Evolution informed scoring improves prognosis in MDS
MYELODYSPLASTIC syndromes may be more accurately stratified using a molecular evolution framework that integrates genomic trajectories with clinical data, improving prediction of both leukaemia free survival and overall survival.
Myelodysplastic syndromes are clonal haematopoietic stem cell disorders characterised by multistep molecular evolution and highly variable risks of progression to acute leukaemia. This heterogeneity creates challenges for clinicians attempting to estimate prognosis and guide treatment decisions, highlighting the need for more precise risk assessment tools.
Researchers developed ProgEvo, a framework designed to infer molecular evolutionary trajectories and combine these findings with established clinical variables to improve prognostic accuracy.
Building An Evolution Informed Model
ProgEvo was trained using data from 2,519 patients in cBioPortal and validated in two independent cohorts comprising 2,043 patients from Genomed4All and 2,157 patients from the Moffitt Cancer Center cohort.
The framework identified 1,765 gene co-occurrences that were grouped into 45 directional evolutionary routes. Eighteen of these routes were subsequently validated in the external Genomed4All cohort.
Researchers then applied a multivariable feature selection strategy to incorporate evolution consistent variables into the existing International Prognostic Scoring System Molecular model. Five evolution informed variables were selected for inclusion, including two directional routes involving ASXL1 followed by KRAS and SRSF2 followed by NRAS, one co-occurrence involving NRAS and RUNX1, and early mutations affecting ATRX and JAK2.
Improved Classification Of Patient Risk
The resulting model, termed IPSS M Evo, demonstrated improved discrimination for both leukaemia free survival and overall survival compared with the existing scoring framework.
Importantly, more than 40% of patients in the Genomed4All cohort were reassigned to different risk categories using the evolution informed approach. The improved performance of the model was subsequently confirmed in the independent validation cohort from the Moffitt Cancer Center.
Researchers concluded that ProgEvo supports the integration of molecular evolution into clinical prognostic frameworks and may enable more personalised risk assessment in myelodysplastic syndromes.
Clinicians can access web based tools to calculate the IPSS M Evo score and compare mutational profiles with cohort derived evolutionary trajectories through Evolution Graphs Tool and IPSS-M-Evo Calculator.
Reference
Civettini I et al. Prognostic score for myelodysplastic syndromes based on molecular evolution. NEJM Evid. 2026;5(6):DOI: 10.1056/EVIDoa2500124.
