Less‑intensive therapy in newly diagnosed acute myeloid leuekmia (AML) in India: Real-world adoption and prognostic utility of ELN‑LI and ELN‑2022 classification

Abstract

Abstract Treatment of AML in India remains challenging. Delivery of intensive chemotherapy (IC) is constrained by costs and health system factors (Philip C et al.BJH 2015). Many newly diagnosed (ND) patients cannot access standard IC, and those who do often face high early mortality from multidrug-resistant infections (Jain H et al.JGO 2025) Consequently, less intensive (LI) hypomethylating agent (HMA) regimens, with or without venetoclax (Ven),are adopted to bridge select patients to IC,or when IC is unaffordable or impractical. Although approved primarily for older patients, LI is being increasingly used even among younger adults who cannot receive IC. However, real-world data on LI adoption and outcomes are limited. Identifying with this goal, we used the Indian Acute Leukemia Research Database (INwARD; established in 2018 by the Hematology Cancer Consortium) to provide a descriptive analysis of the profile of AML treatment post-introduction of LI in our population. Registry data (1st January 2020 through 31st January 2025) collected from 21 member institutions via a prospectively maintained central online data management system were analyzed for treatment and survival outcomes. Survival and follow-up were updated as of 5th of July 2025. A total of 4,845 adults (≥18 years) with ND AML were registered; 3,127 (64.5%) received treatment and 1,718 (35.5%) did not. Among the treated, 1,728 (55.3%) received IC and 1,399 (44.7%) patients received only LI [864 (61.8%) were <60 and 535 (38.3%) ≥60 years]. Azacitidine monotherapy in 657 (46.9%) was the commonest LI. On comparison of IC vs LI, gender (male [923 (53.51) vs 792 (56.73)]), hemoglobin [78.8 (22.1) vs 76.8 (20) g/L], platelet count [44 (21, 91) vs 43 (21, 90) ×10^9/L], and the prevalence of FLT3-ITD and NPM1 mutations were comparable (FLT3-ITD 18.8% vs 18.3%, p=0.789; NPM1 29.0% vs 25.3%, p=0.056). However, patients receiving IC were younger (mean 40.2 vs 53.5 years), had higher presenting white blood cell count (median 15.2 vs 9.9 ×10^9/L), and were more likely to have favourable-risk and less likely to have adverse-risk ELN categories vs those receiving LI (favourable 38.9% vs 18.9%; adverse 14.8% vs 28.8%; p<0.001). OS and EFS were 42 months (95% CI 35.6–48.4) and 15 months (95% CI 12.6–17.4), respectively (p<0.001). In those who received LI, 494 (35.3%) patients had sufficient data to assign risk stratification by ELN 2022 and ELN LI criteria. The median age was 55 years (range, 18–88), 288 patients (58.3%) were male; 325 (65.8%) were <60 years and 169 (34.2%) ≥60 years. Stratified by ELN 2022, 92 patients (18.6%) were favorable risk, 208 (42.1%) intermediate, and 194 (39.2%) adverse risk. The most frequent mutations noted were NPM1 (33.1%), FLT3 ITD (22.4%), IDH2 (18.6%), DNMT3A (16.0%), RUNX1 (14.5%), TET2 (12.2%), NRAS (11.5%), ASXL1 (10.7%), IDH1 (10.7%), and TP53 (10.4%). By the ELN LI model, 279 patients (56.48%) were favorable, 174 (35.22%) intermediate, and 41 (8.30%) adverse risk. Median follow-up was 4.92 months, and 94 patients in this cohort were followed up for >1 year. There were 140 deaths. The median (95% CI) EFS and OS were 18 months (14.166–21.834) and 20 months (14.537–25.463), respectively. The 2022 ELN risk stratification did not differentiate OS across favorable, intermediate, and adverse groups (log-rank p=0.406); median (95% CI) OS was 22 (14.220–29.780) vs 20 (14.419–25.581) vs 20 (14.537–25.463) months, respectively. In contrast, the ELN LI model separated the adverse group (log-rank p=0.002): median OS was 20 (12.814–27.186) vs 22 (15.535–28.465) vs 8 (2.640–13.360) months for favorable, intermediate, and adverse groups, respectively, though it did not discriminate between favorable and intermediate. An exploratory sequential BATTing subgroup–identification approach highlighted TP53 mutated disease as the poorest surviving subgroup in our cohort. In conclusion, treatment rates among adults with ND AML in India is rising with nearly half receiving LI therapy. The genomic landscape of this LI-treated cohort mirrors global reports. ELN-2022 did not stratify overall survival, whereas ELN-LI identified a distinctly adverse group. Limitations include the short follow-up, lack of a standard protocol for LI/HMA-ven, and the inherent constraints of registry analyses. Even so, our real-world analysis shows more patients are now receiving treatment following the introduction of LI in our population. Our data support using ELN-LI for risk communication.