Prognostic Significance of Low Copy Number FLT3 and NPM1 mrd As Detected By Ultra-Sensitive Next Generation Sequencing

Background:

Measurable residual disease (MRD), the presence of leukemia cells below the threshold of morphologic detection, has become an increasingly important component of prognostication and disease monitoring in acute myeloid leukemia (AML). Over time, with the increasing affordability and availability of technologies such as ultra-sensitive next generation sequencing (NGS), depth of MRD detection has increased dramatically. However, the dynamics of MRD over time and clinical significance of MRD detection, particularly at low copy numbers, remain poorly understood. Thus, to improve our understanding of the clinical relevance of low copy number MRD detection in NPM1 and FLT3-ITD mutated AML, we retrospectively analyzed the clinical outcomes of a large cohort of AML patients treated at Stanford Health Care between 2018 and 2024 who had MRD testing for NPM1 and/or FLT3-ITD via ultra-sensitive NGS.

Methods:

Adult patients with NPM1 and/or FLT3-ITD mutated AML at diagnosis and at least one NGS-based MRD test performed by Invivoscribe were included in this retrospective cohort. The Invivoscribe NPM1 and FLT3-ITD MRD assays utilize multiplex polymerase chain reaction followed by NGS. Data, including AML diagnosis, NPM1 and FLT3-ITD mutational status, MRD testing results, and clinical outcomes were collected from the Stanford Health Care electronic medical record. For analyses based on MRD detection depth, patients were grouped into “no” MRD, “low level” MRD (defined as depth of detection at the level of 10^-5 or below), and “high level” MRD (any test with MRD detection greater than 10^-5). This distinction was based upon 10^-5 as the current, clinically-validated lower limit of MRD detection by the Invivoscribe assays. Outcomes assessed included overall survival (OS) and relapse-free survival (RFS). Subsequent MRD testing results were also evaluated for those patients with a “low level” MRD test. All statistical analyses were performed using R statistical software (v 4.3.1), and a p-value < 0.05 was defined as statistically significant.

Results:

A total of 125 patients and 571 MRD tests were included in the study. The NPM1 group included 66 patients and 181 MRD tests, and the FLT3-ITD group included 103 patients and 377 MRD tests (44 patients had both NPM1 and FLT3-ITD mutated AML and were included in both groups). For patients with NPM1 mutated AML, OS was similar across “high level,” “low level,” and “no” MRD groups (p=0.9); however, RFS was significantly worse in the “high level” MRD group compared to the “no” MRD and “low level” MRD groups (p=0.01). Notably, there was no difference in RFS between the “no” and “low level” MRD groups. These findings were recapitulated in patients with FLT3-ITD mutated AML: there was no significant difference in OS between the “high level,” “low level,” and “no” MRD groups (p=0.3), but the “high level” MRD group had significantly worse RFS than both the “no” and “low level” MRD groups (p < 0.001). Again, no difference in RFS was seen in “no” MRD vs “low level” MRD groups. For all patients with an MRD test with a value in the “low level” range and with follow up MRD testing available (NPM1 = 15, FLT3-ITD = 21), subsequent MRD testing was negative in most patients (NPM1: 11/15 [73%], FLT3-ITD: 12/21 [57%]). However, a smaller proportion of patients either remained MRD positive in the “low level” range (NPM1: 2/15 [13%], FLT3-ITD: 4/21 [19%]) or later developed “high level” MRD (NPM1: 2/15 [13%], FLT3-ITD: 5/21 [24%]).

Conclusions:

To our knowledge, our study is the first to assess the clinical impact of MRD detected below the clinically-validated limit of detection by ultra-sensitive NGS. Findings from our retrospective study were consistent with previously published data suggesting the presence of “high level” MRD (i.e. above the current clinically-validated limit of detection) is associated with increased risk of AML relapse in both NPM1 and FLT3-ITD mutated AML. Interestingly, however, in both NPM1 and FLT3-ITD mutated AML, low copy number MRD was not associated with either worse OS or RFS in our cohort. While most patients with “low-level” MRD positivity later have negative MRD testing, there is a subset of patients that go on to develop “high-level” MRD and are thus at increased risk of relapse. These data suggest that MRD positivity below the clinically-validated limit of detection is challenging to interpret, and that longitudinal MRD monitoring is vital for disease surveillance.