Latency between detectable low-level measurable residual disease (MRD) by next-generation sequencing and overt relapse following flow cytometry MRD-negative remission in patients with acute lymphoblastic leukemia: Implications for the optimal frequency of MRD monitoring

Abstract Background: In patients (pts) with acute lymphoblastic leukemia (ALL), MRD by next-generation sequencing (NGS) for IG/TR has become a key marker of response and predicts relapse better than MRD by multiparameter flow cytometry (MFC). While NGS MRD is a helpful prognostic tool, there is uncertainty surrounding both the time from a positive NGS MRD sample to relapse and the optimal NGS MRD monitoring interval for pts in first complete remission (CR1). We therefore sought to retrospectively determine the latency between NGS MRD positivity and overt relapse in pts in CR1 who were MRD-negative by MFC. Methods: Eligible pts had a diagnosis of B-cell or T-cell ALL, experienced first relapse (defined as blasts 5% or extramedullary EM relapse) after MFC MRD negativity, and had available banked samples for NGS MRD assessment by clonoSEQ. Pts with Philadelphia chromosome (Ph)-positive ALL were eligible if they were also MRD negative by PCR for BCR::ABL1 prior to relapse. Available bone marrow samples within 1 year prior to first relapse were retrospectively evaluated for NGS MRD for IG/TR , which has a sensitivity of 1x10-6. Time to relapse was modeled by Kaplan-Meier analysis. Results: 14 pts diagnosed between 2012 and 2018 met inclusion criteria: 3 (21%) had T-cell ALL, 9 (64%) had Ph-negative B-cell ALL, and 2 (14%) had Ph-positive B-cell ALL. The median age was 35 years (19-68). 6 pts (43%) had diploid cytogenetics, 2 (14%) had complex cytogenetics, 2 (14%) were Ph-positive, and 4 (29%) had miscellaneous cytogenetic abnormalities. Of 10 pts tested, 3 (30%) had TP53 mutation. For induction therapy, 8 pts (57%) received a Hyper-CVAD-based regimen (none of whom received frontline blinatumomab or inotuzumab ozogamicin INO), 3 (20%) received a mini-CVD-based regimen (of whom 2 received INO and 1 venetoclax) and 3 (20%) received augmented BFM. Among the 2 pts with Ph-positive ALL, 1 received dasatinib and 1 received ponatinib. All pts achieved MRD negativity by MFC +/- PCR for BCR::ABL1. 4 pts (28%) underwent allogeneic hematopoietic stem cell transplant in CR1. The median time to relapse from diagnosis was 14.6 months (range, 8.6 – 33.1 months), and the median time from the last MFC MRD negative assessment to relapse was 3.6 months (range, 1.0-7.0 months). 3 pts (21%) had isolated central nervous system (CNS) relapse, 1 had isolated non-CNS EM relapse, and 1 had combined bone marrow and non-CNS EM relapse. Among the entire cohort, there were 36 MFC MRD assessments in the 1 year prior to relapse, with a median of 3 per pt (range,1-4). For 25 (69%) of these MFC MRD-negative timepoints, banked specimens were available for NGS MRD assessment, with a median of 2 samples per pt (range, 1-4). Among the 25 NGS MRD samples, 12 (48%) were NGS MRD-positive and 13 (52%) were NGS MRD-negative The median quantified NGS MRD was 9.7 x 10-6 (range, 1 to 6053 x 10-6). 3 positive samples (25%) were 1 x 10-6. Overall, 10 pts (71%) were found to be NGS MRD-positive prior to relapse. 2 of 3 pts with isolated CNS relapse (67%) and both pts with non-CNS EM relapse had positive NGS MRD. The median quantified NGS MRD 6 months prior to relapse was 2 x 10-6 (range, 1 to 10x 10-6), while the median quantified NGS MRD within 6 months of relapse was 86 x 10-6 (range, 1 to 6053 x 10-6) (p=0.014). In both pts with 1 NGS MRD-positive sample, the quantification of NGS MRD increased prior to relapse. The median time to relapse after the last NGS MRD-negative sample was 4.8 months (range, 3.5 –9.5 months); no pt relapsed within 3 months of an NGS MRD-negative sample. In contrast, the median time from NGS MRD-positivity to relapse was 3.6 months (range, 1.0 – 10.2 months). Conclusion: Low-level MRD detectable by NGS for IG/TR was observed in the majority of pts with ALL in MFC MRD-negative CR1 prior to first relapse. The median time between NGS MRD-positivity and overt relapse was 3.6 months, suggesting that rapid intervention is needed in the setting of detectable NGS MRD to prevent relapse. In contrast, no pt relapsed within 3 months of last NGS MRD-negative result. Our data suggest that monitoring of NGS MRD approximately every 3 months would be likely to detect NGS MRD recurrence prior to first relapse in most pts. Additional studies will be needed to identify the optimal frequency of MRD assessment in other settings (e.g. peri-transplant or in pts in later lines of therapy).