Background
Phase 2, open-label, multicentre, single-arm trial (RELAZA-2) at 9 German centres. Adults ≥18 with advanced MDS or AML in morphological CR after chemotherapy or allogeneic HSCT were prospectively screened for MRD over 24 months by quantitative PCR (NPM1, DEK-NUP214, RUNX1-RUNX1T1, CBFb-MYH11) or donor chimaerism (post-alloSCT). Of 198 screened, 60 (30%) developed MRD positivity and 53 (88% of those eligible) began preemptive azacitidine to prevent overt haematological relapse. Published Lancet Oncology 2018.
Interventions and follow up
Regimen: Azacitidine 75 mg/m²/d SC days 1–7 every 4 weeks for 6 cycles (preemptive), then maintenance up to 24 cycles; treatment triggered by MRD positivity in MDS/AML CR
Primary endpoint: Proportion relapse-free and alive at 6 months after starting preemptive treatment
Median follow-up: 13 months (IQR 8.5–22.8) after treatment start
Primary endpoint: Proportion relapse-free and alive at 6 months after starting preemptive treatment
Median follow-up: 13 months (IQR 8.5–22.8) after treatment start
Results
Relapse-free and alive at 6 months (primary, N=53): 58% (31/53), P<.0001 vs null of ≤30%
12-month RFS (MRD-treated): 46% (95% CI 32–59)
12-month RFS (concurrent MRD-negative reference): 88% (95% CI 82–94); HR 6.6 (95% CI 3.7–11.8), P<.0001
MRD response: ~50% achieved MRD negativity by cycle 6; de-escalation offered
12-month RFS (MRD-treated): 46% (95% CI 32–59)
12-month RFS (concurrent MRD-negative reference): 88% (95% CI 82–94); HR 6.6 (95% CI 3.7–11.8), P<.0001
MRD response: ~50% achieved MRD negativity by cycle 6; de-escalation offered
Adverse events
Haematologic (grade 3–4): Neutropenia 85% (45/53) — most common AE; thrombocytopenia and anaemia also reported
Infectious/fatal: One death from infection considered possibly treatment-related; no other unexpected toxicities, consistent with the known azacitidine profile
Infectious/fatal: One death from infection considered possibly treatment-related; no other unexpected toxicities, consistent with the known azacitidine profile
Conclusions
Preemptive azacitidine in MRD-positive AML/MDS patients in morphological remission met its primary endpoint — 58% relapse-free and alive at 6 months, exceeding the 30% historical null. Continuous MRD negativity in untreated patients was highly prognostic (12-month RFS 88%), validating MRD-guided preemptive intervention.
Key Limitations
Single-arm design with a historical null comparator rather than a randomized control; no formal randomized comparison against watchful waiting. Short median follow-up (13 months) limits assessment of durability. Heterogeneous population (mixed MDS and AML, post-chemo and post-alloSCT, multiple MRD markers). Modest sample (N=53 treated). Optimal duration of preemptive therapy and the long-term survival impact remain undefined.
Clinical Context
RELAZA-2 provided proof-of-concept that MRD-guided preemptive azacitidine can delay or prevent overt relapse in MDS/AML, supporting a paradigm of measurable-residual-disease-directed therapy rather than waiting for haematological relapse. ELN 2021/2022 MRD guidelines recognize molecular MRD as prognostic and a basis for preemptive intervention; the approach is most established in NPM1-mutated AML and post-alloSCT chimaerism monitoring. Azacitidine is not regulatory-approved specifically for the MRD-preemptive indication.