A CIBERSORTx-based immune cell scoring system could independently predict the prognosis of patients with myelodysplastic syndromes

Aside from cell intrinsic factors such as genetic alterations, immune dysregulation in the bone marrow (BM) microenvironment plays a role in the development and progression of myelodysplastic syndromes (MDS). However, the prognostic implications of various immune cells in patients with MDS remain unclear. We adopted CIBERSORTx to estimate the relative fractions of 22 subtypes of immune cells in the BM of 316 patients with MDS and correlated the results with clinical outcomes. A lower fraction of unpolarized M0 macrophages and higher fractions of M2 macrophages and eosinophils were significantly associated with inferior survival. An immune cell scoring system (ICSS) was constructed based on the proportion of these 3 immune cells in the BM. The ICSS high-risk patients had higher BM blast counts, higher frequencies of poor-risk cytogenetics, and more NPM1, TP53, and WT1 mutations than intermediate- and low-risk patients. The ICSS could stratify patients with MDS into 3 risk groups with distinct leukemia-free survival and overall survival among the total cohort and in the subgroups of patients with lower and higher disease risk based on the revised International Prognostic Scoring System (IPSS-R). The prognostic significance of ICSS was also validated in another independent cohort. Multivariable analysis revealed that ICSS independently predicted prognosis, regardless of age, IPSS-R, and mutation status. Bioinformatic analysis demonstrated a significant correlation between high-risk ICSS and nuclear factor KB signaling, oxidative stress, and leukemic stem cell signature pathways. Further studies investigating the mechanistic insight into the crosstalk between stem cells and immune cells are warranted.

Automated summary

This study revealed the association between immune cell ratios in the bone marrow of MDS patients and prognosis, with the ICSS scoring system able to stratify patients into high, intermediate, and low-risk groups and independently predict prognosis.