Paper Details

ABSTRACT

BACKGROUND: Asciminib (ASC), a STAMP inhibitor targeting the BCR::ABL1 myristoyl pocket, is increasingly used across treatment lines in chronic myeloid leukemia (CML). Beyond BCR::ABL1 kinase-domain (KD) mutations, somatic mutations in myeloid genes occur in CML, yet their behavior under ASC is unknown. We previously described five distinct dynamics patterns of somatic mutations following tyrosine kinase inhibitor (TKI) therapy (Kim et al., Blood 2017). In newly diagnosed CML-CP, the ASC4FIRST study (Branford et al., ASH 2025) identified ASXL1 mutation as the most frequent additional genomic alteration at baseline, with higher allelic burden on ASC (24% VAF) than investigator-selected TKI (10%) and an association with treatment failure, indicating that epigenetic (ASXL1) clones remain clinically relevant even under ASC. We characterize non-ABL1 somatic mutation burden, variant allele frequency (VAF) dynamics, and pathway-based patterns of somatic mutations in an ASC-treated cohort, and test their longitudinal dynamic patterns.

METHODS/PATIENTS: Eighty-five ASC-treated CML patients (median 2 prior TKI lines, range 0-5; 78/85 prior TKI-exposed) underwent serial single-molecule molecular inversion probe (smMIP) sequencing of a 40-gene myeloid panel at diagnosis/pre-ASC (Dx), 6, 12, and >12 months. Variant positivity was defined at 2% VAF. Genes were grouped into TCGA-AML functional pathways (DNA methylation, chromatin modification, activated signaling, transcription factor, tumor suppressor, spliceosome, cohesin). Per-variant serial VAF trajectories were examined, and unsupervised hierarchical clustering (VAF max-normalized across three timepoints, Dx/6mo/12mo [or >12mo]) was performed as previously described (Kim et al., Blood 2017). Endpoints included treatment failure leading to discontinuation, and progression.

RESULTS: 63/85 cases were sequenced (161 samples). Overall, 38/63 cases (60%) were variant-positive at some timepoint during ASC treatment. Per-sample positivity was Dx 18/59 (31%), 6mo 11/40 (28%), 12mo 11/39 (28%), >12mo 11/23 (48%), and stable over time. By biological pathway, detected-VAF medians were highest for DNA methylation (DNMT3A/TET2; 4.4%) and transcription factor (TF; RUNX1/CEBPA/GATA1/2; 4.3%), intermediate for chromatin modification (ASXL1/EZH2; 3.3%), and lowest for activated signaling excluding ABL1 (FLT3/RAS/PTPN11; 2.2%, near the 2% floor); each pathway's VAF was stable across timepoints, though some clones showed a rising pattern. At the gene level, DNMT3A/TET2 formed the highest-burden clones (median 5.9% at Dx, 6.4% at >12mo; range to 30%), and ASXL1 was consistently detected (median 4.6% at Dx, range to 47%), each stable over time. Unsupervised clustering of 72 higher-confidence clones (≥5% VAF) yielded five patterns: Pattern 1, cleared-after-Dx (n=22; TF-dominant 45%, DNA methylation 27%); Pattern 2, transient-at-6mo (n=16; DNA methylation 31%, TF 19%); Pattern 3, rising/sustained (n=9; mostly epigenetic at 67%, e.g. ASXL1/TET2/IKZF1); Pattern 4, acquired-at-12mo (n=20; DNA methylation 30%, activated signaling 25%, chromatin 20%); and Pattern 5, fluctuating (n=5). Unlike the TKI era, no stable-persistence pattern was observed, and ~85% of variants were detected at a single timepoint. Notably, epigenetic clones expanded to high VAF in individual patients who eventually discontinued ASC for treatment failure (ASXL1 to 42% VAF; DNMT3A to 30% VAF), whereas peak epigenetic-clone VAF was comparable between patients who discontinued or progressed and other patients. Further analysis correlating BCR::ABL1 qPCR transcript reduction with somatic mutation burden dynamics is underway.

CONCLUSION: In ASC-treated CML, non-ABL1 somatic mutations detected by a myeloid smMIP panel are predominantly low-VAF, transient, single-timepoint calls; epigenetic genes (DNMT3A/TET2/ASXL1) form the dominant, higher-VAF, most durable clonal population. We could not observe an evident correlation of somatic mutation burden or dynamics with molecular response or clinical events, while longer follow-up and monitoring are required for firm conclusions. As in the conventional TKI era, we continued to observe a diverse spectrum of somatic mutation dynamics with ASC. Somatic mutation profiles in CML should be interpreted with caution, and longitudinal monitoring should be incorporated even in patients receiving ASC.