AUTHORS
Allyson Maria Freedy, MD, PhD, Danica Wiredja, MD, PhD2, Curtis Lachowiez, MD3, Ann-Kathrin Eisfeld, MD4, Deedra Nicolet, MS5, Jad Othman, MD6, Mael Heiblig, MD7, Anand Ashwin Patel, MD8, Jingjing Zhang, MD, PhD9, Violet Bain10, Grace Bosma, MS11, Diana Abbott, PhD12, Christine M. McMahon, MD13, Mathew G. Angelos, MD, PhD14, Jonathan Gutman, MD15, Marc Schwartz16, Maria Laura Amaya, MD PhD17, Daniel Pollyea, MD18, Andrew T Kent, MD, PhD19
AFFILIATIONS
1 University of Colorado, Aurora, Colorado, United States
2 Children's Hospital of Colorado, Aurora, Colorado, United States
3 Oregon Health & Science University, Portland, Oregon, United States
4 The Ohio State University, Columbus, Ohio, United States
5 The Ohio State University, Columbus, Ohio, United States
6 King's College London, London, United Kingdom
7 Lyon-Sud Hospital, Pierre Bénite, France
8 University of Chicago, Chicago, Illinois, United States
9 University of Colorado, Aurora, Colorado, United States
10 University of Colorado, Aurora, Colorado, United States
11 University of Colorado, Aurora, Colorado, United States
12 University of Colorado, Aurora, Colorado, United States
13 University of Colorado, Aurora, Colorado, United States
14 University of Colorado, Aurora, Colorado, United States
15 University of Colorado, Aurora, Colorado, United States
16 University of Colorado, Aurora, Colorado, United States
17 University of Colorado, Aurora, Colorado, United States
18 University of Colorado, Aurora, Colorado, United States
19 University of Colorado, Aurora, Colorado, United States
ABSTRACT
Background:
Mutations in isocitrate dehydrogenase 2 (IDH2) are observed in ~10% of patients with newly diagnosed acute myeloid leukemia (ND-AML). IDH2 mutations occur at two hotspot residues: arginine 140, most commonly mutated to glutamine (R140Q, 80% of IDH2-mutated cases), and arginine 172, most commonly mutated to lysine (R172K, 20% of IDH2-mutated cases). The two IDH2 variants are increasingly recognized as biologically distinct disease subtypes with distinct co-mutation profiles (Papaemmanuil et al NEJM 2016).
The prognostic impact of specific IDH2 variants in AML remains controversial. One retrospective study demonstrated inferior overall survival (OS) in patients with R172K vs. R140Q variants when treated with intensive chemotherapy (IC, Duchmann et al Blood 2021). However, subsequent studies have failed to reproduce this OS difference (Salman et al Blood Advances 2025). Whether AML with distinct IDH2 variants exhibits differential response and survival to venetoclax combined with azacitidine (ven/aza), remains unknown.
Methods:
We performed a retrospective cohort study of 121 patients with ND-AML with mutations in IDH2 (R140Q n=87, R172K n=34) treated with frontline ven/aza across six institutions. A cohort of 386 patients without IDH2 mutations treated with ven/aza and 48 patients with IDH2 mutations treated with IC were used as comparators. Categorical demographic variables were compared using Fisher’s exact test, Pearson’s Chi-squared test and Wilcoxon rank sum tests. Kaplan-Meier curves were compared using log-rank test. A multivariate Cox regression model was applied to survival data.
Results:
There was no significant difference in median age at diagnosis between patients with R140Q vs. R172K variants (73 vs. 72 years, p-value 0.12). Similar to prior reports, patients with IDH2 R140Q variants showed a greater proportion of co-occurring NPM1, FLT3-ITD and SRSF2 mutations and those with R172K variants showed a greater proportion of co-occurring DNMT3A and BCOR mutations (Middeke et al Blood 2022). Patients with IDH2 R172K vs. R140Q were more often classified as favorable risk as determined by 4-gene classifier (82% vs. 62%, p-value 0.002). This difference was attributed to the higher proportion of co-occurring FLT3-ITD mutations in the R140Q cohort, rendering R140Q patients more frequently intermediate risk. There was no significant difference in allogeneic cell transplantation rates between groups (R140Q 20% vs. R172K 29%, p-value 0.2).
Patients with R172K vs. R140Q treated with ven/aza had increased overall response rates (CR+CRi+CRh+MLFS+PR) (97% vs. 79%, p-value 0.01). All evaluable patients with R172K variants responded following ven/aza therapy; 6 patients with IDH2 R140Q variants were refractory.
This difference in response rates correlated with improved OS in patients with R172K vs. R140Q variants treated with ven/aza (median OS not reached vs. 458 days, p-value < 0.0001). The OS of patients with IDH2 R140Q was not statistically different than patients without IDH2 mutations treated with ven/aza (median OS 458 days vs. 301 days, p-value 0.1). Consistent with prior reports, there was no OS difference between patients with IDH2 R140Q vs. R172K variants when treated with IC (24-month OS 75% vs 70%, log-rank p-value 0.9).
OS remained superior in patients with R172K vs. R140Q variants even within the subgroup of patients classified as favorable risk (12-month OS: 96% vs. 59%, log-rank p-value < 0.0001). Cox multivariate regression accounting for IDH2 variant, age, gender, and genetic risk group demonstrated a statistically significant reduced risk of death in patients with IDH2 R172K vs. R140Q variants (HR:0.15, 95% CI 0.05-0.42, p-value < 0.001).
Conclusion:
When treated with frontline ven/aza, AML patients with IDH2 R172K variants demonstrated increased response rates and improved survival compared to IDH2 R140Q variants. Multivariate analysis showed this was independent of co-occurrence of intermediate risk-defining mutations (particularly FLT3-ITD). The OS differences were not observed with IC-treatment, highlighting that treatment selection and prognostic risk modeling may differ based upon variant level information. Correlative studies are in progress to understand the mechanism behind our observation.
DISCLOSURES
Allyson Maria Freedy, MD, PhD: No relevant conflicts of interest to declare.