Impact of Co-Mutations, Monocytic Differentiation, and Choice of Hypomethylating Agent Combined with Venetoclax on the Predictive Value of Molecular Prognostic Risk Signature in Newly Diagnosed AML Patients

Hypomethylating agents together with venetoclax (HMA/Ven) is an evolving standard of care for treatment of newly diagnosed patients (pts) with acute myeloid leukemia (AML). The recently proposed molecular prognostic risk signature (mPRS) aims to help better predict response to HMA/Ven. TP53^mut pts make up the Lower Benefit (LB) group, FLT3-ITD^mut or N/KRAS^mut pts, the Intermediate Benefit group (IB), and all others a Higher Benefit (HB) group. Here, we further explore how co-mutations, FAB classification, and HMA choice affect outcomes mPRS subgroups.

We conducted a retrospective review of pts >18 years old with AML treated with frontline HMA/Ven between December 2017 and March 2024. Pt records were reviewed for demographics, mutations, cytogenetics, FAB class, HMA choice, and outcomes. OS was compared using the Kaplan-Meier method.

A total of 641 pts were prescribed HMA and screened; 91 met the inclusion criteria and had follow up information available. The median age at start of treatment was 72.7 (36.1-90.0) years. 61 (67%) were male, 16 (17.6%) pts had a history of myeloid malignancy and 9 (9.9%) had therapy related AML. Adverse risk cytogenetics by ELN22 were noted in 35 (38.5%) pts.

One pt did not have mutation data available. The average number of mutations per pt was 4 (<2, n=22; 3-5, n=51; >6, n=18). 42/90 (46.7%) patients had mutations in TP53, FLT3-ITD, or N/KRAS. TP53 was mutated in 22/90 (24.4%), FLT3-ITD in 10/90 (11.1%), and N/KRAS in 10/90 (11.1%) pts. NPM1 was mutated in 16 (17.7%) pts; 9 were in IB group. IDH1 was mutated in 9 (10.0%) pts and IDH2 in 15 (16.7%); 5 and 10 pts, respectively, were in the HB cohort. 33 pts in HB group had ELN22 defined myelodysplasia related mutations (MRM).

Monocytic differentiation was identified in 38 (41.8%) pts, including 23 (25.3%) classified as FAB M4 and 15 (16.5%) as M5. Of 38 pts, 20, 10, and 8 were in mPRS HB, IB, LB groups, respectively.

60 (65.9%) pts were treated with azacitidine (Aza); 31 (34.1%) were treated with decitabine (Dec). 46 (50.5%) pts received <3 cycles of therapy. Median follow up was 22.5 mo. Median OS for the entire cohort was 7.8mo. With respect to HMA used, there was no difference in the OS in Aza and Dec groups, respectively (8.4mo vs 7.8mo; p=0.82). When risk stratified per mPRS, OS for HB group was 11.8mo (n=49), IB group 5.7mo (n=20), and LB group 4.0mo (n=22), p<0.001.

Within HB group, survival of IDH1^mut pts was surprisingly short compared to other HB mutant groups. For example, OS among IDH1^mut vs IDH2^mut pts was 3.2mo and 12.5mo, respectively (p=0.026). The survival of IDH2^mut pts was similar to that of other HB pts (12.5mo vs 12.6mo; p=0.53).

In HB pts, MRM as a whole were associated with worse, although not statistically significant, OS (10.5mo with MRM vs 30.2mo without MRM, p=0.11). However, OS in ASXL1^mut pts (n=18) was dependent on co-mutations. ASXL1^mut pts with NPM1 or IDH2 co-mutations (n=6) showed significantly improved the OS (22.5mo) compared to ASXL1^mut/NPM1^wt/IDH2^wt (n=12) (2.9mo; p=0.035). A similar trend was observed with RUNX1 and IDH2 co-mutations, with OS of 22.5mo vs 10.1mo (p=0.35) for RUNX1/IDH2^mut (n=4) and RUNX1^mut/IDH2^wt (n=11) respectively.

Within IB group, the OS differed significantly for NPM1^mut vs NPM1^wt pts, 15.3mo vs 3.9mo; p=0.024, respectively. NPM1 mutation showed a greater impact on OS in N/KRAS^mut pts (OS unreached, n=3 for NPM1^mut vs 1.4mo, n=7 for NPM1^wt, p=0.015) than in FLT3-ITD^mut pts (OS 5.4mo, n=6 for NPM1^mut vs 5.9mo, n=3 for NPM1^wt; p=0.7).

With respect to monocytic differentiation, there was a numerical, but not statistically significant difference in OS in FAB M4, M5, and non-M4/M5 cohorts (3.6mo, 7.8mo, and 9.4mo, respectively; p=0.31).

In this single institution retrospective study with limited sample size we found that mPRS classifies pts well into 3 distinct groups with significantly different outcomes to therapy with HMA/Ven. Furthermore, our findings suggest NPM1 and IDH2 mutations may favorably abrogate molecular risk within mPRS which impacts prognosis, while, surprisingly, IDH1 mutations were associated with worse outcomes. We show that choice of HMA does not impact outcomes. Monocytic differentiation trends toward poorer outcomes with the use of HMA/Ven, consistent with previous reports. Context matters, it is critical to discern idiosyncrasies of those patients who do not respond to the yet evolving standard of care treatments as these are imperative to guiding future advancements.