Deciphering the role of immune system dysfunction in patients with Myelodysplastic Syndrome to refine the prognostic assessment and predict treatment response

Myelodysplastic syndromes (MDS) are a heterogeneous group of myeloid neoplasms characterized by a variable risk of progression to acute myeloid leukemia (AML). Since the pathogenetic mechanisms are still unknown, the diagnosis of MDS is challenging and risk stratification, mainly based on aspecific clinical parameters, fails to exactly predict the prognosis and the response to the current available frontline therapies, by means of hypomethylating agents (HMAs). Considering the increasingly role of immune dysregulation in MDS pathogenesis and progression, we performed a comprehensive analysis of T, Natural Killer (NK) and myeloid cells in bone marrow (BM) and peripheral blood (PB) of 286 MDS and AML post-MDS patients before and after HMA treatments by using high-dimensional flow cytometry. Data analysis was performed implementing an unsupervised pipeline which combines Phenograph and HDBSCAN algorithms. We identified 5 groups of MDS patients based on their immunological profile. These immunological groups, characterized by different grade of immune dysfunction and inflammation, exhibited distinct prognosis and response to HMA therapy. Moreover, since patients classified within the same category according to existing prognostic scoring systems were subdivided into different immunological groups, we developed a tool to assign a score to the immune dysfunction of MDS patients. This score was integrated to existing prognostic models to refine prognostic assessment and better predict therapy response. Since we observed that the perturbated immune environment found in the BM was also present in the PB, we set-up a non-invasive immune monitoring approach based on pre-formulated antibody panels for implementation in the clinical work-up of MDS patients. Together our data provide evidence that the evaluation of the immune profile of MDS patients is of utmost importance in the clinical decision-making process, since it could improve patient risk stratification and the prediction of HMA treatment response.