STAG2-Mutated MDS: Not Only Loss-of-Function?
Introduction
STAG2, a core component of the cohesin complex, is recurrently mutated in myelodysplastic neoplasms (MDS) and associated with high-risk disease and adverse prognosis in MDS.
Although incorporated into the Molecular International Prognostic Scoring System (IPSS-M) model, the clinical impact of distinct STAG2 mutation subtypes remains unclear.
Aim
This study aimed to characterize the prognostic relevance of STAG2 mutational features in MDS:
in order to improve risk stratification beyond current scoring systems.
Methods
We analyzed 4 independent datasets (Cleveland Clinic, Genome4all, MDS Unit Florence, IPSS-M database), comprising 3008 MDS patients. STAG2-mutated cases were stratified by: • Mutation typology assigned based on published functional evidence from in vitro and in vivo experimental models (Cancer Knowledgebase – Genomenon); • Variant Allele Frequency, dichotomized at the cohort median (41.2%).
Results
Cohort (Table 1) • 258/3008 patients carried STAG2 mutations. • MDS with increased blasts was the most frequent diagnosis. • IPSS-M risk distribution was largely represented by high-risk categories. Mutation burden • 96% of STAG2mut patients had at least 1 co-mutation. • Median: 5 mutations per case (range 0-13). • 46 patients (17.8%) had more than one STAG2 mutation. • 74 patients (28.7%) carried STAG2 LOF mutations and 182 patients (70.5%) UND [N.B. cases with ≥1 LOF alteration were classified as LOF despite additional UND variants]. Prognostic impact into IPSS-M category (Fig. 1a) • HR MDS (Very High, High and Moderate High categories): LOF patients exhibited significantly worse OS compared to UND (median OS: 15 vs.28 months; p <0.001). • The survival disadvantage was consistent across High (OS: p = 0.010; LFS: p = 0.024) and Very High (OS: p = 0.001; LFS: p = 0.021) subgroups. Co-mutational profile (Fig. 1b) • ASXL1 was the most frequent STAG2 co-mutated gene (64.3%). • ASXL1 was the only gene significantly enriched in LOF patients compared to UND within the Higher group (86.6% vs. 62.7%; FDR = 0.004). The co-mutation landscape did not differ significantly between STAG2_und and STAG2_LOF patients in the IPSS-M High and Very High categories. VAF did not correlate with mutation typology, nor did it independently predict survival outcomes across any endpoint.
Conclusions
These findings suggest that STAG2 mutational subtype characterization may provide higher prognostic stratification within high-risk PSS-M categories, supporting its integration into refined clinical risk algorithms for MDS.