🧬 NPM1-mutated AML: HARMONY Co-mutational Risk Model — Deep Dive ⸻ 🧠 Why this study matters •NPM1-mut AML is classically favorable, but outcomes vary widely 📉 •ELN 2022 only accounts for FLT3-ITD ➜ incomplete risk capture •HARMONY study (Leukemia 2026) builds a genotype-driven, AML-specific risk model 🧬  ⸻ 🧪 Cohorts analyzed •🟢 Training: 1001 intensively treated NPM1-mut AML (trial-based) •🔵 Internal validation: 762 real-world patients •🟣 External validation (UK-NCRI): 585 •All adult, intensive chemo, with NGS panels ⸻ 🧬 Key co-mutations explored •🔁 FLT3-ITD •🧩 DNMT3A •🧫 IDH1 / IDH2 (grouped) •🧬 TET2 ➡️ Selected based on frequency + prognostic relevance ⸻ 🧱 Foundational biology •Early clonal events: DNMT3A, TET2, IDH1/2, STAG2 •Late events: NPM1, FLT3-ITD, RAS pathway •Strong co-occurrence: •FLT3-ITD ↔ DNMT3A (≈60%) ⚠️ •“Triple-mutated” AML = NPM1 + FLT3-ITD + DNMT3A ⸻ 🔥 The HARMONY Risk Algorithm (stepwise) 1️⃣ FLT3-ITD present? ••DNMT3A-mut ➜ 🔴 Adverse •DNMT3A-wt: •IDH-mut ➜ 🟢 Favorable •IDH-wt ➜ 🟡 Intermediate 2️⃣ FLT3-ITD absent? •DNMT3A-mut: •IDH-mut ➜ 🟡 Intermediate •IDH-wt ➜ 🟢 Favorable •DNMT3A-wt: •TET2-mut ➜ 🟡 Intermediate •TET2-wt ➜ 🟢 Favorable ⸻ 🏷️ Final HARMONY Risk Groups •🟢 Favorable: ≈52% •🟡 Intermediate: ≈25% •🔴 Adverse: ≈23% ⸻ 📊 Clinical outcomes (training cohort) •🟢 Favorable: •Median OS: 14.4 years •RFS: Not reached •🟡 Intermediate: •Median OS: 2.2 years •RFS: 1.2 years •🔴 Adverse: •Median OS: 0.9 years •RFS: 0.6 years 📉 p < 0.001 across groups ⸻ 🔄 Impact vs ELN 2022 •🔁 42.7% reclassified compared with ELN 2022 •Favorable ➜ Intermediate •Intermediate ➜ Adverse (most impactful) •🎯 Better discrimination: •5-yr OS AUC: 0.695 (HARMONY) vs 0.635 (ELN) ⸻ 🧓 Works even in older adults (>60 y) •🟢 OS ~3.5 yrs •🟡 OS ~1.1 yrs •🔴 OS ~0.6 yrs ➡️ Age-agnostic genomic stratification ✅ ⸻ 🧬 Other mutations (exploratory) •❌ TP53-mut ➜ very poor OS •❌ RUNX1-mut ➜ adverse •⚖️ SRSF2-mut ➜ intermediate •🟢 STAG2 / RAD21-mut ➜ better OS (when FLT3-ITD absent) ⸻ 🩸 Allo-HSCT insights •🟢 Favorable: no clear OS benefit from HSCT in CR1 •🟡 Intermediate: moderate benefit •🔴 Adverse: highest OS benefit from allo-HSCT in CR1 🚨 ➡️ Direct transplant-decision relevance ⸻ 📌 Bottom line 🧠 NPM1-mut AML ≠ one-risk-fits-all 🧬 Co-mutation context is everything 🚀 HARMONY model: •Outperforms ELN 2022 •Reclassifies ~40–45% of patients •Sharpens HSCT decision-making •Sets the stage for genotype + MRD integrated risk 📈 ⸻ 📚 Source •Leukemia 2026 — “Unravelling co-mutational patterns with prognostic implications in NPM1-mutated adult AML (HARMONY study)”