P497: clinical outcomes of npm1-mutated acute myeloid leukemia (aml) patients with measurable residual disease (mrd) relapse after achievement of a first mrd-negative remission

Background: Persisting MRD is a strong prognostic factor in NPM1-mutated (mut) AML patients (pts) and a new risk factor according to the European LeukemiaNet (ELN) 2022 classification. Data on MRD relapse in morphologic complete remission (CR) after achievement of MRD-negativity (neg) in these pts is limited. Aims: To evaluate outcomes of NPM1-mut AML pts suffering MRD relapse after having achieved a MRD-neg CR. Methods: Among 82 NPM1-mut AML pts, 67 achieved a MRD-neg CR at ≥1 timepoint. Of those, 31 pts had a MRD relapse during/after 7 + 3-based chemotherapy (Ctx, n=29) or upfront azacytidine/venetoclax (aza/ven, n=2) and were retrospectively analyzed. Sequential NPM1 MRD analyses were performed in bone marrow or peripheral blood samples using custom digital droplet PCR assays. NPM1/ABL1 levels ≥0.01% were defined as MRD-pos. Median follow-up after MRD relapse was 1.4 years. Results: The median time from MRD-neg CR to MRD relapse was 93 (range 14-932) days. At the time of MRD relapse, 7 pts (23%) were still on treatment, 12 pts (39%) were under active surveillance after completion of Ctx and 12 pts (39%) had undergone allogeneic hematopoietic stem cell transplantation (HSCT). At MRD relapse first pos NPM1/ABL1 levels were median 0.036 (range 0.01-18.9)% and increased to a maximum of median 2.09 (range 0.01-122.4)% while still in morphologic CR. 17 pts (55%) suffered morphologic relapse after a median of 125 (range 28-510) days after MRD relapse and 2 pts (6%) suffered non-relapse mortality (NRM). Cumulative incidence of relapse (CIR) and overall survival (OS) 2 years after MRD relapse were 64% and 59%, respectively (Fig. 1A). CIR (P=.88) and OS (P=.60, Fig. 1B) after MRD relapse did not differ between pts with favorable (i.e. no FLT3-ITD) or intermediate ELN2022 (i.e. concurrent FLT3-ITD, Fig. 1B) at diagnosis, or between pts who did or did not receive a HSCT prior to MRD relapse (CIR P=.65, OS P=.90). All 7 pts who were still on treatment at MRD relapse continued their therapy (3 under consolidation, 1 under maintenance, 3 directly before HSCT). Of the pts under surveillance, 11 received no treatment for their MRD relapse (5 eventually suffered morphologic relapse), and 13 received preemptive treatment (5 within a clinical trial, 3 donor lymphocyte infusion, 2 aza/ven, 2 aza, 1 ven) after a median of 45 (range 26-181) days after MRD relapse (9 relapsed, 3 converted to MRD-neg CR, 1 died from NRM). Both, the NPM1/ABL1 levels at MRD relapse and the maximum in morphologic CR were significantly higher in pts receiving preemptive MRD-guided therapy than in those who did not (P=.03 and P=.001, respectively, Fig. 1C). There was no different CIR (P=.44) or OS (P=.10) in AML pts who continued their (planned) treatment, received a preemptive MRD-guided therapy, or surveillance alone (Fig. 1D). Of note, 12 pts received aza/ven-based treatments, 2 for MRD relapse and 10 later after morphologic relapse. Of the 11 ven-naïve pts, 10 achieved a second MRD-neg CR (4 after 1 cycle, 6 after 2 cycles), and 1 a morphologic CR after 1 cycle despite prior aza exposure in 7/10 pts. Summary/Conclusion: Outcomes of NPM1-mut AML pts with MRD relapse did not differ with regard to ELN2022 risk at diagnosis or prior to HSCT. In this retrospective analysis with heterogeneous preemptive treatments there was no benefit from MRD-guided therapy vs therapy for morphologic relapse. However, response rates to aza/ven in ven-naïve pts were promising irrespective of MRD/morphologic relapse or prior aza exposure. Our data underlines the need of studies testing if MRD-guided aza/ven improves outcomes in NPM1-mut AML pts with (MRD) relapse.Keywords: Molecular relapse, Minimal residual disease (MRD), AML