Immune profiling identifies distinct immunological subtypes of Myelodysplastic Syndromes with prognostic and therapeutic relevance

Myelodysplastic syndromes (MDS) are heterogeneous myeloid neoplasms with variable risk of progression to acute myeloid leukemia (AML), making prognostication and prediction of response to frontline hypomethylating agents (HMAs) challenging. Current risk stratification models rely mainly on non-specific clinical parameters and insufficiently capture underlying disease biology. Given the growing evidence for immune dysregulation in MDS pathogenesis and progression, we performed comprehensive immune profiling of bone marrow and peripheral blood from 286 MDS and post‑MDS AML patients, both before and after HMA therapy. High-dimensional flow cytometry was used to analyze NK, T, and myeloid cell populations, and data were analyzed using an unsupervised pipeline integrating Phenograph and HDBSCAN algorithms. Five immunological patient groups were identified, reflecting varying degrees of immune dysfunction. Notably, the immune‑dysfunctional groups displayed increased frequencies of terminally differentiated NK cells with high PD‑1 expression, consistent with innate immune exhaustion. These immune groups showed significantly different clinical outcomes and responses to HMA treatment. Importantly, patients with the same conventional prognostic classifications were distributed across multiple immune groups, highlighting limitations of existing scoring systems. To address this, we developed an immune dysfunction score, which, when integrated into established prognostic models, improved risk stratification, and prediction of treatment response. Because similar immune alterations were detectable in both bone marrow and peripheral blood, we established a non‑invasive immune monitoring strategy using standardized antibody panels suitable for routine clinical application. Overall, these findings underscore the importance of immune profiling in MDS, highlighting its ability to refine prognostication and more precisely inform therapeutic decision‑making.