Immune scoring model based on immune cell infiltration to predict prognosis in diffuse large B-cell lymphoma

Area of Science:

• Oncology
• Immunology
• Bioinformatics

Background:

• Diffuse large B-cell lymphoma (DLBCL) exhibits significant genetic heterogeneity impacting pathogenesis and clinical presentation.
• Existing prognostic studies for DLBCL have often overlooked the crucial role of tumor-infiltrating immune cells.
• This research specifically investigates the influence of tumor-infiltrating immune cells on DLBCL prognosis.

Purpose of the Study:

• To identify key immune cell types that predict prognosis in diffuse large B-cell lymphoma (DLBCL).
• To develop and validate a novel prognostic model for DLBCL based on immune cell infiltration.
• To assess the clinical utility of immune cell-based nomograms for predicting long-term survival in DLBCL patients.

Main Methods:

• Utilized the GSE10846 (training) and GSE53786 (validation) datasets from NCBI's Gene Expression Omnibus.
• Employed the CIBERSORT algorithm to quantify immune cell proportions in patient samples.
• Applied univariate Cox analysis, Lasso regression, and random forest methods to identify prognostic immune cell markers, followed by multifactor Cox analysis and model construction.

Main Results:

• Included 539 samples, with 10 immune cell types identified as significantly associated with overall survival (p < .05).
• A prognostic model was constructed using two key immune cell types identified through Lasso regression and random forest analysis, integrated with clinical data.
• The model's predictive performance and reliability were successfully validated across both the training and validation datasets.

Conclusions:

• An immune cell-based prediction model effectively forecasts prognostic outcomes for patients diagnosed with DLBCL.
• Nomogram plots derived from this model can assist clinicians in evaluating the likelihood of long-term survival for DLBCL patients.