Pretransplant MRD detection of fusion transcripts is strongly prognostic in KMT2A-rearranged acute myeloid leukemia Free

TO THE EDITOR:

KMT2A-rearranged (KMT2Ar) acute myeloid leukemia (AML) is generally associated with adverse prognosis, with allogeneic hematopoietic cell transplant (HCT) in first remission (CR1) considered standard of care in most cases.^1,2 Despite HCT, ∼40% of patients are destined to have relapsed disease.³ For those in remission prior to HCT, small retrospective studies suggest higher relapse risk if KMT2Ar molecular measurable residual disease (MRD) is detected prior to transplant.^4-7KMT2Ar is stable at relapse, making it a suitable MRD marker for quantitative monitoring.⁸ One barrier to implementation of KMT2Ar MRD monitoring, however, is breakpoint heterogeneity within the KMT2A gene itself and its association with a diverse array of partner fusion genes. Despite >100 translocation partner genes identified, >70% of KMT2A fusions implicated in AML involve either MLLT3 [t(9;11)], ELL [t(11;19) (q23;p13.1)], AFDN [t(6;11)], or MLLT10 [t(10;11)]. Furthermore, the majority of KMT2A genomic breakpoints occur within the major breakpoint cluster regions of introns 8, 9, 10, and 11.^9,10 

This study analyzes the clinical relevance of pre-HCT KMT2Ar MRD detection by real-time quantitative polymerase chain reaction (qPCR) and the prognostic impact on post-HCT outcome in adults with KMT2Ar AML. We also describe a novel reverse transcription droplet digital PCR (RT-dPCR) assay that allows a large majority of patients with KMT2Ar AML to be quantitatively monitored using a simplified multiplexed assay. The relevance of this study is enhanced by clinical development of menin inhibitors showing promising efficacy in patients with relapsed or refractory KMT2Ar AML and potential to detect and preemptively target KMT2Ar MRD using targeted therapies.^4,11,12 

A cohort of 64 patients treated in Australia, United Kingdom, and Taiwan were included. Patients had KMT2Ar confirmed by karyotyping and fluorescence in situ hybridization or RNA sequencing at diagnosis. Within 100 days pre-HCT and with no intervening therapy between sample collection and transplant conditioning, pre-HCT MRD data were available in 45 patients, and MRD assessment was undertaken on archived samples from 19 patients. Association with clinical outcome was performed through retrospective review. Pre-HCT RNA from bone marrow (n = 59) or peripheral blood (if bone marrow was unavailable; n = 5) were assessed by RT-qPCR (n = 55)¹³ or RT-dPCR (n = 9) using a novel multiplexed assay combining a master mix of 3 forward primers to KMT2A gene and a pooled array of 7 reverse primers (targeting common fusion partner genes MLLT3, ELL, AFDN, and AFF1), permitting quantitation of common KMT2Ar fusion genes in a single assay (see supplemental Methods, available on the Blood website). Reverse primers to MLLT10 and MLLT1 were run separately and not included in the reverse primer pool due to the high degree of breakpoint heterogeneity. Results were expressed as KMT2A::X copies per 100 ABL1. Assay sensitivity was ∼0.001%. Survival outcomes were estimated by Kaplan-Meier for relapse-free survival (RFS, HCT date to relapse or death or last follow-up) and overall survival (OS, to date of death or last follow-up). Cumulative incidence of relapse (CIR) was calculated using cumulative incidence function with nonrelapse mortality as a competing risk. Categorical variables were compared using the Fisher exact test; continuous variables were compared using the Mann-Whitney U test. Cox proportional hazards was used to assess variables impacting survival. Variables with P < .1 on univariate analysis were taken forward to multivariate analysis. The study was approved by human research ethics committees at each center.

Median age was 43 years (17-70), 8% had therapy-related AML, and the remainder had de novo disease (Table 1). All patients had received prior intensive chemotherapy and were in remission at the time of HCT; 78% received transplants in CR1, 49% from an unrelated donor, 60% received T-cell depletion, and 46% received myeloablative conditioning. KMT2A partner genes are summarized in Table 1. Median time between MRD sampling and HCT was 24 days (1-93).

Pre-HCT, 42% (27 of 64) were KMT2Ar MRD-positive, with the median MRD level 0.1200% (range, 0.0056%-35.0394%) (Figure 1A). Apart from a higher proportion of t(11;19)/KMT2A::MLLT1 in the MRD-positive cohort, and t(9;11) and t(10;11) in the MRD-negative cohort, no other baseline characteristics were found to be significantly associated with KMT2Ar MRD-positive status pre-HCT (Table 1).

CIR was significantly increased among patients KMT2Ar MRD-positive vs MRD-negative pre-HCT (2-year CIR 75% vs 25%, P = .0004) (Figure 1B). KMT2Ar as a marker of relapse was highly stable, with all assessable patients (n = 22) retaining the same fusion gene at relapse. With 37 months median follow-up time, survival was significantly inferior in patients MRD-positive vs MRD-negative pre-HCT, with 2-year RFS 17% vs 59% (P = .001) and 2-year OS of 39% vs 67% (P = .012) (Figure 1C-D). Pre-HCT MRD-positive status was the only determinant of inferior RFS (hazard ratio 2.6, 95% confidence interval 1.3-5.1, P = .007) and OS on univariate and multivariate analyses. Two-year RFS if KMT2A::X MRD was negative, 0.001% to <0.1%, 0.1% to <1% or ≥1% were 59%, 42%, 0% or 0%, respectively (Figure 1E). Myeloablative conditioning did not abrogate the inferior survival outcome of pre-HCT MRD-positive status (supplemental Figure 2). Insufficient cases were available to confirm association between co-mutation profile and risk of pre-HCT MRD or survival outcome (supplemental Figure 4). Survival was also significantly inferior for pre-HCT MRD-positive status when analysis was restricted to patients who underwent CR1 transplant (2-year RFS 17% vs 62%, P = .0047), to exclude the confounding adverse impact on survival of transplant in CR2 (Figure 1F).

Survival analysis of 22 patients with t(9;11), categorized as European LeukemiaNet intermediate risk, was explored to determine if pre-HCT MRD status could refine prognostication.¹ Pre-HCT MRD detection of KMT2A::MLLT3 MRD was significantly associated with poor survival, with 2-year RFS and OS both 11%, compared with 68% and 67% for patients with KMT2A::MLLT3 MRD-negative status (P = .032 and .048, respectively) (supplemental Figure 3). Whether patients with KMT2A::MLLT3 MRD-negative status at the end of treatment could be spared CR1 transplant requires evaluation in future larger confirmatory cohorts.

To enable KMT2Ar MRD monitoring, baseline bone marrow or peripheral blood RNA should be routinely stored at diagnosis. KMT2Ar at diagnosis is generally identified by either cytogenetic and fluorescence in situ hybridization (specific intron-exon breakpoint unknown) techniques or through use of either RNA sequencing or a targeted RNA fusion panel. The multiplex RT-dPCR assay presented in this article may be used to monitor patients if the KMT2A fusion partner involves either chromosome 9, 19p13.1, 6, or 4 and a diagnostic screening sample is confirmed to be positive using the multiplexed technique. If the KMT2A fusion partner involves chromosomes 10 or 19p13.3, probes to common break points involving these partner genes are required for RT-dPCR or qPCR monitoring. Patient-specific primer or probes may be required for other rare translocations or breakpoint locations outside of the commonly affected regions.

This study, the largest that we know of, highlights the dismal prognosis associated with KMT2Ar MRD detected pretransplant. With next-generation sequencing-based targeted RNA fusion panels increasingly performed at diagnosis, KMT2Ar detection and quantitative MRD tracking will become more commonplace. We also describe the design and validation of a multiplexed RT-dPCR assay able to monitor ∼60% of KMT2Ar (or ∼80% excluding MLLT10 or MLLT1) occurring in adult AML.⁹ These strategies now pave the way for MRD-directed pre- and post-HCT intervention using menin-targeted therapies, aimed at reducing relapse risk and improving overall patient outcome.