Measurable Residual Disease Detection on Day 30 Post Haematopoietic Stem Cell Transplantation Predicts Clinical Outcome in Acute Myeloid Leukemia

Background: Allogeneic haematopoietic stem cell transplantation (allo-HSCT) is the mainstay of post-remission therapy for acute myeloid leukaemia (AML) but post HSCT relapse is a major cause of treatment failure. Measurable residual disease (MRD) detection in AML has an established role in predicting post-chemotherapy relapse and it informs decision of post-remission therapy. However, its role in post HSCT is less studied. In this study, the prognostic value of MRD detection during early post HSCT, when modulation of graft-versus-leukaemia effect is possible, was investigated.

Methods: Adult AML patients with allo-HSCT performed in Queen Mary Hospital in Hong Kong after 2003 at complete remission were studied. MRD detection was performed from archival samples by droplet digital PCR (ddPCR) at 9 recurrent hotspot mutations, including NPM1 Type A, DNMT3A R882H, IDH1 R132H/C/S, IDH2 R140Q/R172K, FLT3 D835Y and NRAS G12D, or at patient-specific mutations on bone marrow (BM) samples at defined time points after HSCT. The study was approved by the Institutional Review Board (IRB) of Hospital Authority /HKU (HKU/HA HKW UW14-430 and UW14-639)

Results: Day 30 post-HSCT BM samples were available in 123 AML patients undergoing allo-HSCT from unrelated (N=50) and related (N=73) donors. Conditioning regimens were myeloablative in 79 patients and reduced intensity in 44 patients. MRD monitoring was performed at the 9 recurrent hotspots in 117 patients (NPM1 Type A, N=69; DNMT3A R882H, N=33; IDH1 R132, N=15; IDH2 R140Q/R172K, N= 27; FLT3 D835Y, N=11; NRAS G12D, N=14) and at patient-specific mutations in 7 patients. In 47 patients, there were more than 1 MRD markers and patients with detectable mutations in any one marker were considered MRD positive. MRD was positive in 37 (30%) patients. Post-HSCT Day 30 MRD positivity was associated with inferior post-HSCT leukaemia free survival (LFS) (median: 20.86 vs 110.06 months, p-value=0.00044) and overall survival (OS) (median: 45.6 vs 144.66 months, p-value=0.0039). MRD positivity was predictive of post-HSCT 3-year relapse risk in both NPM1 (Hazard Ratio(HR) = 5.2, CI: 2.18-12.44, p value <0.001) and non-NPM1 markers (HR = 2.34, CI: 1.14-4.8, p value = 0.02). Multivariate analysis was performed to investigate factors that affect post-HSCT LFS and OS, including post HSCT Day 30 MRD status, patient clinical and demographic information, European LeukemiaNet (ELN) 2022 risk stratification and status at HSCT. Post-HSCT Day 30 MRD positivity was the only statistically significant factor associated with inferior post HSCT LFS (HR = 2.85, CI: 1.45-5.63, p value = 0.0025) and OS (HR = 2.67, CI: 1.27-5.63, p value = 0.0097).

Conclusions: MRD positivity at day 30 post HSCT BM sample is predictive of early relapse within 3 years for both NPM1 and non-NPM1 mutations. The measurement identifies the group of MRD-positive patients who are at risk of early post-HSCT relapse and allows future investigation of early interventions to prevent post-HSCT relapse.