MPN-288: Molecular Characteristics of IDH2-Mutated, Philadelphia Chromosome-Negative, Advanced-Phase Myeloproliferative Neoplasms Treated with Enasidenib

Context: Approximately 20% of acclerated-phase/blast-phase (AP/BP), Philadelphia chromosome-negative (Ph-negative) myeloproliferative neoplasms (MPNs) have an IDH1 or IDH2 mutation. We previously reported on encouraging clinical outcomes in IDH2-mutated MPN-AP/BP patients treated with enasidenib (an IDH2 inhibitor) (Patel et al, BJH 2020). Mechanisms of resistance to IDH inhibition in AML have been reported but without a specific focus on the MPN-AP/BP population. Objective: Analyze molecular characteristics of IDH2-mutated, Ph-negative MPN-AP/BP patients treated with enasidenib at diagnosis, best response, and relapse. Design/Setting: Retrospective single-center analysis of IDH1/2-mutated, Ph-negative, MPN-AP/BP patients that received treatment with an IDH inhibitor from 2009-2019. Patients: IDH1/2-mutated patients were identified utilizing our institutional next-generation sequencing (NGS) database. 8 MPN-AP/BP patients (7 MPN-BP, 1 MPN-AP) that received an IDH inhibitor were identified; all had an IDH2 mutation and received enasidenib. Main Outcomes: Mutant variant allele frequencies (VAFs) and co-occurring pathogenic mutations in addition to IDH2 and canonical MPN mutations (co-mutations). Results: Median overall survival was 338 days (range 26–1341) and overall response rate using 2012 MPN-BP criteria was 75%. All eight patients had a canonical MPN mutation (7 JAK2, 1 MPL). Patients had a median of two co-mutations (range 1–4). NGS data were available for four patients at best response; IDH2 became undetectable in two patients [one with a partial acute leukemia response (ALR-P), and one with a complete acute leukemia response (ALR-C)], while two had persistence (1 ALR-P, 1 ALR-C). The median number of co-mutations at best response was 1 (range 0–4). Three patients had NGS at time of relapse. One patient's IDH2 mutation became detectable again, one had persistence of the mutation, and one's IDH2 mutation remained undetectable. All three patients acquired additional mutations at time of relapse, with a median of four co-mutations (range 4–5). Of note, canonical MPN mutations persisted at best response and at relapse for each patient. Conclusions: Canonical MPN mutations persist with IDH inhibition, even if the IDH mutation is undetectable, which may lead to acquisition of additional pathogenic mutations at relapse. Combination therapy approaches capable of inducing deep molecular responses of the IDH-mutated clone and MPN clone should be prospectively evaluated.