Immune signatures of bone marrow cells can independently predict prognosis in patients with myelodysplastic syndrome.

Increasing evidence supports the role of the immune microenvironment and associated signalling in the pathogenesis of myelodysplastic syndromes (MDS). Nevertheless, the clinical relevancy of immune signals in patients with MDS remains elusive. To address this, we used single-sample gene-set enrichment analysis to score immune signatures of bone marrow (BM) samples from 176 patients with primary MDS. Enhanced signatures of 'immature dendritic cells' and 'natural killer cells with cluster of differentiation (CD)56bright' were correlated with better overall survival (OS), whilst higher 'CD103+ signature' was associated with reduced survival. An MDS-Immune-Risk (MIR) scoring system was constructed based on the weighted sums derived from Cox regression analysis. High MIR scores were correlated with higher revised International Prognostic Scoring System (IPSS-R) scores and mutations in ASXL transcriptional regulator 1 (ASXL1), Runt-related transcription factor 1 (RUNX1), and tumour protein p53 (TP53). High-score patients had significantly inferior leukaemia-free survival (LFS) and OS than low-score patients. The prognostic significance of MIR scores for survival remained valid across IPSS-R subgroups and was validated in two independent cohorts. Multivariable analysis revealed that a higher MIR score was an independent adverse risk factor for LFS and OS. We further proposed a model with the combination of MIR score and gene mutations to be complementary to IPSS-R for the prognostication of LFS and OS of patients with MDS.