Abstract
Treated AML patients often have measurable residual disease (MRD) due to persisting low-level clones. This study assessed whether residual post-treatment somatic mutations, detected by NGS, were significantly prognostic for subsequent clinical outcomes. AML patients (n = 128) underwent both pre-and post-treatment testing with the same 42-gene MRD-validated NGS assay. After induction, 59 (46%) patients were mutation-negative (0.0024 VAF detection limit) and 69 (54%) had ≥1 persisting NGS-detectable mutation. Compared with NGS-negative patients, NGS-positive patients had shorter overall survival (17 months versus median not reached; P = 0.004; hazard ratio = 2.2 [95% CI: 1.3-3.7]) and a shorter time to relapse (14 months versus median not reached; P = 0.014; HR = 1.9 [95% CI: 1.1-3.1]). Among 95 patients with a complete morphologic remission (CR), 43 (45%) were MRD-positive by NGS and 52 (55%) were MRD-negative. These MRD-positive CR patients had a shorter overall survival (16.8 months versus median not reached; P = 0.013; HR = 2.1 [95% CI: 1.2-3.9]) than did the MRD-negative CR patients. Post-treatment persisting MRD positivity, defined by the same NGS-based test used at diagnosis, is thus a more sensitive biomarker for low-level leukemic clones compared to traditional non-molecular methods and is prognostic of subsequent relapse and death.