Mutant RUNX1, SETBP1 and IDH Mutant Allele Dynamics As Biomarkers of Hematologic Response to IDH Inhibitors in Myeloid Malignancies

Background: Mutations in isocitrate dehydrogenase 1 or 2 (IDH1 or IDH2) are observed in ~20% of patients with acute myeloid leukemia (AML) and 6% of patients with myelodysplastic syndrome (MDS), carrying an intermediate prognosis despite the development of novel IDH1 and IDH2 inhibitors. Clinical trials have demonstrated clinical responses for patients carrying these mutations in relapsed or refractory AML and IDH1 inhibitors are now approved in 1^st line. In patients with relapsed disease, MDS, or chronic myelomonocytic leukemia (CMML), these agents are often used off-label to achieve hematological improvement and freedom from red blood cell or platelet transfusions. However, the genomic landscape of the response to therapy is not well characterized. This study aimed to evaluate the mutational profile of patients with CMML, MDS or AML receiving IDH1 or IDH2-targeted therapy at the time of initiation, response, and relapse.

Methods: We identified all patients with a hematological malignancy treated from 2010 through April 2024 at Columbia University Medical Center with documented IDH1 or IDH2 mutations who had received either ivosidenib (IDH1 inhibitor) or enasidenib (IDH2 inhibitor). Laboratory values, bone marrow biopsy reports and mutational profiles were collected prior to initiation of therapy, during the period of best response, and at time of discontinuation. The objective of the study was to determine gene mutations and variant allele frequencies (VAF) which were associated with hematologic response to therapy, defined as an absolute neutrophil count (ANC) ≥1.0 x 10⁹/L, platelet count ≥100 x 10⁹/L, and red blood cell transfusion independence, defined as the absence of any transfusions for ≥12 weeks within the first 24 weeks. Patients achieving all 3 hematologic parameters were considered to have a complete hematologic response (CHR); those meeting 0, 1, or 2 parameters were defined as having an incomplete hematologic response (iHR). VAF changes were assessed by two-tailed t-test; differences in mutation burden were analyzed by Fisher’s exact test.

Results: We identified 59 IDH-mutant patients treated with an IDH inhibitor, 51% of whom were male. Median age was 76 years (range 43-92). 49 (83%) patients had AML, 7 (12%) with MDS, 1 (2%) with CMML, and 2 (3%) with systemic mastocytosis with an associated hematologic neoplasm (SM-AHN). Of these 59 patients, 25 (42%) carried an IDH1 mutation and 33 (56%) had an IDH2 mutation; 3 patients had co-occurring IDH1 and IDH2 mutations. 27 (46%) were treated with ivosidenib, and 32 (54%) with enasidenib. There was no significant difference in age, initial WBC, ANC, Hgb, or Plt count between IDH-inhibitor treated groups. 33 patients (56%) received concomitant therapy, most commonly a hypomethylating agent. Median duration of treatment was 153 days (range 13–2023).

In total, 23 of 59 patients achieved a CHR while on therapy (39.0%); there was no significant difference in rates of CHR by IDH inhibitor (p>0.05). Ivosidenib-treated patients with CHR had a significantly greater reduction in mutant IDH VAF as compared to those with iHR (mean difference: -11.57% vs + 2.98%; p=0.021) which was not observed in enasidenib-treated patients (+4.13% vs. +5.10%; p>0.05); Mutant IDH VAF at initiation was not significantly different between CHR vs. iHR with either IDH-inhibitor group (p>0.05).

The most common co-mutations in the CHR group were SRSF2 (42.9% of tested patients) and DNMT3A (28.6%). The most common co-mutations in the iHR group were DNMT3A (47.8%), SRSF2 (36.0%), RUNX1 (34.6%), SETBP1 (30.0%) and BCOR (25%). Notably, no patients with a CHR were found to have a RUNX1, SETBP1 or BCOR mutation (0/16, 0/12, and 0/16 respectively). Patients with a RUNX1 or SETBP1 mutation were significantly associated with an iHR (p=0.0078 and 0.0396, respectively); a BCOR mutation trended towards significance with an iHR (p=0.0793).

Conclusion: Many patients achieve meaningful hematologic responses to IDH inhibitors. Reductions in mutant IDH1 VAF, but not IDH2, may be associated with a CHR; however initial VAF is not predictive of response to therapy. Patients with co-mutations in RUNX1, SETBP1 or BCOR may be associated with iHR and may be less likely to benefit from therapy. The impact of the co-mutational landscape on clonal contraction in response to inhibitor therapy as a driver of hematological response will be investigated.

BC and AS are co-first authors with equal contribution