Venetoclax-Based Measurable Residual Disease Directed Therapy in Acute Myeloid Leukemia Is Associated with Emergent TP53 Microclones in Remission

Venetoclax (VEN) combined with low-dose cytarabine (LDAC) is highly effective in acute myeloid leukemia (AML) with either measurable residual disease (MRD) or oligoblastic (5-15% blasts) relapse (VALDAC study) (Tiong et al, JCO 2024). We have previously shown that VEN-based therapy enhances selection of TP53 deficient leukemic clones with TP53 variants observed in 32% of patients (pts) at relapse (DiNardo et al, Blood 2020 and Thijssen et al, Blood 2021). In order to further understand the origins of these clones, we aimed to sensitively assess the clonal hematopoietic landscape in remission in pts with AML treated with VEN-LDAC.

Baseline samples from 18 pts from the VALDAC study, at MRD (n=10) or oligoblastic (n=8) relapses were sequenced using highly sensitive targeted NGS (TwinStrand Duplex Sequencing^TM AML-29 Panel; sensitivity approximately 0.01% variant allele frequency [VAF]). 330 variants were detected in these 18 pts with a median of 18 variants per pt (range 4–50). Most variants were present at very low VAF (median VAF of 0.063% [IQR 0.032, 0.267]). VAF >1% was observed in 12% vs 20% of the variants among pts with MRD vs oligoblastic relapses (p=0.03), respectively. Frequently mutated genes at study entry before VEN-LDAC exposure were DNMT3A (94%), TET2 (89%), ASXL1 (72%) and TP53 (50%).

We then assessed samples from 22 pts whilst in morphological remission, tested at a median of 2.1 months post-VEN-LDAC therapy. This revealed 518 variants with a median of 19.5 variants per pt (range 4-60) and median VAF 0.057% (IQR 0.028, 0.191). Frequently mutated genes post-VEN-LDAC therapy were DNMT3A (100%), TET2 (91%), TP53 (77%), ASXL1 (64%) and CBL (50%). Comparing pre- (n=18) and post-VEN (n=22) timepoints, NPM1 mutations were most frequently cleared as expected (39% pre-VEN-LDAC vs 4.5% post-VEN-LDAC, p=0.01). The frequency for other variants were TP53 (50% vs 77%, p=0.10) and CBL (22% vs 50%, p=0.10). Strikingly, TP53 variants ≥0.1% VAF were more commonly identified post-VEN (50% of pts), compared to pre-VEN (5.6%) (p=0.004).

We were able to assess clonal dynamics in 12 pts with available paired screening and post cycle 1 (n=2) or 2 (n=10) samples. While the median number of variants per pt (18 vs 16) and VAF (0.059% vs 0.072%) were similar between the 2 timepoints, 120 (56%) out of the 216 baseline variants were not detected post-VEN-LDAC, whereas 120 new variants emerged following therapy; net change ranged from -14 to +18 variants per pt. Gene variants most frequently cleared [no. var/ no. pt] involved NPM1 (5/6 [83%]), RAD21 (5/9 [56%]) and PHF6 (6/12 [50%]). 0/2 IDH1 and 1/3 IDH2 variants became undetectable. Excluding DTA variants, the most frequently gained mutations post VEN-LDAC were TP53 (41/9 [no. var/ no. pt]), RUNX1 (8/5), CBL (5/5) and GATA2 (5/5).

We next focused on 104 persisting or emerging TP53 variants in 17/22 pts post-VEN. The median number of TP53 variants per pt was 5 (range 0-19). Median VAF was 0.041%, 27 variants were ≥0.1%, and none was above 1% (range 0.013-0.733). Notably the persistence/acquisition of the TP53 “microclones” was not significantly correlated with relapse-free or overall survival (total 9 deaths, including 2 non-relapse deaths) in 22 evaluable pts with a median follow-up of 27.3 months (range 17.6-35.5). In 6 pts that subsequently experienced morphologic relapse and assessed by standard unique molecular index (UMI)-based targeted NGS panel, only 2 TP53 variants (1 in each of 2 pts) were detected at 0.6% and 0.9% VAF.

We have previously reported the frequent emergence of BAX-mutated clonal hematopoiesis post-VEN (Tiong et al, ASH 2023). Among 20 pts with available TP53 (TwinStrand) and BAX (by UMI-NGS) mutation data, BAX mutations were detected in 7 pts (35%), all with concurrent TP53 microclones.

In conclusion, TP53 microclones (0.01-1% VAF) were detectable in early AML relapse and rapidly enriched following VEN-LDAC therapy in remission (77%). This, along with significant enrichment of BAX mutations (35% of pts) and clearance of NPM1, RAD21 and PHF6, defines a unique molecular landscape shaped by the selective pressure from VEN-based therapy. These data give insights into resistance of the hematopoietic compartment to VEN-based therapy. The expansion of specific TP53 and/or BAX-mutated clones at relapse likely involves additional aberrations and requires further study.