Beyond age: Genomic features and treatment outcomes in older versus younger adults with Acute Myeloid Leukemia

Abstract

Abstract Background: Median overall survival (mOS) in patients ≥60 years with newly diagnosed acute myeloid leukemia (AML) remains <10%, driven by adverse cytogenetics, high-risk mutations, and limited tolerance to intensive therapy. High-intensity induction chemotherapy-such as anthracycline with cytarabine (7+3), Vyxeos, or cladribine/idarubicin/cytarabine with venetoclax (CLIA-Ven)-remains the standard of care for medically fit patients. Hypomethylating agents with venetoclax (HMA+Ven) is preferred for older/unfit adults and increasingly used in younger patients ineligible for intensive therapy. However, real-world comparisons across age groups remain limited. We evaluated molecular features, treatment patterns, and outcomes in a large, single-institution cohort. Methods We performed retrospective analysis of 459 newly diagnosed AML patients (treated January 2015-February 2025), comparing those ≥60 vs. <60 years. EMR data included demographics, cytogenetics (karyotype, FISH), mutations, 30-gene NGS panel, treatment, and outcomes. Induction regimens were grouped as CLIA-Ven, 7+3 ± FLT3 inhibitors, Vyxeos, HMA-Ven, low-dose cytarabine (LDAC) ± Ven, or other. Mutations were classified by function: epigenetic (ASXL1, BCOR, EZH2, DNMT3A, TET2), RNA splicing (SF4B1, U2AF1, SRSF2, ZRSR2), transcription factor (RUNX1, GATA), signaling (CEBPA, CALR, CBL, FLT3, KRAS, NRAS, KIT, PTPN11, JAK), and IDH, STAG2, and NPM1. Statistical analysis was performed using R 4.4.0. Patients were censored at allogeneic stem cell transplant (SCT) for survival analysis. Primary endpoints were complete remission with or without count recovery (CR/CRi) and mOS, with subgroup analyses by age, cytogenetic risk, and SCT status. Results Of 459 patients, 237 (52%) were ≥60 and 222 (48%) were <60 years. Gender distribution was similar (58% vs. 54% male, p=0.4). Older patients more often had epigenetic (57% vs. 37%), TP53 (10% vs. 5%), and RNA splicing mutations (7.3% vs. 3.3%), with fewer signaling mutations (8% vs. 35%) (p<0.01 for all). Transcription factor mutations were similar (3.4% vs. 3.3%). ELN 2022 classification revealed more poor-risk disease in older patients (31% vs. 21%, p<0.01). Overall CR/CRi after first/second induction was significantly lower in older patients (60% vs. 84%, p<0.01). HMA+Ven was used more in the older cohort (54% vs. 15%, p<0.01). Relapsed/refractory disease was also higher in older patients (33.3% vs. 20%, p<0.01). SCT was performed less in older patients (13% vs. 37%, p<0.01). Among patients ≥60 years, CR/CRi rates varied by regimen: CLIA-Ven (100%, n=3), 7+3 (84%), Vyxeos (76%), HMA+Ven (62%), and LDAC (14%) (p<0.01). The mOS was 15.4 months (mo) (95% CI: 13.3–30) in patients <60 years vs. 8 mo (95% CI: 7.2–12) in those ≥60 years (p<0.01). Median relapse-free survival (mRFS) was also superior at 35 mo (95% CI: 22–NR) in older patients vs. 20 mo (95% CI: 14–32) in younger (p=0.001). Amid patients ≥60 years, the longest OS was with CLIA-Ven (mOS 14.3 mo, 95% CI: 6.3–NR), followed by HMA+Ven (11.3 mo, 95% CI: 7.7–14), Vyxeos (11.3 mo, 95% CI: 6.2–NR), 7+3 ± targeted therapy (8.7 mo, 95% CI: 6.3–22), and LDAC ± Ven (6.4 mo, 95% CI: 1.7–NR) (p <0.01). After propensity score matching (PSM) using inverse probability of treatment weighting (IPTW) to balance mutation profile and ELN risk, older patients had inferior mOS (8.8 mo, 95% CI: 7.3–NR) vs. 14.7 mo (95% CI: 11.8–25) in younger (p<0.01). The mRFS was 22 mo for younger (95% CI: 13–NR) vs. 10 mo for older patients (95% CI: 7.7–14). However, after adjusting for treatment regimen, OS and RFS differences by age were no longer significant: mOS 12.4 (95% CI: 10–19) vs. 11 mo (95% CI: 7.3–14), p=0.49; mRFS 14.6 (95% CI: 11–25) vs. 13.4 mo (95% CI: 9–25), p=0.27. Conclusion Older adults with AML often present with poor-risk genetics, receive low-intensity therapy, and have inferior outcomes. However, after adjusting for treatment intensity, age-related differences in OS and RFS diminished, suggesting outcomes are driven more by treatment selection and frailty than age alone. Subsets of older patients, especially those treated with CLIA-Ven or HMA+Ven, achieved meaningful responses, supporting individualized, risk-adapted therapy. SCT remains underutilized in older patients, where selected individuals can achieve durable remission. These findings confirm the prognostic impact of genomics and support risk-stratified therapy to improve outcomes in older adults with AML.