Background
Phase 3, open-label, randomized controlled trial (IDHENTIFY). N=319 patients with R/R IDH2-mutated AML who had received 2–3 prior regimens, including intensive chemotherapy. Enasidenib (AG-221) is a selective inhibitor of mutant IDH2 (R140Q, R172K, and other variants), reducing 2-hydroxyglutarate and restoring epigenetic homeostasis. Previously approved based on phase 1/2 single-arm data (Stein Blood 2017; ORR 40.3%), IDHENTIFY was designed to confirm OS superiority vs conventional care regimens (CCR) in R/R IDH2-mutated AML. Published Leukemia 2023.
Interventions and follow up
Arm A: Enasidenib 100 mg PO once daily until progression or unacceptable toxicity
Arm B: Conventional care regimen (CCR) — investigator's choice: azacitidine, low-dose cytarabine, intermediate-dose cytarabine, or best supportive care (BSC)
Primary endpoint: OS
mFollow up: 23.8 month
Arm B: Conventional care regimen (CCR) — investigator's choice: azacitidine, low-dose cytarabine, intermediate-dose cytarabine, or best supportive care (BSC)
Primary endpoint: OS
mFollow up: 23.8 month
Results
mOS (enasidenib vs CCR): 8.8 vs 9.6 months, HR 1.06 (95% CI 0.83–1.36), P=.79 — NOT SIGNIFICANT
CR rate: 27.1% (enasidenib) vs 10.6% (CCR), P<.001
ORR: 35.5% vs 19.5%, P=.001
EFS: 4.9 vs 2.6 months, HR 0.87 — not significant
Differentiation syndrome: 13.8% (enasidenib) vs 0.6% (CCR)
CR rate: 27.1% (enasidenib) vs 10.6% (CCR), P<.001
ORR: 35.5% vs 19.5%, P=.001
EFS: 4.9 vs 2.6 months, HR 0.87 — not significant
Differentiation syndrome: 13.8% (enasidenib) vs 0.6% (CCR)
Adverse events
Overall: Grade ≥3 AEs 73.1% (enasidenib) vs 75.5% (CCR); discontinuation due to AEs 10.6% vs 5.0%.
IDH-specific: Differentiation syndrome any grade 13.8% vs 0.6% — key safety monitoring requirement.
Hepatic/metabolic: Hyperbilirubinemia 33% vs 10% — enasidenib class effect (indirect/unconjugated via bilirubin conjugation inhibition, not hepatotoxic).
Gastrointestinal: Nausea 55% vs 35% (grade ≥3 2%); diarrhea 32% vs 24%.
Hematologic: Cytopenias balanced across arms.
IDH-specific: Differentiation syndrome any grade 13.8% vs 0.6% — key safety monitoring requirement.
Hepatic/metabolic: Hyperbilirubinemia 33% vs 10% — enasidenib class effect (indirect/unconjugated via bilirubin conjugation inhibition, not hepatotoxic).
Gastrointestinal: Nausea 55% vs 35% (grade ≥3 2%); diarrhea 32% vs 24%.
Hematologic: Cytopenias balanced across arms.
Conclusions
IDHENTIFY did not demonstrate an OS advantage for enasidenib over CCR in R/R IDH2-mutated AML (mOS 8.8 vs 9.6 months, HR 1.06), despite statistically higher CR (27% vs 11%) and ORR (36% vs 20%) rates. The disconnect between response rates and OS — likely reflecting the heterogeneity of the CCR arm (some patients received active chemotherapy) — means enasidenib's clinical role rests on its phase 1/2 approval data, not IDHENTIFY.
Key Limitations
The CCR comparator arm was heterogeneous (aza, LDAC, intermediate-dose Ara-C, or BSC at investigator's choice), allowing some patients to receive active therapies — potentially explaining why the CCR arm performed unexpectedly well (mOS 9.6 months). Cross-over was allowed, diluting OS differences. Phase 3 negative OS results raise questions about the clinical benefit of enasidenib as monotherapy in R/R AML; optimal sequencing, combination approaches, and patient selection criteria need refinement. Hyperbilirubinemia (enasidenib class effect via indirect pathway inhibition — NOT hepatic injury) must be recognized and distinguished from true liver toxicity. Differentiation syndrome education is mandatory.
Clinical Context
Despite IDHENTIFY's negative primary OS endpoint, enasidenib remains FDA-approved for R/R IDH2-mutated AML (accelerated approval August 2017, based on phase 1/2 ORR; full approval sought). Enasidenib occupies a niche in R/R IDH2-mutated AML as a well-tolerated oral option for patients declining or ineligible for intensive salvage chemotherapy. The FDA has not withdrawn approval based on IDHENTIFY. No frontline IDH2 inhibitor combination trial (analogous to AGILE for IDH1) has yet been positive. In IDH2-mutated AML, allo-SCT in CR1 is recommended for eligible patients. Mutations at R172K show different response profiles vs R140Q. ESMO-MCBS: not assigned.