Distinct patterns of mutant ASXL1 over time and their implications for treatment failure (TF) and BCR::ABL1 mutation development in newly diagnosed patients with chronic myeloid leukemia in chronic phase (CML-CP) treated with asciminib (ASC) vs investigator-selected tyrosine kinase inhibitors (IS-TKI) in the ASC4FIRST study

Abstract Introduction: ASC, an allosteric BCR::ABL1 inhibitor, was recently approved in several countries for use in newly diagnosed patients (pts) with CML-CP based on the results from the ASC4FIRST study (NCT04971226). The most frequent additional genomic alteration (AGA) at baseline (BL) reported in this study was ASXL1 mutation (ASXL1+). The allelic burden of ASXL1+ clones at BL was incidentally higher in pts receiving ASC (24% variant allele frequency [VAF]) vs IS-TKI (10% VAF). ASXL1+ at BL was associated with increased risk of TF (as per ELN 2020 criteria). Here, we report an exploratory analysis of the longitudinal changes of ASXL1+ from BL in relation to outcomes until week (wk) 96 follow-up. Methods: Presence of AGAs was determined ​​using a DNA next-generation sequencing panel of 89 genes on diagnostic blood samples collected at BL, wk 48 and end of treatment (EOT). A multivariable Cox proportional hazards regression model with time-dependent covariates adjusted for age, sex and treatment (ASC or IS-TKI) was used to determine the prognostic and predictive value of AGAs considering their changes over time. Pts harboring AGAs were categorized as presenting persistent, eradicated or treatment-emergent clones. Results: Of 405 pts randomized, AGA analysis was performed at BL for 341 pts (ASC, n=169; IS-TKIs, n=172 [imatinib, n=83; second generation TKI, n=89]) with the available samples. AGAs were detected at BL in 21.1% (72/341) of pts (32/169 ASC vs 40/172 IS-TKI). The most frequent AGA detected was ASXL1+ in 11% (39/341) of pts at BL and in 15% (50/341) at any time until wk 96/EOT. Three patients with ASXL1+ at BL had no post-BL genomic data. Other AGAs detected were not included in the analysis due to low frequencies of events for individual mutated genes. On-treatment emergence of at least one new ASXL1+ clone was observed in 12 pts (6 ASC vs 6 IS-TKI). In 6 pts with an emergent ASXL1+ clone at wk 48 (VAF: 1%−10%), major molecular response (MMR) was achieved at the same time. In these 6 cases, clonal hematopoiesis (CH) in non-leukemic clones was suspected. ​For the other 6 pts, ASXL1+ emergence likely took place in leukemic clones. Eradication of BL ASXL1+ was observed in 74% (29/39) of pts irrespective of initial clonal burden (VAF: 1.2%–44%). ASXL1+ clones detectable at BL were persistent at the last evaluation in 6 pts (4 ASC vs 2 IS-TKI). Pts with ≥1 persistent ASXL1+ clone had a higher probability of TF as compared to pts who had on-treatment eradication of BL ASXL1+ clones in either treatment arm. Summarized at pt level, TF was observed in 5/6 pts with persistent vs 4/29 pts with eradicated clones (Fisher test; p=0.03). Treatment-emergent BCR::ABL1 mutations were associated with persistent ASLX1+ in 5/6 cases vs 1/29 cases among pts who had on-treatment eradication of ASXL1+ clones. Overall, detectable ASXL1+ at any time during the study was a prognostic and a predictive factor for TF. For pts with no ASXL1 mutation at any time point (85%; 291/341 pts), the TF rate was considerably lower for ASC (8%; 12/145 pts) than for IS-TKI (28%; 41/146 pts). The overall TF rate was comparable between treatment arms (8/24 on ASC vs 9/26 on IS-TKI) among patients with detectable ASXL1+ at any time. Although the ASC arm showed a higher TF risk difference in ASXL1+ compared to ASXL1-negative pts, this may be due to the higher baseline ASXL1+ clonal burden in the ASC group. In the larger subset of ASXL1-negative pts, the MMR rate at wk 96 was 80% (116/145) for ASC and 51% (74/146) for IS-TKI, while the MMR rate was comparable for pts on ASC (13/24; 54%) vs IS-TKI (12/26; 46%) in the subset with detectable ASXL1+ at any time during the study. Conclusions: ASXL1+ was the most frequent AGA in ASC4FIRST and was associated with higher rates of TF and BCR::ABL1 mutations. The number of emergent ASXL1 mutations on treatment was low and similar in both treatment arms. Most of the ASXL1+ clones were eradicated on treatment and the initial allelic burden at BL was not associated with persistence or eradication. TF risk was lower for pts with eradicated ASXL1+ clones. Data from this exploratory analysis in ASC4FIRST support the use of ASC in newly diagnosed CML-CP pts regardless of ASXL1 mutational status. Additional investigations in larger cohorts and longer follow-up are warranted to further elucidate the risks associated with different patterns of ASXL1+ over time and to develop optimized therapeutic strategies in this population.