AUTHORS
Frederico Honorato Honorato Jr, MD, Carolina Perrone Marques Ferreira, MD2, Aline Raposo Nishimoto, MD3, André Dias Américo Dias Dias Américo, MD4, Alexandre Eiji Kayano, MD5, Amanda Ennes Inacio Dias dos Santos, MD6, Tarcila Santos Datoguia, MD7, Jade Zezzi Martins do Nascimento, MD8, Kira Bucci, MD9, Juliana Dall Agnol Rocha, MD10, Luana Nobrega, MD11, Paloma Martinho Resende, MD12, Nelson Hamerschlak Sr, MD PhD13, Fabio Pires S Santos, MD, PhD14
AFFILIATIONS
1 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
2 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
3 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
4 Beneficiência Portuguesa Hospital, São Paulo, São Paulo, Brazil
5 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
6 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
7 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
8 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
9 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
10 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
11 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
12 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
13 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
14 Einstein Hospital Israelita, São Paulo, São Paulo, Brazil
ABSTRACT
Bernard et al. identified 18 molecular groups with distinct outcomes in myelodysplastic syndromes (MDS) (Blood 2024;144:1617). Whether this taxonomy extends to acute myeloid leukemia with myelodysplasia-related features (AML-MR) and resolves prognostic heterogeneity within this predominantly adverse-risk population remains unknown.
We pooled publicly available data from four adult AML cohorts (UK-NCRI, AMLSG, BeatAML, and TCGA). AML-MR was defined by International Consensus Classification (ICC) genomic criteria; TP53-altered cases were retained in this subgroup to preserve the Bernard taxonomy and permit assignment of the TP53-complex group. World Health Organization (WHO) criteria were applied in a sensitivity analysis. Patients were assigned by a panel-adapted Bernard classifier, with documented approximations at nodes requiring variant-level allelic data. Seventeen groups were represented because the no-event category was not applicable in AML-MR. Because ELN 2022 assigns nearly the entire cohort to a single risk category, the contribution of molecular group to OS prediction was evaluated within the adverse-risk stratum rather than by between-category comparison, using a Cox model adjusted for age and blast percentage (restricted cubic splines) and for cohort. A secondary analysis evaluated whether molecular subgroups were associated with OS after adjusting for ELN 2017 and age. Discrimination was assessed by the Harrell C-index with optimism correction from 200 bootstrap resamples.
We included 1,441 patients with AML-MR; median age was 61 years (interquartile range, 53-70), and 1,166 deaths occurred among 1,440 evaluable patients. Overall, 90.8% were assigned to one of 16 driver-defined molecular groups, whereas 9.2% were molecularly not otherwise specified. Compared to the published data on MDS, there was substantial difference in molecular subgroup distribution: an increase was observed in TP53-complex (10% to 28.6%), IDH-STAG2 (8.9% to 16.2%) and -7/SETBP1 (4.9% to 10.9%), while some subgroups were more uncommon in AML-MR, including SF3B1 (14% to 4.1%), bi-TET2 (13% to 5.3%) and del(5q) (6.9% to 2%). ELN 2022 classified 1,425 patients (98.9%) as adverse risk. Among 1,404 evaluable patients within that stratum, molecular subgroups remained independently associated with overall survival (OS) after adjustment for age, blast percentage, and study (global Wald χ2=163, 16 df, P<.0001). Adding molecular group increased the optimism-corrected C-index from 0.633 to 0.671 (ΔC +0.038). TP53-complex subgroup had an almost threefold higher adjusted hazard than SF3B1 (hazard ratio [HR], 2.92; 95% CI, 2.09-4.07; P<.0001), with a median OS of 6.2 versus 18.6 months. In a secondary analysis, adding molecular group to ELN 2017 plus age increased the optimism-corrected C-index from 0.642 to 0.670 (ΔC +0.028; likelihood ratio χ2=136, 16 df, P<.0001). Findings were concordant under the WHO definition of AML-MR (n=1,328 evaluable; global Wald c2=152, 16 df, P<.0001). Limitations of this study include the retrospective nature of the analysis, heterogeneity of next-generation sequencing panels used and inference of multi-hit status from variant sequencing reads.
An MDS-derived molecular taxonomy is applicable to AML-MR and reveals substantial prognostic heterogeneity within a population predominantly classified as ELN adverse risk. Molecular subgrouping may complement current risk classification by resolving clinically relevant heterogeneity within the ELN adverse-risk category.
DISCLOSURES
Fabio Pires S Santos, MD, PhD: GSK - Honoraria; BeOne - Honoraria; Dr Reddy's - Honoraria; AstraZeneca - Honoraria; Bristol-Myers-Squibb - Honoraria; Novartis - Honoraria; Abbvie - Honoraria; Amgen - Honoraria; Servier - Honoraria; Pint Pharma - Honoraria; Johnson & Johnson - Honoraria; Abbvie - Research Funding; AstraZeneca - Research Funding; Johnson & Johnson - Research Funding; Servier - Research Funding