, subgroup MRD analyses were performed after one, two and three courses of chemotherapy, and patients were classified into the following subgroups:

1. Persistently MRD negative after the first induction regimen for all three courses
2. Persistently MRD positive for all three courses
3. Recurrently MRD positive after MRD negative
4. MRD negative after two courses of chemotherapy
5. MRD negative after 3 courses of chemotherapy

Evaluation endpoints included relapse, survival, and transplant-related mortality.

Study results

Selected patient characteristics are shown in Table 1. In the allo-SCT group, half of the patients received matched sibling donors and one third of patients received haploidentical donors. Over half of patients received first induction therapy with idarubicin hydrochloride and cytarabine (IA), and most patients received intermediate- or high-dose cytarabine as second induction or first consolidation therapy.

Table 1. Selected patient characteristics according to PRT group*

Relapse and survival

A total of 146 patients had relapsed at last follow-up. The key relapse and survival outcomes according to PRT group is presented in Table 2.

• Median time from CR1 to relapse was longer in the allo-SCT group, compared with the CMT group (p = 0.01).
• Incidence of relapse was lower for the allo-SCT group than either the auto-SCT group (hazard ratio [HR] 0.44; p = 0.001) or CMT group (HR 0.21; p <0.001), and lower for the auto-SCT group than the CMT group (HR 0.47; p < 0.001).
• Incidence of transplant-related mortality was higher for the allo-SCT group than for the CMT group (HR 9.36; p = 0.004) or auto-SCT group (HR 3.10; p = 0.04).
• Leukemia-free survival (LFS) and overall survival (OS) rates were comparable between the allo-SCT and auto-SCT groups but improved for the allo-SCT group (LFS: HR 0.38; p < 0.001; OS: HR 0.47; p < 0.001) and auto-SCT group (LFS: HR 0.53; p = 0.002; OS: HR 0.54; p = 0.006) compared with the CMT group.
• Graft-versus-host-disease-free, relapse-free survival (GRFS) was better for the auto-SCT group than for the CMT group (HR 0.54; p = 0.002) or the allo-SCT group (HR 0.65; p = 0.02).

Table 2. Relapse and survival outcomes by PRT group*

According to multivariate analysis, requiring two cycles of chemotherapy to achieve CR (versus one cycle) and MRD after three cycles of chemotherapy were risk factors for relapse, LFS, and OS in the entire population. Allo-SCT had a favorable association with relapse, LFS, and OS, compared with auto-SCT or CMT.

Association between MRD and PRT selection

• For patients who were persistently MRD negative after the first induction regimen, better GRFS was seen for the CMT group (HR 0.35; p = 0.03) and the auto-SCT group (HR 0.07; p = 0.01) compared with the allo-SCT group.
• Patients who were persistently MRD positive and received allo-SCT had
  • lower cumulative incidence of relapse, compared with CMT and auto-SCT (HR 0.16, p < 0.001; HR 0.25, p < 0.001, respectively), and
  • better LFS and OS, compared with CMT (LFS: HR 0.19, p < 0.001; OS: HR 0.30, p < 0.001) and auto-SCT (LFS: HR 0.35, p = 0.004; OS: HR 0.54, p = 0.04).
• Patients with recurrent MRD who received allo-SCT had similarly lower incidence of relapse, and better LFS and OS, compared with CMT and auto-SCT.
• Patients who were MRD negative after two courses of chemotherapy and received auto-SCT had
  • lower cumulative incidence of relapse, compared with CMT and allo-SCT (HR 0.25, p = 0.01; HR 0.08, p <0.001, respectively), and
  • better LFS and OS compared with CMT (LFS: HR 0.26, p = 0.004, OS: HR 0.22, p = 0.005) and allo-SCT (LFS: HR 0.21, p = 0.001; OS: HR 0.25, p = 0.001).
• For patients who were MRD negative after three courses of chemotherapy, those who received allo-SCT had lower cumulative incidence of relapse and better LFS compared with those who received CMT (HR 0.10, p = 0.04; HR 0.18, p = 0.01, respectively). Outcomes for the allo-SCT and auto-SCT groups were comparable.

Conclusion

In the entire population, post-remission treatment with allo-SCT and auto-SCT were advantageous with respect to incidence of relapse, LFS, and OS, compared with chemotherapy. Whilst LFS and OS were comparable for the allo-SCT and auto-SCT groups, allo-SCT had a lower incidence of relapse but higher transplant-related mortality than auto-SCT.

Based on the results of the dynamic MRD subgroup analysis, the authors suggest that chemotherapy and auto-SCT might be preferable for patients who are persistently MRD negative, whereas they recommend allo-SCT for patients who are persistently MRD positive or have recurrent MRD. Auto-SCT might be favored before allo-SCT for patients who are MRD negative after two courses of chemotherapy, whereas allo-SCT might be favored for patients who are MRD negative after three courses of chemotherapy.

The authors note the retrospective nature of this study, and a prospective multicenter trial is underway to validate these findings.