P509: clonal hematopoiesis-associated mutations as measurable residual disease markers in acute myeloid leukemia patients following allogeneic stem cell transplantation

Background: Clonal hematopoiesis (CH)-associated mutations (mut) are frequent in acute myeloid leukemia (AML), appear early in leukemogenesis, and often persist in remission after chemotherapy. Following allogeneic hematopoietic stem cell transplantation (HSCT), which replaces the patients’ hematopoiesis, CH mut may present useful measurable residual disease (MRD) markers. Aims: To evaluate patient specific CH-associated mut as MRD markers in AML patients (pts) after allogeneic HSCT. Methods: We analyzed 31 AML pts with CH-associated mut present at diagnosis (DNMT3A: n=17; SRSF2: n=9; IDH2: n=7; ASXL1: n=4; TET2: n=2, and JAK2: n=1), 23 pts had 1, 7 had 2, and 1 patient had 3 CH mut. All received a myeloablative (36%), reduced-intensity (48%), or non-myeloablative (16%) HSCT (median age 58, range 32-72 years). 84% were transplanted in complete remission (CR) or CR with incomplete recovery (CRi). European Leukemia Net (ELN) risk was 23% favorable, 37% intermediate, and 40% adverse. Mut-specific digital droplet PCR assays were developed using a competitive probe-approach. MRDpos was defined as variant allele frequency (VAF) ≥2% for ASXL1 & ≥0.05% for all other analyzed mut. 209 samples with a median of 6 (range 1-13) blood or bone marrow samples per patient were available after HSCT with a median of 1.2 years follow-up time for pts alive. Results: Prior to HSCT, 89% of pts remained CH-associated mut positive in CR/CRi. 11 pts relapsed after HSCT (35%), all with the same CH-associated mut present at diagnosis (median VAF at relapse 15.6, range 0.2-41.2%). Of those, 6 pts had at least 1 MRD sample available prior to relapse. In 5 pts, impeding relapse was predicted by a CH-associated mut MRD conversion with a median VAF of 0.27 (range 0.06-1.0)% at a median of 103 (range 14-199) days prior to relapse. The lead time was longest for a DNMT3A mut patient (199 days) & shortest for a SRSF2 mut patient (14 days). For the patient relapsing without prior CH MRD positivity, the last available sample was taken 169 days before relapse & was CH MRDneg. 20 pts retained remission during follow-up. Of those, 17 pts remained CH MRDneg in all samples (median samples per analyzed mutation 6, range 1-13). 2 pts had 2 MRDpos samples directly after HSCT but became MRDneg in subsequent samples. 1 patient received Enasidenib maintenance after HSCT with fluctuating CH-MRD values after HSCT (SRSF2 & IDH2, VAFs <0.05-0.34% & <0.05-0.14%, respectively). Of the pts with MRD samples within the first year after HSCT available, pts with at least 1 MRDpos sample after HSCT had a significantly higher cumulative incidence of relapse (CIR, P=.008, Figure 1A) & shorter relapse-free survival (RFS, P=.02, Figure 1B). 8 pts had concurrent NPM1 mut (n=4) or RUNX1 mut (n=4), which showed concordant MRD results in 7 pts (4 remained MRDneg, 3 concurrently converted MRDpos). 1 patient suffered a NPM1 MRDneg relapse which was predicted by 2 re-detected CH-associated mut (DNMT3A mut & IDH2 mut, 112 & 199 days prior to relapse, respectively, Figure 1C). Image:Summary/Conclusion: Prior to HSCT most CH-associated mut remain detectable in pts in CR/CRi. However, pts with at least one CH MRDpos sample following allogeneic HSCT had higher CIR & shorter RFS. All relapsing pts were positive for the same CH-associated mut present at diagnosis. 5 of 6 showed increasing VAFs prior to hematologic relapse & the majority of pts retaining remission remained CH MRDneg. While CH-associated mut in chemotherapy treated AML pts have limited MRD value, these mut are feasible MRD markers after HSCT & may inform relapse preventing therapy decisions.