Outcomes following MRD relapse in AML patients monitored by droplet digital PCR

Abstract Introduction: European Leukemia Network (ELN) guidelines recommend PCR-based measurable residual disease (MRD) monitoring for patients with acute myeloid leukemia (AML) with NPM1 mutation (NPM1mut), t(8;21)/RUNX1::RUNX1T1 or inv(16)/CBFB::MYH11 during and after treatment. This may enable early detection of MRD relapse and intervention prior to morphologic relapse. MRD testing for these AML subtypes has been performed for patients treated at the Leukemia and Bone Marrow Transplant (L/BMT) Program of British Columbia (BC) since 2022. We analyzed patient outcomes following MRD relapse and treatment in this patient cohort. Methods: A retrospective study was conducted on consecutive adult patients in BC, Canada diagnosed with AML between April 2022 and October 2024 who received frontline intensive chemotherapy and had ≥2 MRD measurements. Patients not achieving complete response (CR)/CR with incomplete count recovery (CRi) or those receiving non-intensive therapy were excluded. MRD testing was performed by the Cancer Genetic and Genomics Laboratory, BC Cancer using droplet digital PCR (ddPCR). The routine limit of reliable detection for the assay is 10-4.5 or better. Frequency of monitoring followed ELN 2021 guidelines with monthly peripheral blood testing post-consolidation or stem cell transplant (SCT). Overall survival (OS) was defined from date of diagnosis to date of death. MRD relapse was defined as a conversion from a negative to detectable MRD, or an increase of MRD ≥1 order of magnitude (≥1log10). Survival analysis was performed using the Kaplan-Meier survival estimator and the log-rank test was used to test for differences. Results: Among 97 patients who met inclusion criteria, the median age at diagnosis was 57 years (range 19-72); 55 (57%) were male. There were 75 (75%) patients with NPM1mut, 16 (17%) with inv(16), and 6 (6%) with t(8;21). Based on ELN 2022 criteria, 68 (70%) were favourable risk, 28 (29%) intermediate risk, and one (1%) adverse risk. Forty-one (42%) patients underwent allogeneic SCT (alloSCT) in CR1, 8 (8%) in CR2; one had autologous (auto) SCT in CR1. Median follow-up was 22 months (range 4.5-65). The 1-, 2-, and 3-year OS was 94%, 93%, and 71% respectively, with no significant OS difference among AML subtypes. Twenty-two (23%) patients had MRD relapse after chemotherapy, with a median time from CR/CRi to MRD relapse of 7.7 months (range 2-24). Sixteen (73%) received pre-emptive treatment: intermediate or high dose cytarabine (INDAC/HIDAC) (n=11), azacitidine (n=2), azacitidine + venetoclax (n=2), and FLAG-Ida (n=1); 15 (68%) later underwent SCT (14 allo, 1 auto). The overall response rate (ORR) to pre-emptive treatment among the 14 patients evaluable for MRD was 50% (MRD negative n=2, ≥ 3 log10 reduction n=2, 1-3 log10 reduction n=3). One patient relapsed post-treatment (alloHSCT) and died. Among 15 patients without relapse, 14 remain alive. Among 6 patients not receiving pre-emptive therapy, 3 relapsed and died. Median OS of patients with MRD relapse after chemotherapy was not reached; the 1- and 2-year OS was 95% and 67% respectively. The 2-year OS was higher with pre-emptive treatment vs. no treatment group (80% vs. 27%, p=0.012). Thirteen patients had MRD relapse post-alloSCT with a median time from alloSCT to MRD relapse of 5 months (range 2-15). Twelve patients received pre-emptive therapy: azacitidine (n=1), azacitidine + donor lymphocyte infusion (DLI, n=4), azacitidine + venetoclax + DLI (n=1), withdrawal of immunosuppression (n=2), and gilteritinib (n=4). The ORR was 60% among 10 patients with evaluable MRD responses (MRD negative n=5, ≥ 3 log10 reduction n=1). Three patients relapsed post-treatment and died; 8 of the remaining 9 relapse-free patients remain alive. One patient with MRD relapse post-alloSCT received no therapy and died after morphologic relapse. The median OS of patients with MRD relapse after transplant was not reached and the 1- and 2-year OS was 84% and 65% respectively. For the entire cohort, patients with initial MRD relapse had significantly longer OS compared to those with morphologic relapse without prior MRD relapse (p=0.004). Conclusion: Our results suggest that ddPCR-based MRD monitoring per ELN guidelines allows for pre-emptive treatment in most patients. Early results suggest patients receiving treatment for MRD relapse had favorable outcomes compared to those untreated or those with morphologic relapse without prior MRD relapse.