Objective  To establish and preliminary evaluate a prognostic model of diffuse large B-cell lymphoma (DLBCL) based on tu-mor-infiltrating immune cells in the era of immunochemotherapy.

  Methods  The immune cell abundance identifier (ImmuCellAI) algo-rithm was used to calculate the abundance of 24 immune cells in the DLBCL tumor microenvironment, and the least absolute shrink-age and selection operator (LASSO) regression and COX regression were used to screen predictive variables and build the immune cells-based risk scores model (IRS model). IRSC model was constructed by combining IRS model and the clinical factors of the patients. The Kaplan-Meier method and ROC curve were used to evaluate the model, and nomogram was used to calculate the survival rate at differ-ent time points.

  Results  The overall survival time (OS) of the high-risk patients of the IRS model was significantly lower than that of the low-risk group P=1e-15, HR=0.298 (0.2176-0.4082), and the AUC value of ROC curves based on patients' 1-, 3-, and 5-year surviv-al were 0.728, 0.711 and 0.615, respectively, and the risk score of this model was negatively correlated with the efficacy of Immune checkpoint inhibitors (ICPIs). The predictive value of IRSC model was higher than that of IRS model: the prognosis of the high- risk group was significantly worse than that of the low-risk P < 2e-16, HR = 0.17 (0.1143-0.253). The AUC value of ROC curves based on 1-, 3-, and 5-year survival were 0.797, 0.809 and 0.792, respectively.

  Conclusions  The IRS model can well predict the prognosis of DLBCL patients and the efficacy of ICPIs, while the IRSC model has higher prognostic power.