Exploration of biomarkers for predicting the prognosis of patients with diffuse large B-cell lymphoma by machine-learning analysis

Exploration of biomarkers for predicting the prognosis of patients with diffuse large B-cell lymphoma by machine-learning analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Exploration of biomarkers for predicting the prognosis of patients with diffuse large B-cell lymphoma by machine-learning analysis Shifen Wang, Hong Tao, Xingyun Zhao, Siwen Wu, Chunwei Yang, Yuanfei Shi, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5876864/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 20 Aug, 2025 Read the published version in BMC Immunology → Version 1 posted 8 You are reading this latest preprint version Abstract Background: Diffuse large B cell lymphoma (DLBCL), one distinct origin of hematological malignancies, has caused a major public health problem. However, the molecular mechanisms was not been clearly elucidated. The aim is to explore disease-specific diagnostic biomarkers and mechanisms to improve this situation. Methods: Three microarray datasets (GSE25638, GSE12195, GSE12453) were downloaded from the Gene Expression Omnibus (GEO) database. The key genes in DLBCL patients were screened by differential expression genes (DEGs) analysis and weighted gene co-expression network analysis (WGCNA). Functional enrichment analysis and protein-protein interaction (PPI) network construction were employed to reveal DLBCL-related pathogenic molecules and underlying mechanisms. Random forest analysis was adopted for screening candidate biomarkers, and Kaplan Meier survival analysis were constructed to predict the risk of patients. The single‐sample gene set enrichment analysis was used to explore immune cell infiltration in lymphoma. Validation of the hub genes expression was confirmed by RT-PCR and immunohistochemistry (IHC) tests. Results: 95 key genes were obtained from three datasets about DLBCL patients by DEGs and WGCNA. The four hub genes (CXCL9, CCL18, C1QA, CTSC) were screened by random forest analysis and machine learning algorithm. The ROC results showed that the AUC was 1.00 in the training set, and the bootstrap verification was performed for 1000 times in the external validation set, and the AUC size was 0.839. The several pathways were found by Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analysis. The four hub genes were identified to be excellent potential for the survival of DLBCL patients. Dysregulated immune cell infiltrations were observed in DLBCL, as well as positive correlations with the four hub genes, respectively. Validation of the hub genes with high expressions was also demonstrated in DLBCL patients. Conclusion: This study identified four candidate hub genes (CXCL9, CCL18, C1QA, CTSC) that could predict the risk of DLBCL, and CXCL9 may be essential in developing the disease, which provided a new perspective for the molecular mechanism and therapeutic targets for DLBCL. Diffuse Large B Cell Lymphoma (DLBCL) DEGs WGCNA Hub genes Immune cell infiltration Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1. Introduction Diffuse large B cell lymphoma (DLBCL), the most common subtype of non-Hodgkin lymphoma (NHL), is an aggressive cancer that accounts for 30%-40% of NHL[ 1 , 2 ]. Gene expression profiling (GEP) analysis by microarray analysis revealed three molecular subtypes: germinal center B-cell like (GCB), activated B-cell like (ABC), and unclassifiable[ 3 , 4 ]. These 3 DLBCL subtypes have different clinical and biological heterogeneity with varying responses to treatment and prognosis[ 5 ]. However, limited funding and limited availability of GEP technology make it impractical to generalize to clinical DLBCL patients. The Hans algorithm separates diffuse large B cell lymphoma (DLBCL) into two subtypes: Germinal center B-cell like (GCB) and non-GCB[ 6 ], just as with the GEP classification. Rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) has been shown to have a poor prognosis for non-GCB patients[ 7 , 8 ]. Genetic and functional genomic studies have revealed oncogenic driver pathways in DLBCL[ 9 , 10 ], leading to the exploration of potential targets for the treatment of DLBCL[ 11 ]. Understanding the immunological components of DLBCL has grown more crucial as novel immune-targeting treatments have been created. The pathogenesis of DLBCL involves a complex interaction between tumor cells and tumor microenvironment (TME), and immune cells shape the tumor microenvironment of DLBCL[ 12 ]. There are still many unanswered questions regarding the immunology of DLBCL, such as the immune system and the underlying mechanisms of DLBCL development. This study aimed to analyze the prognostic role and immune infiltration of potential markers of DLBCL in DLBCL by bioinformatics analysis, and to elucidate the possible mechanisms. Our findings suggest that CXCL9, CCL18, C1QA, and CTSC are prognostic biomarkers and poor prognostic indicators in DLBCL patients. In summary, our study provides evidence that four key genes (CXCL9, CCL18, C1QA, CTSC) are important in the development and progression of DLBCL and suggests that these genes may be novel biomarkers and new therapeutic targets for DLBCL. The potential prognostic biomarkers of DLBCL were screened by bioinformatics analysis. We explored the value of four key genes in DLBCL and their molecular mechanisms, providing a basis for future immunotherapy and precision medicine. The predictive performance was assessed using survival analysis and receiver operating characteristic curve analysis. To further understand the role of TME in the pathogenesis of DLBCL patients, we analyzed the relationship between immune cell subtypes and clinical prognosis in the DLBCL cohort. In addition, we examined the correlation between four key genes and the immune process in DLBCL. 2. Materials and methods 2.1 Microarray data The study flowchart is shown in Fig. 1 . Three microarray datasets (GSE25638, GSE12195, GSE12453) were retrieved from the NCBI Gene Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/ ) database. Dataset GSE25638 included data on 26 DLBCL patients and 13 controls. Dataset GSE12195 included data on 73 DLBCL patients and 10 controls. Dataset GSE12453 included data on 11 DLBCL patients and 25 controls. 2.2 Data processing and differentially expressed genes (DEGs) identification DEGs were calculated between disease and control groups with it via the “Limma” package in R software. For DLBCL, adjusted P value 1 were used to identify the DEGs. Next, the difference analysis results for each group were presented using the heatmap and volcano plot. In the three plots, blue indicated low expression, and yellow indicated high. 2.3 Functional enrichment analysis and protein-protein interaction (PPI) network For identifying the biological functions and signaling pathways of the shared genes, the R package (R4.3.1) "clusterProfiler" performed an enrichment analysis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) pathways. The shared genes were entered into the website ( https://cn.string-db.org/ ) in order to create the protein interaction relationship data. The data were then imported into the Cytoscape software for computation, prediction, and visualization. 2.4 Significant module identification via WGCNA Weighted gene co-expression network analysis (WGCNA) is one of the most popular algorithms for computing large volumes of data, which evaluates gene expression correlations and visualizes co expression networks through clustering and modularization. Data were checked to identify the outliers in the samples by hierarchical clustering. There were no offending samples in the DLBCL dataset (GSE25638) and no samples were removed. The “pickSoftThreshold” package function was utilized to screen out the optimal soft threshold. Through a topological overlap matrix (TOM) analysis, the adjacency matrix was clustered and categorized genes with similar expression patterns into gene modules, that is, the module eigengenes (MEs). In addition, the correlations between MEs and clinic traits were calculated via Pearson’s correlation test and the strongest positive correlation was selected for further analysis. 2.5 Machine learning algorithm Previous reviews reported that the prediction model of random forest showed better diagnostic efficiency than other models[ 13 , 14 ], so the random forest algorithm was selected for variable screening in this study. These three DLBCL datasets were subjected to the random forest machine learning algorithm in order to further uncover potential biomarkers. The most significant factors were determined via random forest analysis using a decision tree method[ 15 ]. Random forest (RF) model is developed from decision trees. RF can generate hundreds or even thousands of trees. The samples of each tree are extracted from the bags in a defined set by the Bootstrap method as the training samples, and the remaining data, called Out-Of-Bag data (OOB), is used as the test samples. The principle of screening variables is mainly to rank the importance of features according to the error generated before and after rearranging features. Based on the OOB test set, the prediction accuracy of the original features and the randomly arranged features after permutation was compared to measure the feature importance. We were able to identify DLBCL characteristic genes by filtering the shared genes using this approach. In order to determine the optimal number of trees, a random forest model with 500 trees on the discovery queue is established, and cross-validation error is used. Next, the top 10 genes were sequenced according to their significance, and their distributions were plotted. The final outcome was determined for each DLBCL dataset by setting the significance threshold at 0.5. By using this algorithm, key genes were selected from three DLBCL datasets, and we took the intersection of the algorithm results from each dataset. 2.6 Nomogram construction and receiver operating characteristic (ROC) evaluation The nomogram was constructed based on the four hub genes by using the “rms” package. Each gene’s relative expression level corresponds to a score based on the nomogram. To evaluate the predictive value of four candidate biomarker, we generated ROC curves and calculated the area under the curve (AUC) and 95% confidence interval (CI). ROC curve was performed to determine whether the nomogram-based decision was conducive to DLBCL diagnosis. The bootstrap algorithm was used to validate the results on the external dataset GSE83632. The optimal AUC for predicting the risk of DLBCL was > 0.7. 2.7 Survival Analysis Log-rank tests and Kaplan-Meier (KM) graphs were used to assess how the groups' overall survival differed from one another. These were plotted and analyzed in the survminer package and the survival package in R. Each survival curve's cut-offs for patient classification are given in the figure legends. Using the survival package in R, multivariate Cox regression was carried out while taking the patient's sex, tumor stage, and grade into consideration. 2.8 Immune cell infiltration For RNA samples from different tissue types, immune cell subpopulation infiltration scores in the tumor microenvironment are computed using the single-sample gene set enrichment analysis (ssGSEA) algorithm[ 16 ]. The algorithm "ssGSEA" can converted the normalized gene expression matrix into gene enrichment score matrix. The analysis's 28 immune cells' specific gene maps came from a recent publication ( Supplementary Table 2 ) [ 17 ]. The proportion of each immune cell in each sample was visualized from the barplot. A boxplot was used to compare the expression of the differences for each immune cell between the two groups. A heatmap displaying the correlation of different immune cells in DLBCL pathogenesis was constructed using the R package "corrplot". The correlation between characteristic genes and immune cell infiltration is also examined using the ssGSEA algorithm and constructed using the R package "ggcorrplot". 2.9 Isolation of human PBMCs and RT-PCR A total of 13 samples were collected from the First Affiliated Hospital, Zhejiang University School of Medicine, including 13 DLBCL samples and 5 healthy controls. Using EDTA anticoagulant blood and FicollPaque density gradient centrifugation, human peripheral blood mononuclear cells (PBMCs) were separated. TRIzol reagent was used to extract total RNA from the PBMCs of DLBCL patients and healthy controls, and the Prime ScriptTM RT Reagent Kit (TaKaRa, China) was used to create cDNA through reverse transcription. The internal reference was β-Actin. The SYBR Green PCR Kit (TaKaRa, China) was used for quantitative realtime polymerase chain reaction (RT-PCR) in accordance with the manufacturer's instructions. Furthermore, the relative expression was calculated by the 2 −△△CT method. All reactions were repeated in triplicate. The primer sequences of target genes are displayed in Supplementary Table 1, which are from public website ( https://www.origene.com/ ). 2.10 Immunohistochemical analysis The lymph node tissues of DLBCL patients and healthy controls were collected from the Department of Pathology, the First Affiliated Hospital of Zhejiang University School of Medicine. To construct the tissue microarray (TMA), a 3-mm-diameter core of the cancer tissue area was generated. The brief steps for immunohistochemistry (IHC) were as follows: First, slides made from the TMA were deparaffinized in xylene and rehydrated in ethanol, and the antigens were extracted for IHC tests with specific antibodies. Second, the slides were blocked with goat serum for 1 h, and after washing, the slides were further stained with a diluted anti-human CXCL9 antibody (Abcam, ab290643). Then, 3,3′-diaminobenzidine (DAB) was used to stain the tissues after 24 h. The IHC results of CXCL9 staining scores were semiquantified by an immunoreactive scoring (IRS) system, which has been previously described[ 18 ]. 2.11 Statistical analysis R software and corresponding packages were used for calculations and analysis. We used Wilcoxon rank sum test and Kruskal-Wallis test to determine differences between two or more groups. Spearman’s method used to analyze the correlation of variables. ROC curve analysis was used to determine the diagnostic value of biomarkers. Survival curves were plotted via the Kaplan-Meier log rank test. In the prognostic assessment analysis, a two-sided P value of < 0.05 was considered statistically significant. Of note, in Fig. 7 B, as the data distribution did not conform to the normal distribution, we used the wilicox test to compare the difference in the abundance of immune cells between the tumor group and the control group. 3. Results 3.1 Screening of DEGs from GSE25638、GSE12195 and GSE12453 Using data from the open NCBI GEO database, we obtained the Series Matrix File data file for GSE25638、GSE12195 and GSE12453 for DLBCL, respectively ( Table 1 ) . A total of 39 sets of transcriptome data, including 13 from normal samples and 26 from DLBCL samples in GSE25638. The DEGs between the two groups of samples were identified using the limma package. A total of 2003 DEGs (including 1642 upregulated and 361 downregulated genes) were identified in DLBCL (Fig. 2 A). In this cohort, we observed more up-regulated genes and fewer down-regulated genes in DLBCL samples compared to normal samples. A total of 36 sets of transcriptome data, including 25 from normal samples and 11 from DLBCL samples in GSE12453. A total of 411 DEGs (including 333 upregulated and 78 downregulated genes) were identified in DLBCL (Fig. 2 B). In this cohort, we also observed more up-regulated genes and fewer down-regulated genes in DLBCL samples compared to normal samples. However, a total of 83 sets of transcriptome data, including 10 from normal samples and 73 from DLBCL samples in GSE12195. In this cohort, we observed that there was little difference between up-regulated and down-regulated genes in DLBCL samples compared to normal samples (Fig. 2 C), which may be related to the key factors of immune microenvironment status. Then, we intersected the upregulated DEGs of the two datasets (GSE12195 and GSE12453), resulting in a total of 222 intersecting DEGs (Fig. 2 E and 2 F). Table 1 Basic information of GEO datasets used in the study. ID GSE series Disease Samples Sourcetypes Platform 1 GSE25638 DLBCL 26 DLBCL patients and 13 normal donor controls lymphoid tissue GPL570 2 GSE12195 DLBCL 73 DLBCL patients and 10 normal donor controls lymphoid tissue GPL570 3 GSE12453 DLBCL 11 DLBCL patients and 25 normal donor controls lymphoid tissue GPL570 Note: GSE, Gene expression omnibus series; DLBCL, Diffuse large B cell lymphoma; GPL, Gene expression omnibus platform. 3.2 Co-expression network construction and key module identification for WGCNA To further explore the key gene associations of DLBCL, WGCNA was performed to identify the most relevant gene modules in addition to analyzing the differential expression genes. The PCA score plot showed significant differences between the disease and control groups in DLBCL (Fig. 3 A), which could be used for subsequent analysis. We first selected a suitable soft threshold β value that met the scale-free conditions (Fig. 3 B and 3 C). The thresholds β were set to 20 for the dataset (GSE25638) and then the tom matrix was used to identify gene modules (Fig. 3 C). Then the TOM was converted from the adjacent matrix and turned into the dissTOM (dissTOM = 1-TOM) for hierarchical clustering. Dynamic tree shearing clustered the similarity genes based on the topological overlap and then divided them into various modules (Fig. 3 D). A total of 5 gene modules were found, among which the ME grey module and the ME turquoise module had the strongest positively relationship with DLBCL (1034 genes, R = 0.95/0.9, p = 3e-20/4e-15) (Fig. 2 D). On this basis, the two set of genes was compared with the 222 up-regulated intersection DEGs of GSE12195 and GSE12453 identified above, and a total of 95 intersection genes were obtained (Fig. 2 E). These 95 genes were considered for our subsequent studies. 3.3 Functional enrichment analysis of the pathogenic genes involved in DLBCL To reveal the potential pathogenic genes and underlying mechanism in DLBCL, we conducted pathway analysis on the 95 shared genes from these three datasets. The findings revealed that the DLBCL-related genes were primarily enriched in pathways such as inflammatory response, secretory granule, and collagen binding (Fig. 4 A). These genes of DLBCL were mainly enriched in mineral absorption and pertussis (Fig. 4 B). GO analysis showed that BP terms were significantly enriched included "inflammatory response", "biological process involved in interspecies interaction between organisms", and "defense response". Moreover, "secretory granule", "extracellular region", and " extracellular space" were enriched for CC, while the MF of DEGs were highly associated with "collagen binding", "CXCR3 chemokine receptor binding ", and "peptidase regulator binding " (Fig. 4 A). The results of KEGG analysis showed that the DLBCL shared genes were mainly enriched in CXCR chemokine receptor binding and chemokine activity (Fig. 4 C). The top five pathways identified by gene set enrichment analysis (GSEA) were “Circadian rhythm”, “Drug metabolism-cytochrome P450”, “Phenylalanine metabolism”, “Taurine and hypotaurine metabolism” and “Tyrosine metabolism” (Fig. 4 C). The tail five pathways were “Biosynthesis of amino acids”, “Cell cycle”, “DNA replication”, “Primary immunodeficiency” and “Proteasome” (Fig. 4 D). Meanwhile, GSEA analysis also showed that these genes were mainly enriched in CXCR chemokine receptor binding and chemokine activity (Fig. 4 E). Then, the PPI showing the shared genes in the DLBCL dataset (Fig. 4 F). Based on the 95 shared genes, a PPI network was preliminarily constructed to select hub genes for DLBCL. After eliminating DEGs with poor interaction, 68 genes were retained (Fig. 4 F). 3.4 Identify potential shared diagnostic genes based on machine learning algorithms To further explore the diagnostic value of DEGs in DLBCL, we use the machine learning algorithms. Based on the above 95 shared genes mentioned above, we employed learning approaches for the three lymphoma datasets to confirm variables significantly associated DLBCL. Gene importance was calculated by random forest for DEGs. The 10 most important genes from the random forest results for the three lymphoma datasets are shown in Fig. 5 A-C. Then, a subset of 4 overlapping hub genes (CXCL9, CCL18, C1QA, CTSC) were selected from the random forest machine learning algorithms for the three lymphoma datasets (Fig. 5 D). 3.5 Diagnostic value and prognostic correlation of DLBCL hub genes Compared with the control, the predictive nomogram was constructed, whereby each gene’s relative expression corresponded to a score, and the total score was obtained by the summation of the score of each gene (Supplementary Fig. 1A). ROC curve analysis showed good predictive performance of each gene as follows (Supplementary Fig. 1B): CXCL9 (AUC: 1.000); CCL18 (AUC:1.000); C1QA (AUC: 1.000); CTSC (AUC: 1.000). ROC analysis of the nomogram yielded an AUC of 1.000, demonstrating a high predictive value for DLBCL. Then, in order to determine the performance of the screened markers, we performed 1000 bootstrap validation on the external dataset GSE83632, and the average AUC under the 1000 ROC curves was 0.839 (95%CI: 0.774–0.904) (supplementary Fig. 2), which further proved the predictive efficacy of these four candidate genes. Moreover, we analyzed the prognostic predictive value of 4 hub genes in the two subgroups of DLBCL. Kaplan–Meier curves showed that high expression groups of CXCL9, CCL18, C1QA and CTSC had significantly shorter overall survival than their low expression groups, there was statistical difference (Fig. 6 A). Kaplan-Meier survival analysis of the two subtypes of DLBCL indicated that, overall, the GCB subgroup exhibited a prominent advantage of median survival time, whereas the non-GCB subgroup presented with a worse prognosis (Fig. 6 B). 3.6 Immune infiltration analysis of shared diagnostic genes Considering that DLBCL are characterized by a high immune response. The abundances of immune cells in different groups were analyzed with ssGSEA. In each group, the proportion of 28 immune cells was shown as a bar plot ( Supplementary Table 2 ). Generally, the barplot (Fig. 7 A) clearly illustrated that the proportion of activated CD8 + T cells, Myeloid-derived suppressor cells (MDSCs), activated CD8 + T cells, central memory CD4 + T cells, and effector memory CD8 + T cells were higher in DLBCL. Compared with the control samples, the activated B cell, eosinophils, immature B cells, and neutrophils were increased in the DLBCL samples (Fig. 7 B). In contrast, B cells and granulocytes correlated negatively. Moreover, the relationship between biomarkers and immune cell contents was investigated. In DLBCL samples, NK cells and T cells were significantly positively correlated with 4 hub genes (Fig. 8 C). CXCL9 was significantly positively correlated with CD56 bright NK cells and activated CD4 + T cells, negative for macrophages immature B cells, neutrophil and eosinophil (Fig. 8 C). Moreover, T cell subsets are also involved in immune regulation of four key genes, such as regulatory T cells, T follicular helper cells and Th2 cells. It appears that immune function is crucial to the development of DLBCL (Fig. 7 C). 3.7 Validation of the four optimal key genes The relative expression of four optimal key genes was verified with RT-PCR in DLBCL patients and healthy controls. Compared with the control group, the expression of CXCL9 (Fig. 8 A), CCL18 (Fig. 8 B), CTSC (Fig. 8 C) and C1QA (Fig. 8 D) was significantly upregulated in DLBCL patients (all P < 0.05), which was consistent with the results of this bioinformatic analysis. Among them, CXCL9 was the most significantly differentially expressed among the four genes(P < 0.01) (Fig. 8 A). Immunohistochemical (IHC) analysis revealed that, compared with that in the lymph node tissues of the healthy group, CXCL9 protein was highly expressed in DLBCL tissues (Fig. 8 E). 4. Discussion The Limma R package provides a comprehensive solution for analyzing gene expression data [58]. In the present study, 222 DEGs were identified between DLBCL and control groups. WGCNA is a systematic biological technique for analyzing gene association patterns among samples. The most relevant correlation between gene modules and phenotypes can be calculated by building a gene co-expression network and determining related gene clusters. Herein, we found 1034 significant module genes of DLBCL and applied Limma analysis and WGCNA to screen the gene clusters of connected, shared, and correlating genes in DLBCL. The intersection of DEGs regarding DLBCL (n = 222) and module genes regarding DLBCL (n = 1034) yielded 95 common risk genes associated with DLBCL. To elucidate the roles of these 95 DEGs in DLBCL, KEGG pathway and GO analysis were performed. These DEGs were mainly enriched in KEGG pathways "Mineral absorption”, “Pertussis”, “Lysosome”, “Complement and coagulation cascades” and “Fluid shear stress and atherosclerosis”, suggesting a correlation with immunoinflammatory responses. Mineral absorption was confirmed to be one of the most significantly enriched pathways in the proteome and phosphoproteome, which is involved in the process of mammalian immunosuppression[ 19 ]. Pertussis and lysosome have also been associated with immune responses dysfunction[ 20 , 21 ]. A previous study has confirmed that complement and coagulation cascades are one of the pathophysiological pathways in the pathogenesis of acquired immunodeficiency syndrome-related non-Hodgkin lymphoma[ 22 ]. Recent studies have further shown that the complement and coagulation cascade pathways play a key role in the anti-tumor process of 10-hydroxy-2-decenoic acid (10-HAD), which provides a potential therapeutic approach for lymphoma treatment[ 23 ]. Significantly enriched GO terms associated with BP included "inflammatory response", "immune response", "defense response", “immune system process”, which are well-established integral components of DLBCL pathogenesis[ 24 , 25 ]. Moreover, the CC and MF analysis of DEGs were closely related to inflammatory response and innate and adaptive immune response. These shared DEGs were primarily enriched in CXCR chemokine receptor binding and chemokine activity, according to GSEA analysis. Taken together, these findings demonstrate that these 95 DEGs were correlated with DLBCL pathogenesis. 95 DEGs in DLBCL were mapped to the PPI network to further explore their potential interplay, and 68 DEGs were identified. Using the machine learning method (random forest), four hub DEGs (CXCL9, CCL18, C1QA, CTSC) were identified. We constructed a nomogram and evaluated its predictive value for DLBCL in DLBCL patients. First, we identified four key immune-associated genes (CXCL9, CCL18, C1QA, CTSC) for constructing our nomogram. Next, we conducted validation which showed that the four hub genes had a good predictive value of DLBCL patients, indicating their potential roles during the DLBCL progression. The expression levels of four hub genes were significantly correlated with the prognosis of DLBCL. We further investigated the role of four hub genes in the survival of patients with DLBCL. Kaplan-Meier survival curve showed that the increased mRNA levels of four hub genes were significantly associated with poor prognosis of DLBCL patients (P < 0.05). Furthermore, we discovered a correlation between the genesis of DLBCL and the levels of the four hub genes. A major factor in the categorization of DLBCL, which is linked to the prognosis of patients and a hot issue in the pathological mechanism of DLBCL, is the cell of origin. GCB-DLBCL originate from the germinal center (GC) stage, while nonGCB-DLBCL originates from the ABC/post-GC stage, among which nonGCB-DLBCL has the worse prognosis. The potential of gene level as a prognostic marker for subtype categorization is demonstrated by our discovery that four hub genes' expression levels may further separate GCB-DLBCL/nonGCB-DLBCL into subgroups with low expression with favorable prognosis and high expression with bad prognosis. Current evidence suggests that the CXCL9 is produced by lymphoma cells and can elicit cytotoxic responses via CXCR3, which promotes tissue necrosis and vascular damage by recruiting cytotoxic lymphocytes[ 26 ]. Our findings are in accordance with earlier research that shown CXCL9 is significantly expressed in DLBCL tissues and cell lines and is strongly associated with patient survival and clinical progression[ 27 ]. Notably, Epstein-Barr virus (EBV)-negative DLBCL does not secrete chemokines such as CXCL9 in contrast to EBV-positive DLBCL patients [ 26 ], and we speculate that CXCL9 promotes DLBCL progression and may be involved in the immune-modulatory mechanism of DLBCL, which is consistent with the results of our KM curve analysis. Another study reported that in the pathological mechanism of combination therapy of IL15 receptor agonists with CART in relapsed/refractory B malignancies, an increase in chemokines including CXCL9 drives the trafficking of lymphocytes to tissues, which is associated with a high durable response rate[ 28 ]. However, little is known about CXCL9's actions and underlying mechanism in DLBCL. There is controversy regarding the association between CCL18 and the occurrence and severity of DLBCL[ 29 , 30 ]. It is generally believed that the biological functions of CCL18 and its co-expressed genes are prominent in cell migration, proliferation and apoptosis[ 31 ]. Tumor-associated macrophages secrete many chemokines into the tumor microenvironment, especially a large amount of CCL18, which is one of the important mechanisms of the immunosuppressive nature of the tumor microenvironment and cancer immune evasion[ 32 ]. Investigating the precise regulation mechanism of CCL18 in DLBCL is crucial since chemokines have anti-cancer and pro-cancer effects in the tumor microenvironment. The C1QA gene is known to encode the A-chain polypeptide of serum complement subcomponent C1q. C1QA is empirically associated with the outcome of rituximab treatment of DLBCL. Jin et al. found that DLBCL patients with homozygous A -allele carriers of the gene C1qA [276] had better overall response and higher complete response, and overall survival in receiving R-CHOP[ 33 ]. Of note, C1qA may originate from other non-tumor cells in DLBCL tissue. Blood monocytes and tumor-associated macrophages also expressed C1qA and may affect tumor cells through it. More research is needed to determine whether exogenous C1qA protein level is affected by Rituximab. C1QA is derived from rituximab resistant DLBCL cells under rituximab stress and is regulated by METTL3 and YTHDF2-mediated m6A methylation[ 34 ]. Notably, C1QA may originate from other non-tumor cells in DLBCL tissues such as blood monocytes and tumor-associated macrophages[ 34 ]. However, it is unclear whether rituximab affects the levels of exogenous C1QA protein. A lysosomal cysteine protease called cathepsin C (CTSC), commonly referred to as dipeptidylpeptidase I, is necessary for the catalytic activation of several serine proteases[ 35 ]. In a conjoined gene study of non-Hodgkin B-cell lymphoma cell line KPU-UH1, novel conjoined genes, including CTSC-RAB38, were detected and may be genetic biomarkers for lymphoma[ 36 ]. Recent studies have found that tumor-secreted protease cathepsin C (CTSC) promotes tumour metastasis by regulating recruitment of neutrophils and formation of neutrophil extracellular traps (NETs)[ 37 ]. Validation by RT-PCR, hub gene expression in DLBCL was statistically significant when compared to healthy control, consistent with previous analyses. IHC tests further demonstrated that DLBCL had high levels of CXCL9 protein expression. The proportion immune cells have a significant impact on DLBCL. A high percentage of PD1 + CD8 + and PD-L1 + T cells in the TME was reported to predict poor survival in DLBCL in one study, but T cells with high expression of the immunological checkpoint cytotoxic T-lymphocyte-associated protein 4 (CTLA4) had a positive prognosis[ 38 ]. A significant percentage of DLBCL in our investigation was attributed to activated T cells, NK cells, and macrophages, which were consistent with previous single-cell sequencing results[ 39 ]. Immunosuppressive cells were also found to be high in DLBCL in our study, such as Treg cells, MDSCs, and macrophages. It is speculated that immune-promoting and immune-suppressive mechanisms play a key role in the pathogenesis of DLBCL. Then, we examined the relationship between the four identified genes (CXCL9, CCL18, C1QA, CTSC) and immunity. The four hub genes positively correlated with most immune-related cells. However, there are many limitations associated with the current manuscript. First, this study relies on bioinformatics analysis on public datasets, which could generate some inconsistencies in the results because of its retrospective nature. Moreover, due to the sample size limitation of our study, the accuracy of the results needs to be further verified in larger independent cohorts. Furthermore, additional molecular experiments are required to investigate the underlying mechanisms. 5. Conclusion In summary, through bioinformatics analysis such as DEGs and WGCNA, machine learning algorithms, and GO and KEGG enrichment analysis, four co-hub genes (CXCL9, CCL18, C1QA, CTSC) were screened out by the system. We developed a nomogram model to predict the risk of DLBCL and identified four key genes with prognostic value, which laid the foundation for further study of potential key candidate genes in DLBCL patients. In addition, we also found significant differences in immune cell abundance in DLBCL patients. The possible diagnostic or therapeutic value of these immune cells and genes in DLBCL warrants further investigation. Abbreviations DLBCL Diffuse large B cell lymphoma GEO Gene Expression Omnibus DEGs Differentially expressed genes WGCNA Weighted gene co-expression network analysis PPI Protein-protein interaction RT-qPCR real time quantitative PCR IHC Immunohistochemistry KEGG Kyoto Encyclopedia of Genes and Genomes GO Gene Ontology NHL Non-Hodgkin lymphoma GEP Gene expression profiling GCB germinal center B-cell like ABC activated B-cell like CHOP Rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone TME Tumor microenvironment TOM Topological overlap matrix Mes module eigengenes ROC Receiver operating characteristic curves AUC Area under the curve CI confidence interval KM Kaplan-Meier ssGSEA single-sample gene set enrichment analysis PBMCs Peripheral blood mononuclear cells RT-PCR Real-time polymerase chain reaction TMA Tissue microarray GSEA Gene set enrichment analysis Declarations Acknowledgments We would like to thank the contributions of public databases such as the GEO to human medicine. Moreover, we sincerely appreciate the highly qualified native English speaking editors at American Journal Experts (AJE) for providing reputable English language editing service for our manuscript. Authors ’ contributions DC and SW designed the study. SW, XZ, SWu, CY and DC performed the data analysis. SWu, CY and DC participated in the collection and detection of clinical samples from patients and healthy controls. SW and DC prepared the figures and manuscript. In addition, HT, YS, ZX and DC critically revised the manuscript. All authors contributed to the manuscript and approved the submitted version. Funding This work was supported by the Key Research and Development Plan of Department of Science and Technology of Zhejiang Province (2024C03156), National Natural Science Foundation of China (81871709, 82270175, 82300213), Natural Science Foundation of Fujian Province of China (2021J02040), National Key Clinical Specialty Discipline Construction Program (2021-76), Fujian Provincial Clinical Research Center for Hematological Malignancies (2020Y2006). Data availability statement Data is provided within the manuscript or supplementary information files. Ethics approval and consent to participate This project was conducted in accordance with the guidelines of the Helsinki Declaration, and approved by the Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine. All involved participants provided written informed consent. Consent for publication Not applicable. Clinical Trial Number Not applicable. References Alaggio R, Amador C, Anagnostopoulos I, The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours. Lymphoid Neoplasms, Leukemia[J], : 2022;36(7):1720–1748. http://dx.doi.org/10.1038/s41375-022-01620-2 Campo E, Jaffe ES, Cook JR, et al. The International Consensus Classification of Mature Lymphoid Neoplasms: a report from the Clinical Advisory Committee. Blood[J]. 2022;140(11):1229–53. http://dx.doi.org/10.1182/blood.2022015851 . Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature[J]. 2000;403(6769):503–11. http://dx.doi.org/10.1038/35000501 . Rosenwald A, Wright G, Leroy K, et al. Molecular diagnosis of primary mediastinal B cell lymphoma identifies a clinically favorable subgroup of diffuse large B cell lymphoma related to Hodgkin lymphoma. J experimental medicine[J]. 2003;198(6):851–62. http://dx.doi.org/10.1084/jem.20031074 . Lossos IS, Czerwinski DK, Alizadeh AA, et al. Prediction of survival in diffuse large-B-cell lymphoma based on the expression of six genes. New Engl J medicine[J]. 2004;350(18):1828–37. http://dx.doi.org/10.1056/NEJMoa032520 . Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood[J]. 2004;103(1):275–82. http://dx.doi.org/10.1182/blood-2003-05-1545 . Fu K, Weisenburger DD, Choi WW, et al. Addition of rituximab to standard chemotherapy improves the survival of both the germinal center B-cell-like and non-germinal center B-cell-like subtypes of diffuse large B-cell lymphoma. J Clin oncology: official J Am Soc Clin Oncology[J]. 2008;26(28):4587–94. http://dx.doi.org/10.1200/jco.2007.15.9277 . Meyer PN, Fu K, Greiner TC, et al. Immunohistochemical methods for predicting cell of origin and survival in patients with diffuse large B-cell lymphoma treated with rituximab. J Clin oncology: official J Am Soc Clin Oncology[J]. 2011;29(2):200–7. http://dx.doi.org/10.1200/jco.2010.30.0368 . Reddy A, Zhang J, Davis NS, et al. Genetic and Functional Drivers of Diffuse Large B Cell Lymphoma. Cell[J]. 2017;171(2):481–e494. 15.http://dx.doi.org/10.1016/j.cell.2017.09.027 . Schmitz R, Wright GW, Huang DW et al. Genetics and Pathogenesis of Diffuse Large B-Cell Lymphoma, The New England journal of medicine[J] 2018;378(15):1396–407 http://dx.doi.org/10.1056/NEJMoa1801445 Melani C, Lakhotia R, Pittaluga S et al. Combination Targeted Therapy in Relapsed Diffuse Large B-Cell Lymphoma, The New England journal of medicine[J] 2024;390(23):2143–55 http://dx.doi.org/10.1056/NEJMoa2401532 Liu Y, Zhou X, Wang X. Targeting the tumor microenvironment in B-cell lymphoma: challenges and opportunities. J Hematol oncology[J]. 2021;14(1):125. http://dx.doi.org/10.1186/s13045-021-01134-x . Nickson D, Meyer C, Walasek L, et al. Prediction and diagnosis of depression using machine learning with electronic health records data: a systematic review. BMC Med Inf Decis making[J]. 2023;23(1):271. http://dx.doi.org/10.1186/s12911-023-02341-x . Tiruneh SA, Vu TTT, Rolnik DL et al. Machine Learning Algorithms Versus Classical Regression Models in Pre-Eclampsia Prediction: A Systematic Review, Current hypertension reports[J] 2024;26(7):309–23 http://dx.doi.org/10.1007/s11906-024-01297-1 Gregory GA, Robinson TIG, Linklater SE, et al. Global incidence, prevalence, and mortality of type 1 diabetes in 2021 with projection to 2040: a modelling study, The lancet. Diabetes endocrinology[J]. 2022;10(10):741–60. http://dx.doi.org/10.1016/s2213-8587(22)00218-2 . Barbie DA, Tamayo P, Boehm JS, et al. Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1. Nature[J]. 2009;462(7269):108–12. http://dx.doi.org/10.1038/nature08460 . Charoentong P, Finotello F, Angelova M, et al. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell reports[J]. 2017;18(1):248–62. http://dx.doi.org/10.1016/j.celrep.2016.12.019 . Cao H, Zhao J, Chen Z, et al. Loss of LIMCH1 predicts poor prognosis in patients with surgically resected Lung Adenocarcinoma: A study based on Immunohistochemical Analysis and Bioinformatics. J Cancer[J]. 2021;12(1):181–9. http://dx.doi.org/10.7150/jca.47883 . Wang J, Wang X, Zhang C, et al. Microplastics induce immune suppression via S100A8 downregulation. Ecotoxicol Environ safety[J]. 2022. http://dx.doi.org/10.1016/j.ecoenv.2022.113905 . 242:113905. Gros F, Muller S. The role of lysosomes in metabolic and autoimmune diseases, Nature reviews. Nephrology[J]. 2023;19(6):366–83. http://dx.doi.org/10.1038/s41581-023-00692-2 . Miguelena Chamorro B, De Luca K, Swaminathan G, et al. Bordetella bronchiseptica and Bordetella pertussis: Similarities and Differences in Infection, Immuno-Modulation, and Vaccine Considerations. Clin Microbiol reviews[J]. 2023;36(3):e0016422. http://dx.doi.org/10.1128/cmr.00164-22 . Zhuang K, Zhang Y, Mo P, et al. Plasma proteomic analysis reveals altered protein abundances in HIV-infected patients with or without non-Hodgkin lymphoma. J Med virology[J]. 2022;94(8):3876–89. http://dx.doi.org/10.1002/jmv.27775 . Tian Y, Liu X, Wang J, et al. Antitumor Effects and the Potential Mechanism of 10-HDA against SU-DHL-2 Cells, Pharmaceuticals (Basel. Switzerland)[J]. 2024;17(8). http://dx.doi.org/10.3390/ph17081088 . Manfroi B, De Grandis M, Moreaux J et al. The microenvironment of DLBCL is characterized by noncanonical macrophages recruited by tumor-derived CCL5, Blood advances[J] 2021;5(21):4338–51 http://dx.doi.org/10.1182/bloodadvances.2021004203 Nielsen KR, Steffensen R, Bendtsen MD, et al. Inherited Inflammatory Response Genes Are Associated with B-Cell Non-Hodgkin's Lymphoma Risk and Survival. PloS one[J]. 2015;10(10):e0139329. .http://dx.doi.org/10.1371/journal.pone.0139329 . Higuchi T, Hashida Y, Matsuo K, et al. EBV-positive pyothorax-associated lymphoma expresses CXCL9 and CXCL10 chemokines that attract cytotoxic lymphocytes via CXCR3. Cancer science[J]. 2023;114(6):2622–33. http://dx.doi.org/10.1111/cas.15782 . Ruiduo C, Ying D, Qiwei W. CXCL9 promotes the progression of diffuse large B-cell lymphoma through up-regulating β-catenin. Biomed pharmacotherapy = Biomedecine pharmacotherapie[J]. 2018;107:689–95. http://dx.doi.org/10.1016/j.biopha.2018.07.171 . Srinagesh H, Jackson C, Shiraz P, et al. A phase 1 clinical trial of NKTR-255 with CD19-22 CAR T-cell therapy for refractory B-cell acute lymphoblastic leukemia. Blood[J]. 2024;144(16):1689–98. http://dx.doi.org/10.1182/blood.2024024952 . Chen Q, Zheng T, Lan Q, et al. Single-nucleotide polymorphisms in genes encoding for CC chemokines were not associated with the risk of non-Hodgkin lymphoma, Cancer epidemiology, biomarkers & prevention: a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive. Oncology[J]. 2013;22(7):1332–5. http://dx.doi.org/10.1158/1055-9965.Epi-13-0328 . Zhou Q, Huang L, Gu Y et al. The expression of CCL18 in diffuse large B cell lymphoma and its mechanism research, Cancer biomarkers: section A of Disease markers[J] 2018;21(4):925–34 http://dx.doi.org/10.3233/cbm-171097 Korbecki J, Grochans S, Gutowska I, et al. CC Chemokines in a Tumor: A Review of Pro-Cancer and Anti-Cancer Properties of Receptors CCR5, CCR6, CCR7, CCR8, CCR9, and CCR10 Ligands. Int J Mol sciences[J]. 2020;21(20). http://dx.doi.org/10.3390/ijms21207619 . Korbecki J, Olbromski M, Dzięgiel P. CCL18 in the Progression of Cancer. Int J Mol sciences[J]. 2020;21(21). http://dx.doi.org/10.3390/ijms21217955 . Jin X, Ding H, Ding N, et al. Homozygous A polymorphism of the complement C1qA276 correlates with prolonged overall survival in patients with diffuse large B cell lymphoma treated with R-CHOP. J Hematol oncology[J]. 2012. 5:51.http://dx.doi.org/10.1186/1756-8722-5-51 . Li J, Zhu Z, Zhu Y, et al. METTL3-mediated m6A methylation of C1qA regulates the Rituximab resistance of diffuse large B-cell lymphoma cells. Cell death discovery[J]. 2023;9(1):405. http://dx.doi.org/10.1038/s41420-023-01698-2 . Korkmaz B, Caughey GH, Chapple I, et al. Therapeutic targeting of cathepsin C: from pathophysiology to treatment. Pharmacol therapeutics[J]. 2018;190:202–36. http://dx.doi.org/10.1016/j.pharmthera.2018.05.011 . Matsumoto Y, Tsukamoto T, Chinen Y, et al. Detection of novel and recurrent conjoined genes in non-Hodgkin B-cell lymphoma. J Clin experimental hematopathology: JCEH[J]. 2021;61(2):71–7. http://dx.doi.org/10.3960/jslrt.20033 . Xiao Y, Cong M, Li J, et al. Cathepsin C promotes breast cancer lung metastasis by modulating neutrophil infiltration and neutrophil extracellular trap formation. Cancer cell[J]. 2021;39(3):423–37. e7.http://dx.doi.org/10.1016/j.ccell.2020.12.012 . Xu-Monette ZY, Xiao M, Au Q, et al. Immune Profiling and Quantitative Analysis Decipher the Clinical Role of Immune-Checkpoint Expression in the Tumor Immune Microenvironment of DLBCL. Cancer Immunol research[J]. 2019;7(4):644–57. http://dx.doi.org/10.1158/2326-6066.Cir-18-0439 . Ye X, Wang L, Nie M, et al. A single-cell atlas of diffuse large B cell lymphoma. Cell reports[J]. 2022;39(3):110713. http://dx.doi.org/10.1016/j.celrep.2022.110713 . Additional Declarations No competing interests reported. Supplementary Files SupplementaryFigureLegendsandTables.docx Cite Share Download PDF Status: Published Journal Publication published 20 Aug, 2025 Read the published version in BMC Immunology → Version 1 posted Editorial decision: Revision requested 05 May, 2025 Reviews received at journal 27 Apr, 2025 Reviews received at journal 15 Apr, 2025 Reviewers agreed at journal 14 Apr, 2025 Reviewers agreed at journal 04 Apr, 2025 Reviewers invited by journal 04 Apr, 2025 Submission checks completed at journal 02 Apr, 2025 First submitted to journal 30 Mar, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5876864","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":438486023,"identity":"1b24afa8-d27c-4b0a-961f-842415b78d00","order_by":0,"name":"Shifen Wang","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Shifen","middleName":"","lastName":"Wang","suffix":""},{"id":438486026,"identity":"47f8092c-7c97-4436-9b44-c484afa1d419","order_by":1,"name":"Hong Tao","email":"","orcid":"","institution":"Suzhou Vocational Health College","correspondingAuthor":false,"prefix":"","firstName":"Hong","middleName":"","lastName":"Tao","suffix":""},{"id":438486028,"identity":"a82dd58e-d794-4434-9442-f704c565b46b","order_by":2,"name":"Xingyun Zhao","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Xingyun","middleName":"","lastName":"Zhao","suffix":""},{"id":438486029,"identity":"b2643dad-5e76-4c5b-8d24-19eb3c76d6ba","order_by":3,"name":"Siwen Wu","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Siwen","middleName":"","lastName":"Wu","suffix":""},{"id":438486030,"identity":"4c49ef4e-1b25-47ea-8cf6-e569b36257fc","order_by":4,"name":"Chunwei Yang","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Chunwei","middleName":"","lastName":"Yang","suffix":""},{"id":438486031,"identity":"a41e922a-d737-4cc5-bb9b-f5e5af5ffbfd","order_by":5,"name":"Yuanfei Shi","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Yuanfei","middleName":"","lastName":"Shi","suffix":""},{"id":438486032,"identity":"bb30b439-07ae-431f-b106-6a9041bdfed9","order_by":6,"name":"Zhenshu Xu","email":"","orcid":"","institution":"Fujian Medical University Union Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zhenshu","middleName":"","lastName":"Xu","suffix":""},{"id":438486035,"identity":"a49e9de6-a543-40a0-bd45-7e203f934ec3","order_by":7,"name":"Dawei Cui","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1ElEQVRIie3RMQuCQBTA8ReBLoeuJ9R3uBAicOir3BHoks0NDgeBq9/G1gvBlgtXR6e2qGhpik5bmg7dgu4Pj3vD+00HYDL9YkINBUBu98B4APH4INJGPksP4hzlrGmSYuJX4oxhGzBun4SWeHLtE1oWaC5EiEFGjKMN1RIi4hwzrsiBh3iUFoxjRPSkuuyfLfF3oMirD6njHFpCrJbwHsSrrw9MywhhCauFWvwUrfXEqUJ2fybB0s0kq29JMM1sqSdfIdp9ptX3XmWLAccmk8n0T70BamlG7XgYD3UAAAAASUVORK5CYII=","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Dawei","middleName":"","lastName":"Cui","suffix":""}],"badges":[],"createdAt":"2025-01-22 02:53:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5876864/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5876864/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12865-025-00738-z","type":"published","date":"2025-08-20T16:13:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":80063396,"identity":"53e2b501-7c4c-4021-a963-de9a6f3fe4cc","added_by":"auto","created_at":"2025-04-07 12:43:43","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":308751,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFlow chart of this study design. \u003c/strong\u003eDEGs, differentially expressed genes; WGCNA, weighted gene co-expression network analysis; GO: gene ontology; GSEA, gene set enrichment analysis; KEGG: kyoto encyclopedia of genes and genomes; ROC, receiver operating characteristic curve; KM survival analysis, Kaplan-Meier survival analysis; PBMCs, peripheral blood mononuclear cells; IHC, immunohistochemistry.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/6ac4e8fad002421e36bc88bf.png"},{"id":80063400,"identity":"d2818bba-babd-473f-85e1-53db04b6a99e","added_by":"auto","created_at":"2025-04-07 12:43:43","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":431405,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDLBCL data processing and DEGs identification. \u003c/strong\u003eA. The volcano plot shows all DEGs between DLBCL and controls in the GSE25638. B. The volcano plot shows all DEGs between DLBCL and controls in the GSE12195. C. The volcano plot shows all DEGs between DLBCL and controls in the GSE12453. Significantly upregulated or downregulated DEGs were marked in orange and lake blue, respectively, gray dots represent nonsignificant genes. D. Heat map of DEGs in DLBCL (dataset GSE25638). Blue refers to control and yellow to DLBCL. E. Venn diagram of genes shared by up-regulated DEGs in the three DLBCLS, where WGCNA module genes were used for the GSE25638 dataset. A total of 95 genes were overlapped.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/25a93b2482310f7f9aaf9bed.png"},{"id":80063404,"identity":"95f89aab-a18b-4f6b-ad59-98f6c39e2083","added_by":"auto","created_at":"2025-04-07 12:43:44","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":182689,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of modules linked to clinical features of DLBCL by WGCNA. \u003c/strong\u003eA. PCA score plots of control (green) and DLBCL (red) samples in GSE25638. B. Cluster dendrogram of DLBCL highly connected genes in key modules. C. Determination of soft-threshold power for DLBCL. D. Heatmap of module–trait relationships in DLBCL. Digits in the boxes were the correlations (up) and corresponding \u003cem\u003eP\u003c/em\u003e-value (down).\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/385f0883f50c5c2308a7ec29.png"},{"id":80063408,"identity":"90f9db4c-84e8-4a34-a9a2-5c1c7fdf7c0b","added_by":"auto","created_at":"2025-04-07 12:43:44","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":671154,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFunctional analysis of DEGs of DLBCL related DEGs. \u003c/strong\u003eA. GO annotation analysis of the 95 shared DEGs in DLBCL. B-C. KEGG enrichment analysis of the 95 shared DEGs in DLBCL. D. GSEA of the top 5 enriched pathways in the 95 shared DEGs of DLBCL. E. GSEA of the tail 5 enriched pathways in the 95 shared DEGs of DLBCL. F. The protein-protein interaction (PPI) network of 95 DLBCL-related DEGs was visualized via Cytoscape. The remaining 27 DEGs were eliminated due to lack of interaction.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/0503d009e9bb6d93586df1fc.png"},{"id":80063405,"identity":"4de5d8f1-b92f-4263-9692-e02b4c7c25c2","added_by":"auto","created_at":"2025-04-07 12:43:44","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":140286,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of candidate 10 biomarkers by machine learning. \u003c/strong\u003eA. dataset GEO25638, B. dataset GEO12195, C. dataset GEO12453 and ranked based on the importance score of the random forest algorithm. D. The random forest algorithm was applied to the common genes of three DLBCL datasets to obtain four hub genes of DLBCL.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/087a146d18777dfbdcbf4548.png"},{"id":80063403,"identity":"03a6fa84-b08e-4887-b0c2-8bf27854b03d","added_by":"auto","created_at":"2025-04-07 12:43:43","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":230587,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSurvival impact of four key genes in DLBCL patients. \u003c/strong\u003eA. The survival probability for expression levels of four hub genes (CXCL9, CCL18, C1QA, and CTSC) in DLBCL patients were examined using Kaplan-Meier analysis. Red curves indicate high gene expression and blue curves indicate low gene expression. B. The survival probability for expression levels of four hub genes (CXCL9, CCL18, C1QA, and CTSC) in two subtypes of DLBCL patients were examined using Kaplan-Meier analysis. Red/green curves indicate high gene expression and blue/purple curves indicate low gene expression.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/d94b6561140fb9e82fc8d86d.png"},{"id":80063421,"identity":"942d25bb-f02e-4d4b-a4eb-c6556560ca71","added_by":"auto","created_at":"2025-04-07 12:43:44","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":457264,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAlterations in Immune Cell Infiltration Landscape in DLBCL patients.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA. The bar plot displays the percentage of 28 immune cells in health control and DLBCL samples. Green refers to control and red to DLBCL.B. The boxplot compares the expression of immune cells between DLBCL and health controls, which were used the wilicox test. Green refers to health controls and red to DLBCL. C. Correlation analysis of 28 immune cell infiltrations with four hub DEGs, which were calculated using Spearman method. *P \u0026lt; 0.05; **P \u0026lt; 0.01; ***P \u0026lt; 0.001; ****P \u0026lt; 0.0001.\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/22c7632eb48052fdb59a4f83.png"},{"id":80064753,"identity":"a801515b-8086-4c6c-8e20-3068a6d2dcb0","added_by":"auto","created_at":"2025-04-07 12:59:44","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":666671,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe relative expressions of optimal key genes were validated by RT-PCR and IHC. \u003c/strong\u003eA-D. The expressions of CXCL9 (A), CCL18 (B), CTSC (C) and C1QA (D) between DLBCL patients and healthy control (HC). E. Representative images of immunohistochemical staining of lymph node tissue from a control healthy group (left) and DLBCL (right). *P \u0026lt; 0.05; **P \u0026lt; 0.01; ***P \u0026lt; 0.001; ****P \u0026lt; 0.0001.\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/d87fb0267ad3c501e93843f4.png"},{"id":89847139,"identity":"bdbba965-cdb6-4af8-8c54-a2ea7a57b5f8","added_by":"auto","created_at":"2025-08-25 16:41:03","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4571540,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/7fed337d-7e2b-4a2f-9abb-a6addd65c72e.pdf"},{"id":80064141,"identity":"f48e5e16-0582-44ce-beeb-9dfdc02790ec","added_by":"auto","created_at":"2025-04-07 12:51:43","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":264320,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigureLegendsandTables.docx","url":"https://assets-eu.researchsquare.com/files/rs-5876864/v1/f5419b7a39384773899728aa.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Exploration of biomarkers for predicting the prognosis of patients with diffuse large B-cell lymphoma by machine-learning analysis","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eDiffuse large B cell lymphoma (DLBCL), the most common subtype of non-Hodgkin lymphoma (NHL), is an aggressive cancer that accounts for 30%-40% of NHL[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Gene expression profiling (GEP) analysis by microarray analysis revealed three molecular subtypes: germinal center B-cell like (GCB), activated B-cell like (ABC), and unclassifiable[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. These 3 DLBCL subtypes have different clinical and biological heterogeneity with varying responses to treatment and prognosis[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. However, limited funding and limited availability of GEP technology make it impractical to generalize to clinical DLBCL patients. The Hans algorithm separates diffuse large B cell lymphoma (DLBCL) into two subtypes: Germinal center B-cell like (GCB) and non-GCB[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], just as with the GEP classification. Rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) has been shown to have a poor prognosis for non-GCB patients[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Genetic and functional genomic studies have revealed oncogenic driver pathways in DLBCL[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], leading to the exploration of potential targets for the treatment of DLBCL[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Understanding the immunological components of DLBCL has grown more crucial as novel immune-targeting treatments have been created. The pathogenesis of DLBCL involves a complex interaction between tumor cells and tumor microenvironment (TME), and immune cells shape the tumor microenvironment of DLBCL[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. There are still many unanswered questions regarding the immunology of DLBCL, such as the immune system and the underlying mechanisms of DLBCL development.\u003c/p\u003e \u003cp\u003eThis study aimed to analyze the prognostic role and immune infiltration of potential markers of DLBCL in DLBCL by bioinformatics analysis, and to elucidate the possible mechanisms. Our findings suggest that CXCL9, CCL18, C1QA, and CTSC are prognostic biomarkers and poor prognostic indicators in DLBCL patients. In summary, our study provides evidence that four key genes (CXCL9, CCL18, C1QA, CTSC) are important in the development and progression of DLBCL and suggests that these genes may be novel biomarkers and new therapeutic targets for DLBCL.\u003c/p\u003e \u003cp\u003eThe potential prognostic biomarkers of DLBCL were screened by bioinformatics analysis. We explored the value of four key genes in DLBCL and their molecular mechanisms, providing a basis for future immunotherapy and precision medicine. The predictive performance was assessed using survival analysis and receiver operating characteristic curve analysis. To further understand the role of TME in the pathogenesis of DLBCL patients, we analyzed the relationship between immune cell subtypes and clinical prognosis in the DLBCL cohort. In addition, we examined the correlation between four key genes and the immune process in DLBCL.\u003c/p\u003e"},{"header":"2. Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Microarray data\u003c/h2\u003e \u003cp\u003eThe study flowchart is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Three microarray datasets (GSE25638, GSE12195, GSE12453) were retrieved from the NCBI Gene Expression Omnibus (GEO; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) database. Dataset GSE25638 included data on 26 DLBCL patients and 13 controls. Dataset GSE12195 included data on 73 DLBCL patients and 10 controls. Dataset GSE12453 included data on 11 DLBCL patients and 25 controls.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Data processing and differentially expressed genes (DEGs) identification\u003c/h2\u003e \u003cp\u003eDEGs were calculated between disease and control groups with it via the \u0026ldquo;Limma\u0026rdquo; package in R software. For DLBCL, adjusted P value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and |log2 FC|\u0026gt;1 were used to identify the DEGs. Next, the difference analysis results for each group were presented using the heatmap and volcano plot. In the three plots, blue indicated low expression, and yellow indicated high.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Functional enrichment analysis and protein-protein interaction (PPI) network\u003c/h2\u003e \u003cp\u003eFor identifying the biological functions and signaling pathways of the shared genes, the R package (R4.3.1) \"clusterProfiler\" performed an enrichment analysis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) pathways. The shared genes were entered into the website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cn.string-db.org/\u003c/span\u003e\u003cspan address=\"https://cn.string-db.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) in order to create the protein interaction relationship data. The data were then imported into the Cytoscape software for computation, prediction, and visualization.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Significant module identification via WGCNA\u003c/h2\u003e \u003cp\u003eWeighted gene co-expression network analysis (WGCNA) is one of the most popular algorithms for computing large volumes of data, which evaluates gene expression correlations and visualizes co expression networks through clustering and modularization. Data were checked to identify the outliers in the samples by hierarchical clustering. There were no offending samples in the DLBCL dataset (GSE25638) and no samples were removed. The \u0026ldquo;pickSoftThreshold\u0026rdquo; package function was utilized to screen out the optimal soft threshold. Through a topological overlap matrix (TOM) analysis, the adjacency matrix was clustered and categorized genes with similar expression patterns into gene modules, that is, the module eigengenes (MEs). In addition, the correlations between MEs and clinic traits were calculated via Pearson\u0026rsquo;s correlation test and the strongest positive correlation was selected for further analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Machine learning algorithm\u003c/h2\u003e \u003cp\u003ePrevious reviews reported that the prediction model of random forest showed better diagnostic efficiency than other models[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e], so the random forest algorithm was selected for variable screening in this study. These three DLBCL datasets were subjected to the random forest machine learning algorithm in order to further uncover potential biomarkers. The most significant factors were determined via random forest analysis using a decision tree method[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Random forest (RF) model is developed from decision trees. RF can generate hundreds or even thousands of trees. The samples of each tree are extracted from the bags in a defined set by the Bootstrap method as the training samples, and the remaining data, called Out-Of-Bag data (OOB), is used as the test samples. The principle of screening variables is mainly to rank the importance of features according to the error generated before and after rearranging features. Based on the OOB test set, the prediction accuracy of the original features and the randomly arranged features after permutation was compared to measure the feature importance. We were able to identify DLBCL characteristic genes by filtering the shared genes using this approach. In order to determine the optimal number of trees, a random forest model with 500 trees on the discovery queue is established, and cross-validation error is used. Next, the top 10 genes were sequenced according to their significance, and their distributions were plotted. The final outcome was determined for each DLBCL dataset by setting the significance threshold at 0.5. By using this algorithm, key genes were selected from three DLBCL datasets, and we took the intersection of the algorithm results from each dataset.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Nomogram construction and receiver operating characteristic (ROC) evaluation\u003c/h2\u003e \u003cp\u003eThe nomogram was constructed based on the four hub genes by using the \u0026ldquo;rms\u0026rdquo; package. Each gene\u0026rsquo;s relative expression level corresponds to a score based on the nomogram. To evaluate the predictive value of four candidate biomarker, we generated ROC curves and calculated the area under the curve (AUC) and 95% confidence interval (CI). ROC curve was performed to determine whether the nomogram-based decision was conducive to DLBCL diagnosis. The bootstrap algorithm was used to validate the results on the external dataset GSE83632. The optimal AUC for predicting the risk of DLBCL was \u0026gt;\u0026thinsp;0.7.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Survival Analysis\u003c/h2\u003e \u003cp\u003eLog-rank tests and Kaplan-Meier (KM) graphs were used to assess how the groups' overall survival differed from one another. These were plotted and analyzed in the survminer package and the survival package in R. Each survival curve's cut-offs for patient classification are given in the figure legends. Using the survival package in R, multivariate Cox regression was carried out while taking the patient's sex, tumor stage, and grade into consideration.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Immune cell infiltration\u003c/h2\u003e \u003cp\u003eFor RNA samples from different tissue types, immune cell subpopulation infiltration scores in the tumor microenvironment are computed using the single-sample gene set enrichment analysis (ssGSEA) algorithm[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. The algorithm \"ssGSEA\" can converted the normalized gene expression matrix into gene enrichment score matrix. The analysis's 28 immune cells' specific gene maps came from a recent publication (\u003cb\u003eSupplementary Table\u0026nbsp;2\u003c/b\u003e) [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. The proportion of each immune cell in each sample was visualized from the barplot. A boxplot was used to compare the expression of the differences for each immune cell between the two groups. A heatmap displaying the correlation of different immune cells in DLBCL pathogenesis was constructed using the R package \"corrplot\". The correlation between characteristic genes and immune cell infiltration is also examined using the ssGSEA algorithm and constructed using the R package \"ggcorrplot\".\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Isolation of human PBMCs and RT-PCR\u003c/h2\u003e \u003cp\u003eA total of 13 samples were collected from the First Affiliated Hospital, Zhejiang University School of Medicine, including 13 DLBCL samples and 5 healthy controls. Using EDTA anticoagulant blood and FicollPaque density gradient centrifugation, human peripheral blood mononuclear cells (PBMCs) were separated. TRIzol reagent was used to extract total RNA from the PBMCs of DLBCL patients and healthy controls, and the Prime ScriptTM RT Reagent Kit (TaKaRa, China) was used to create cDNA through reverse transcription. The internal reference was β-Actin. The SYBR Green PCR Kit (TaKaRa, China) was used for quantitative realtime polymerase chain reaction (RT-PCR) in accordance with the manufacturer's instructions. Furthermore, the relative expression was calculated by the 2\u003csup\u003e\u0026minus;△△CT\u003c/sup\u003e method. All reactions were repeated in triplicate. The primer sequences of target genes are displayed in Supplementary Table\u0026nbsp;1, which are from public website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.origene.com/\u003c/span\u003e\u003cspan address=\"https://www.origene.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Immunohistochemical analysis\u003c/h2\u003e \u003cp\u003eThe lymph node tissues of DLBCL patients and healthy controls were collected from the Department of Pathology, the First Affiliated Hospital of Zhejiang University School of Medicine. To construct the tissue microarray (TMA), a 3-mm-diameter core of the cancer tissue area was generated. The brief steps for immunohistochemistry (IHC) were as follows: First, slides made from the TMA were deparaffinized in xylene and rehydrated in ethanol, and the antigens were extracted for IHC tests with specific antibodies. Second, the slides were blocked with goat serum for 1 h, and after washing, the slides were further stained with a diluted anti-human CXCL9 antibody (Abcam, ab290643). Then, 3,3\u0026prime;-diaminobenzidine (DAB) was used to stain the tissues after 24 h. The IHC results of CXCL9 staining scores were semiquantified by an immunoreactive scoring (IRS) system, which has been previously described[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Statistical analysis\u003c/h2\u003e \u003cp\u003eR software and corresponding packages were used for calculations and analysis. We used Wilcoxon rank sum test and Kruskal-Wallis test to determine differences between two or more groups. Spearman\u0026rsquo;s method used to analyze the correlation of variables. ROC curve analysis was used to determine the diagnostic value of biomarkers. Survival curves were plotted \u003cem\u003evia\u003c/em\u003e the Kaplan-Meier log rank test. In the prognostic assessment analysis, a two-sided P value of \u0026lt;\u0026thinsp;0.05 was considered statistically significant. Of note, in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e7\u003c/span\u003eB, as the data distribution did not conform to the normal distribution, we used the wilicox test to compare the difference in the abundance of immune cells between the tumor group and the control group.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Screening of DEGs from GSE25638、GSE12195 and GSE12453\u003c/h2\u003e \u003cp\u003eUsing data from the open NCBI GEO database, we obtained the Series Matrix File data file for GSE25638、GSE12195 and GSE12453 for DLBCL, respectively \u003cb\u003e(\u003c/b\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. A total of 39 sets of transcriptome data, including 13 from normal samples and 26 from DLBCL samples in GSE25638. The DEGs between the two groups of samples were identified using the limma package. A total of 2003 DEGs (including 1642 upregulated and 361 downregulated genes) were identified in DLBCL (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). In this cohort, we observed more up-regulated genes and fewer down-regulated genes in DLBCL samples compared to normal samples. A total of 36 sets of transcriptome data, including 25 from normal samples and 11 from DLBCL samples in GSE12453. A total of 411 DEGs (including 333 upregulated and 78 downregulated genes) were identified in DLBCL (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). In this cohort, we also observed more up-regulated genes and fewer down-regulated genes in DLBCL samples compared to normal samples. However, a total of 83 sets of transcriptome data, including 10 from normal samples and 73 from DLBCL samples in GSE12195. In this cohort, we observed that there was little difference between up-regulated and down-regulated genes in DLBCL samples compared to normal samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eC), which may be related to the key factors of immune microenvironment status. Then, we intersected the upregulated DEGs of the two datasets (GSE12195 and GSE12453), resulting in a total of 222 intersecting DEGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eE and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBasic information of GEO datasets used in the study.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eID\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGSE series\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDisease\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSamples\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSourcetypes\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePlatform\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGSE25638\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDLBCL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26 DLBCL patients and 13 normal donor controls\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003elymphoid tissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGPL570\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGSE12195\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDLBCL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e73 DLBCL patients and 10 normal donor controls\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003elymphoid tissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGPL570\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGSE12453\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDLBCL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e11 DLBCL patients and 25 normal donor controls\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003elymphoid tissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGPL570\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNote: GSE, Gene expression omnibus series; DLBCL, Diffuse large B cell lymphoma; GPL, Gene expression omnibus platform.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Co-expression network construction and key module identification for WGCNA\u003c/h2\u003e \u003cp\u003eTo further explore the key gene associations of DLBCL, WGCNA was performed to identify the most relevant gene modules in addition to analyzing the differential expression genes. The PCA score plot showed significant differences between the disease and control groups in DLBCL (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e3\u003c/span\u003eA), which could be used for subsequent analysis. We first selected a suitable soft threshold β value that met the scale-free conditions (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e3\u003c/span\u003eB and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). The thresholds β were set to 20 for the dataset (GSE25638) and then the tom matrix was used to identify gene modules (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). Then the TOM was converted from the adjacent matrix and turned into the dissTOM (dissTOM\u0026thinsp;=\u0026thinsp;1-TOM) for hierarchical clustering. Dynamic tree shearing clustered the similarity genes based on the topological overlap and then divided them into various modules (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). A total of 5 gene modules were found, among which the ME grey module and the ME turquoise module had the strongest positively relationship with DLBCL (1034 genes, R\u0026thinsp;=\u0026thinsp;0.95/0.9, p\u0026thinsp;=\u0026thinsp;3e-20/4e-15) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eD). On this basis, the two set of genes was compared with the 222 up-regulated intersection DEGs of GSE12195 and GSE12453 identified above, and a total of 95 intersection genes were obtained (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eE). These 95 genes were considered for our subsequent studies.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Functional enrichment analysis of the pathogenic genes involved in DLBCL\u003c/h2\u003e \u003cp\u003eTo reveal the potential pathogenic genes and underlying mechanism in DLBCL, we conducted pathway analysis on the 95 shared genes from these three datasets. The findings revealed that the DLBCL-related genes were primarily enriched in pathways such as inflammatory response, secretory granule, and collagen binding (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). These genes of DLBCL were mainly enriched in mineral absorption and pertussis (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). GO analysis showed that BP terms were significantly enriched included \"inflammatory response\", \"biological process involved in interspecies interaction between organisms\", and \"defense response\". Moreover, \"secretory granule\", \"extracellular region\", and \" extracellular space\" were enriched for CC, while the MF of DEGs were highly associated with \"collagen binding\", \"CXCR3 chemokine receptor binding \", and \"peptidase regulator binding \" (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). The results of KEGG analysis showed that the DLBCL shared genes were mainly enriched in CXCR chemokine receptor binding and chemokine activity (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). The top five pathways identified by gene set enrichment analysis (GSEA) were \u0026ldquo;Circadian rhythm\u0026rdquo;, \u0026ldquo;Drug metabolism-cytochrome P450\u0026rdquo;, \u0026ldquo;Phenylalanine metabolism\u0026rdquo;, \u0026ldquo;Taurine and hypotaurine metabolism\u0026rdquo; and \u0026ldquo;Tyrosine metabolism\u0026rdquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). The tail five pathways were \u0026ldquo;Biosynthesis of amino acids\u0026rdquo;, \u0026ldquo;Cell cycle\u0026rdquo;, \u0026ldquo;DNA replication\u0026rdquo;, \u0026ldquo;Primary immunodeficiency\u0026rdquo; and \u0026ldquo;Proteasome\u0026rdquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). Meanwhile, GSEA analysis also showed that these genes were mainly enriched in CXCR chemokine receptor binding and chemokine activity (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eE).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThen, the PPI showing the shared genes in the DLBCL dataset (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eF). Based on the 95 shared genes, a PPI network was preliminarily constructed to select hub genes for DLBCL. After eliminating DEGs with poor interaction, 68 genes were retained (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003eF).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Identify potential shared diagnostic genes based on machine learning algorithms\u003c/h2\u003e \u003cp\u003eTo further explore the diagnostic value of DEGs in DLBCL, we use the machine learning algorithms. Based on the above 95 shared genes mentioned above, we employed learning approaches for the three lymphoma datasets to confirm variables significantly associated DLBCL. Gene importance was calculated by random forest for DEGs. The 10 most important genes from the random forest results for the three lymphoma datasets are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-C. Then, a subset of 4 overlapping hub genes (CXCL9, CCL18, C1QA, CTSC) were selected from the random forest machine learning algorithms for the three lymphoma datasets (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e5\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Diagnostic value and prognostic correlation of DLBCL hub genes\u003c/h2\u003e \u003cp\u003eCompared with the control, the predictive nomogram was constructed, whereby each gene\u0026rsquo;s relative expression corresponded to a score, and the total score was obtained by the summation of the score of each gene (Supplementary Fig.\u0026nbsp;1A). ROC curve analysis showed good predictive performance of each gene as follows (Supplementary Fig.\u0026nbsp;1B): CXCL9 (AUC: 1.000); CCL18 (AUC:1.000); C1QA (AUC: 1.000); CTSC (AUC: 1.000). ROC analysis of the nomogram yielded an AUC of 1.000, demonstrating a high predictive value for DLBCL. Then, in order to determine the performance of the screened markers, we performed 1000 bootstrap validation on the external dataset GSE83632, and the average AUC under the 1000 ROC curves was 0.839 (95%CI: 0.774\u0026ndash;0.904) (supplementary Fig.\u0026nbsp;2), which further proved the predictive efficacy of these four candidate genes. Moreover, we analyzed the prognostic predictive value of 4 hub genes in the two subgroups of DLBCL. Kaplan\u0026ndash;Meier curves showed that high expression groups of CXCL9, CCL18, C1QA and CTSC had significantly shorter overall survival than their low expression groups, there was statistical difference (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). Kaplan-Meier survival analysis of the two subtypes of DLBCL indicated that, overall, the GCB subgroup exhibited a prominent advantage of median survival time, whereas the non-GCB subgroup presented with a worse prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e6\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Immune infiltration analysis of shared diagnostic genes\u003c/h2\u003e \u003cp\u003eConsidering that DLBCL are characterized by a high immune response. The abundances of immune cells in different groups were analyzed with ssGSEA. In each group, the proportion of 28 immune cells was shown as a bar plot (\u003cb\u003eSupplementary Table\u0026nbsp;2\u003c/b\u003e). Generally, the barplot (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e7\u003c/span\u003eA) clearly illustrated that the proportion of activated CD8\u003csup\u003e+\u003c/sup\u003e T cells, Myeloid-derived suppressor cells (MDSCs), activated CD8\u003csup\u003e+\u003c/sup\u003e T cells, central memory CD4\u003csup\u003e+\u003c/sup\u003e T cells, and effector memory CD8\u003csup\u003e+\u003c/sup\u003e T cells were higher in DLBCL. Compared with the control samples, the activated B cell, eosinophils, immature B cells, and neutrophils were increased in the DLBCL samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e7\u003c/span\u003eB). In contrast, B cells and granulocytes correlated negatively. Moreover, the relationship between biomarkers and immune cell contents was investigated. In DLBCL samples, NK cells and T cells were significantly positively correlated with 4 hub genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC). CXCL9 was significantly positively correlated with CD56\u003csup\u003ebright\u003c/sup\u003e NK cells and activated CD4\u003csup\u003e+\u003c/sup\u003e T cells, negative for macrophages immature B cells, neutrophil and eosinophil (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC). Moreover, T cell subsets are also involved in immune regulation of four key genes, such as regulatory T cells, T follicular helper cells and Th2 cells. It appears that immune function is crucial to the development of DLBCL (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e7\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Validation of the four optimal key genes\u003c/h2\u003e \u003cp\u003eThe relative expression of four optimal key genes was verified with RT-PCR in DLBCL patients and healthy controls. Compared with the control group, the expression of CXCL9 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA), CCL18 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB), CTSC (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC) and C1QA (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eD) was significantly upregulated in DLBCL patients (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), which was consistent with the results of this bioinformatic analysis. Among them, CXCL9 was the most significantly differentially expressed among the four genes(P\u0026thinsp;\u0026lt;\u0026thinsp;0.01) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA). Immunohistochemical (IHC) analysis revealed that, compared with that in the lymph node tissues of the healthy group, CXCL9 protein was highly expressed in DLBCL tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eE).\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe Limma R package provides a comprehensive solution for analyzing gene expression data [58]. In the present study, 222 DEGs were identified between DLBCL and control groups. WGCNA is a systematic biological technique for analyzing gene association patterns among samples. The most relevant correlation between gene modules and phenotypes can be calculated by building a gene co-expression network and determining related gene clusters. Herein, we found 1034 significant module genes of DLBCL and applied Limma analysis and WGCNA to screen the gene clusters of connected, shared, and correlating genes in DLBCL. The intersection of DEGs regarding DLBCL (n\u0026thinsp;=\u0026thinsp;222) and module genes regarding DLBCL (n\u0026thinsp;=\u0026thinsp;1034) yielded 95 common risk genes associated with DLBCL.\u003c/p\u003e \u003cp\u003eTo elucidate the roles of these 95 DEGs in DLBCL, KEGG pathway and GO analysis were performed. These DEGs were mainly enriched in KEGG pathways \"Mineral absorption\u0026rdquo;, \u0026ldquo;Pertussis\u0026rdquo;, \u0026ldquo;Lysosome\u0026rdquo;, \u0026ldquo;Complement and coagulation cascades\u0026rdquo; and \u0026ldquo;Fluid shear stress and atherosclerosis\u0026rdquo;, suggesting a correlation with immunoinflammatory responses. Mineral absorption was confirmed to be one of the most significantly enriched pathways in the proteome and phosphoproteome, which is involved in the process of mammalian immunosuppression[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Pertussis and lysosome have also been associated with immune responses dysfunction[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. A previous study has confirmed that complement and coagulation cascades are one of the pathophysiological pathways in the pathogenesis of acquired immunodeficiency syndrome-related non-Hodgkin lymphoma[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Recent studies have further shown that the complement and coagulation cascade pathways play a key role in the anti-tumor process of 10-hydroxy-2-decenoic acid (10-HAD), which provides a potential therapeutic approach for lymphoma treatment[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Significantly enriched GO terms associated with BP included \"inflammatory response\", \"immune response\", \"defense response\", \u0026ldquo;immune system process\u0026rdquo;, which are well-established integral components of DLBCL pathogenesis[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Moreover, the CC and MF analysis of DEGs were closely related to inflammatory response and innate and adaptive immune response. These shared DEGs were primarily enriched in CXCR chemokine receptor binding and chemokine activity, according to GSEA analysis. Taken together, these findings demonstrate that these 95 DEGs were correlated with DLBCL pathogenesis.\u003c/p\u003e \u003cp\u003e95 DEGs in DLBCL were mapped to the PPI network to further explore their potential interplay, and 68 DEGs were identified. Using the machine learning method (random forest), four hub DEGs (CXCL9, CCL18, C1QA, CTSC) were identified. We constructed a nomogram and evaluated its predictive value for DLBCL in DLBCL patients. First, we identified four key immune-associated genes (CXCL9, CCL18, C1QA, CTSC) for constructing our nomogram. Next, we conducted validation which showed that the four hub genes had a good predictive value of DLBCL patients, indicating their potential roles during the DLBCL progression. The expression levels of four hub genes were significantly correlated with the prognosis of DLBCL. We further investigated the role of four hub genes in the survival of patients with DLBCL. Kaplan-Meier survival curve showed that the increased mRNA levels of four hub genes were significantly associated with poor prognosis of DLBCL patients (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Furthermore, we discovered a correlation between the genesis of DLBCL and the levels of the four hub genes. A major factor in the categorization of DLBCL, which is linked to the prognosis of patients and a hot issue in the pathological mechanism of DLBCL, is the cell of origin. GCB-DLBCL originate from the germinal center (GC) stage, while nonGCB-DLBCL originates from the ABC/post-GC stage, among which nonGCB-DLBCL has the worse prognosis. The potential of gene level as a prognostic marker for subtype categorization is demonstrated by our discovery that four hub genes' expression levels may further separate GCB-DLBCL/nonGCB-DLBCL into subgroups with low expression with favorable prognosis and high expression with bad prognosis.\u003c/p\u003e \u003cp\u003eCurrent evidence suggests that the CXCL9 is produced by lymphoma cells and can elicit cytotoxic responses via CXCR3, which promotes tissue necrosis and vascular damage by recruiting cytotoxic lymphocytes[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Our findings are in accordance with earlier research that shown CXCL9 is significantly expressed in DLBCL tissues and cell lines and is strongly associated with patient survival and clinical progression[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Notably, Epstein-Barr virus (EBV)-negative DLBCL does not secrete chemokines such as CXCL9 in contrast to EBV-positive DLBCL patients [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], and we speculate that CXCL9 promotes DLBCL progression and may be involved in the immune-modulatory mechanism of DLBCL, which is consistent with the results of our KM curve analysis. Another study reported that in the pathological mechanism of combination therapy of IL15 receptor agonists with CART in relapsed/refractory B malignancies, an increase in chemokines including CXCL9 drives the trafficking of lymphocytes to tissues, which is associated with a high durable response rate[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. However, little is known about CXCL9's actions and underlying mechanism in DLBCL. There is controversy regarding the association between CCL18 and the occurrence and severity of DLBCL[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. It is generally believed that the biological functions of CCL18 and its co-expressed genes are prominent in cell migration, proliferation and apoptosis[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Tumor-associated macrophages secrete many chemokines into the tumor microenvironment, especially a large amount of CCL18, which is one of the important mechanisms of the immunosuppressive nature of the tumor microenvironment and cancer immune evasion[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Investigating the precise regulation mechanism of CCL18 in DLBCL is crucial since chemokines have anti-cancer and pro-cancer effects in the tumor microenvironment. The C1QA gene is known to encode the A-chain polypeptide of serum complement subcomponent C1q. C1QA is empirically associated with the outcome of rituximab treatment of DLBCL. Jin et al. found that DLBCL patients with homozygous A -allele carriers of the gene \u003cem\u003eC1qA\u003c/em\u003e[276] had better overall response and higher complete response, and overall survival in receiving R-CHOP[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. Of note, C1qA may originate from other non-tumor cells in DLBCL tissue. Blood monocytes and tumor-associated macrophages also expressed \u003cem\u003eC1qA\u003c/em\u003e and may affect tumor cells through it. More research is needed to determine whether exogenous C1qA protein level is affected by Rituximab. C1QA is derived from rituximab resistant DLBCL cells under rituximab stress and is regulated by METTL3 and YTHDF2-mediated m6A methylation[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Notably, C1QA may originate from other non-tumor cells in DLBCL tissues such as blood monocytes and tumor-associated macrophages[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. However, it is unclear whether rituximab affects the levels of exogenous C1QA protein. A lysosomal cysteine protease called cathepsin C (CTSC), commonly referred to as dipeptidylpeptidase I, is necessary for the catalytic activation of several serine proteases[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. In a conjoined gene study of non-Hodgkin B-cell lymphoma cell line KPU-UH1, novel conjoined genes, including CTSC-RAB38, were detected and may be genetic biomarkers for lymphoma[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. Recent studies have found that tumor-secreted protease cathepsin C (CTSC) promotes tumour metastasis by regulating recruitment of neutrophils and formation of neutrophil extracellular traps (NETs)[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. Validation by RT-PCR, hub gene expression in DLBCL was statistically significant when compared to healthy control, consistent with previous analyses. IHC tests further demonstrated that DLBCL had high levels of CXCL9 protein expression.\u003c/p\u003e \u003cp\u003eThe proportion immune cells have a significant impact on DLBCL. A high percentage of PD1\u003csup\u003e+\u003c/sup\u003e CD8\u003csup\u003e+\u003c/sup\u003e and PD-L1\u003csup\u003e+\u003c/sup\u003e T cells in the TME was reported to predict poor survival in DLBCL in one study, but T cells with high expression of the immunological checkpoint cytotoxic T-lymphocyte-associated protein 4 (CTLA4) had a positive prognosis[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. A significant percentage of DLBCL in our investigation was attributed to activated T cells, NK cells, and macrophages, which were consistent with previous single-cell sequencing results[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. Immunosuppressive cells were also found to be high in DLBCL in our study, such as Treg cells, MDSCs, and macrophages. It is speculated that immune-promoting and immune-suppressive mechanisms play a key role in the pathogenesis of DLBCL. Then, we examined the relationship between the four identified genes (CXCL9, CCL18, C1QA, CTSC) and immunity. The four hub genes positively correlated with most immune-related cells.\u003c/p\u003e \u003cp\u003eHowever, there are many limitations associated with the current manuscript. First, this study relies on bioinformatics analysis on public datasets, which could generate some inconsistencies in the results because of its retrospective nature. Moreover, due to the sample size limitation of our study, the accuracy of the results needs to be further verified in larger independent cohorts. Furthermore, additional molecular experiments are required to investigate the underlying mechanisms.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eIn summary, through bioinformatics analysis such as DEGs and WGCNA, machine learning algorithms, and GO and KEGG enrichment analysis, four co-hub genes (CXCL9, CCL18, C1QA, CTSC) were screened out by the system. We developed a nomogram model to predict the risk of DLBCL and identified four key genes with prognostic value, which laid the foundation for further study of potential key candidate genes in DLBCL patients. In addition, we also found significant differences in immune cell abundance in DLBCL patients. The possible diagnostic or therapeutic value of these immune cells and genes in DLBCL warrants further investigation.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDLBCL\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDiffuse large B cell lymphoma\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGEO\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eGene Expression Omnibus\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDEGs\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDifferentially expressed genes\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eWGCNA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eWeighted gene co-expression network analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePPI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eProtein-protein interaction\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRT-qPCR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ereal time quantitative PCR\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIHC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eImmunohistochemistry\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eKEGG\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eKyoto Encyclopedia of Genes and Genomes\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGO\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eGene Ontology\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eNHL\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eNon-Hodgkin lymphoma\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGEP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eGene expression profiling\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGCB\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003egerminal center B-cell like\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eABC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eactivated B-cell like\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCHOP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eRituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTME\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eTumor microenvironment\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTOM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eTopological overlap matrix\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMes\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emodule eigengenes\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eROC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eReceiver operating characteristic curves\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAUC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eArea under the curve\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003econfidence interval\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eKM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eKaplan-Meier\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003essGSEA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003esingle-sample gene set enrichment analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePBMCs\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePeripheral blood mononuclear cells\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRT-PCR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eReal-time polymerase chain reaction\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTMA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eTissue microarray\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGSEA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eGene set enrichment analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank the contributions of public databases such as the GEO to human medicine. Moreover, we sincerely appreciate the highly qualified native English speaking editors at American Journal Experts (AJE) for providing reputable English language editing service for our manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u003c/strong\u003e\u003cstrong\u003e\u0026rsquo;\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDC and SW designed the study. SW, XZ, SWu, CY and DC performed the data analysis. SWu, CY and DC participated in the collection and detection of clinical samples from patients and healthy controls. SW and DC prepared the figures and manuscript. In addition, HT, YS, ZX and DC critically revised the manuscript. All authors contributed to the manuscript and approved the submitted version. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Key Research and Development Plan of Department of Science and Technology of Zhejiang Province (2024C03156), National Natural Science Foundation of China (81871709, 82270175, 82300213), Natural Science Foundation of Fujian Province of China (2021J02040), National Key Clinical Specialty Discipline Construction Program (2021-76), Fujian Provincial Clinical Research Center for Hematological Malignancies (2020Y2006).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData is provided within the manuscript or supplementary information files.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis project was conducted in accordance with the guidelines of the Helsinki Declaration, and approved by the Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine. All involved participants provided written informed consent.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Trial Number\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNot applicable.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAlaggio R, Amador C, Anagnostopoulos I, The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours. Lymphoid Neoplasms, Leukemia[J], : 2022;36(7):1720\u0026ndash;1748.\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1038/s41375-022-01620-2\u003c/span\u003e\u003cspan address=\"10.1038/s41375-022-01620-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCampo E, Jaffe ES, Cook JR, et al. The International Consensus Classification of Mature Lymphoid Neoplasms: a report from the Clinical Advisory Committee. Blood[J]. 2022;140(11):1229\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1182/blood.2022015851\u003c/span\u003e\u003cspan address=\"10.1182/blood.2022015851\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature[J]. 2000;403(6769):503\u0026ndash;11. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1038/35000501\u003c/span\u003e\u003cspan address=\"10.1038/35000501\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRosenwald A, Wright G, Leroy K, et al. Molecular diagnosis of primary mediastinal B cell lymphoma identifies a clinically favorable subgroup of diffuse large B cell lymphoma related to Hodgkin lymphoma. J experimental medicine[J]. 2003;198(6):851\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1084/jem.20031074\u003c/span\u003e\u003cspan address=\"10.1084/jem.20031074\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLossos IS, Czerwinski DK, Alizadeh AA, et al. Prediction of survival in diffuse large-B-cell lymphoma based on the expression of six genes. New Engl J medicine[J]. 2004;350(18):1828\u0026ndash;37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1056/NEJMoa032520\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa032520\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood[J]. 2004;103(1):275\u0026ndash;82. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1182/blood-2003-05-1545\u003c/span\u003e\u003cspan address=\"10.1182/blood-2003-05-1545\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu K, Weisenburger DD, Choi WW, et al. Addition of rituximab to standard chemotherapy improves the survival of both the germinal center B-cell-like and non-germinal center B-cell-like subtypes of diffuse large B-cell lymphoma. J Clin oncology: official J Am Soc Clin Oncology[J]. 2008;26(28):4587\u0026ndash;94. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1200/jco.2007.15.9277\u003c/span\u003e\u003cspan address=\"10.1200/jco.2007.15.9277\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeyer PN, Fu K, Greiner TC, et al. Immunohistochemical methods for predicting cell of origin and survival in patients with diffuse large B-cell lymphoma treated with rituximab. J Clin oncology: official J Am Soc Clin Oncology[J]. 2011;29(2):200\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1200/jco.2010.30.0368\u003c/span\u003e\u003cspan address=\"10.1200/jco.2010.30.0368\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eReddy A, Zhang J, Davis NS, et al. Genetic and Functional Drivers of Diffuse Large B Cell Lymphoma. Cell[J]. 2017;171(2):481\u0026ndash;e494. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e15.http://dx.doi.org/10.1016/j.cell.2017.09.027\u003c/span\u003e\u003cspan address=\"15.10.1016/j.cell.2017.09.027\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmitz R, Wright GW, Huang DW et al. Genetics and Pathogenesis of Diffuse Large B-Cell Lymphoma, The New England journal of medicine[J] 2018;378(15):1396\u0026ndash;407\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1056/NEJMoa1801445\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa1801445\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMelani C, Lakhotia R, Pittaluga S et al. Combination Targeted Therapy in Relapsed Diffuse Large B-Cell Lymphoma, The New England journal of medicine[J] 2024;390(23):2143\u0026ndash;55\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1056/NEJMoa2401532\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa2401532\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Zhou X, Wang X. Targeting the tumor microenvironment in B-cell lymphoma: challenges and opportunities. J Hematol oncology[J]. 2021;14(1):125. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1186/s13045-021-01134-x\u003c/span\u003e\u003cspan address=\"10.1186/s13045-021-01134-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNickson D, Meyer C, Walasek L, et al. Prediction and diagnosis of depression using machine learning with electronic health records data: a systematic review. BMC Med Inf Decis making[J]. 2023;23(1):271. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1186/s12911-023-02341-x\u003c/span\u003e\u003cspan address=\"10.1186/s12911-023-02341-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTiruneh SA, Vu TTT, Rolnik DL et al. Machine Learning Algorithms Versus Classical Regression Models in Pre-Eclampsia Prediction: A Systematic Review, Current hypertension reports[J] 2024;26(7):309\u0026ndash;23\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1007/s11906-024-01297-1\u003c/span\u003e\u003cspan address=\"10.1007/s11906-024-01297-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGregory GA, Robinson TIG, Linklater SE, et al. Global incidence, prevalence, and mortality of type 1 diabetes in 2021 with projection to 2040: a modelling study, The lancet. Diabetes endocrinology[J]. 2022;10(10):741\u0026ndash;60. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1016/s2213-8587(22)00218-2\u003c/span\u003e\u003cspan address=\"10.1016/s2213-8587(22)00218-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarbie DA, Tamayo P, Boehm JS, et al. Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1. Nature[J]. 2009;462(7269):108\u0026ndash;12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1038/nature08460\u003c/span\u003e\u003cspan address=\"10.1038/nature08460\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCharoentong P, Finotello F, Angelova M, et al. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell reports[J]. 2017;18(1):248\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1016/j.celrep.2016.12.019\u003c/span\u003e\u003cspan address=\"10.1016/j.celrep.2016.12.019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCao H, Zhao J, Chen Z, et al. Loss of LIMCH1 predicts poor prognosis in patients with surgically resected Lung Adenocarcinoma: A study based on Immunohistochemical Analysis and Bioinformatics. J Cancer[J]. 2021;12(1):181\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.7150/jca.47883\u003c/span\u003e\u003cspan address=\"10.7150/jca.47883\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang J, Wang X, Zhang C, et al. Microplastics induce immune suppression via S100A8 downregulation. Ecotoxicol Environ safety[J]. 2022. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1016/j.ecoenv.2022.113905\u003c/span\u003e\u003cspan address=\"10.1016/j.ecoenv.2022.113905\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 242:113905.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGros F, Muller S. The role of lysosomes in metabolic and autoimmune diseases, Nature reviews. Nephrology[J]. 2023;19(6):366\u0026ndash;83. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1038/s41581-023-00692-2\u003c/span\u003e\u003cspan address=\"10.1038/s41581-023-00692-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiguelena Chamorro B, De Luca K, Swaminathan G, et al. Bordetella bronchiseptica and Bordetella pertussis: Similarities and Differences in Infection, Immuno-Modulation, and Vaccine Considerations. Clin Microbiol reviews[J]. 2023;36(3):e0016422. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1128/cmr.00164-22\u003c/span\u003e\u003cspan address=\"10.1128/cmr.00164-22\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhuang K, Zhang Y, Mo P, et al. Plasma proteomic analysis reveals altered protein abundances in HIV-infected patients with or without non-Hodgkin lymphoma. J Med virology[J]. 2022;94(8):3876\u0026ndash;89. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1002/jmv.27775\u003c/span\u003e\u003cspan address=\"10.1002/jmv.27775\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTian Y, Liu X, Wang J, et al. Antitumor Effects and the Potential Mechanism of 10-HDA against SU-DHL-2 Cells, Pharmaceuticals (Basel. Switzerland)[J]. 2024;17(8). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.3390/ph17081088\u003c/span\u003e\u003cspan address=\"10.3390/ph17081088\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eManfroi B, De Grandis M, Moreaux J et al. The microenvironment of DLBCL is characterized by noncanonical macrophages recruited by tumor-derived CCL5, Blood advances[J] 2021;5(21):4338\u0026ndash;51\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1182/bloodadvances.2021004203\u003c/span\u003e\u003cspan address=\"10.1182/bloodadvances.2021004203\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNielsen KR, Steffensen R, Bendtsen MD, et al. Inherited Inflammatory Response Genes Are Associated with B-Cell Non-Hodgkin's Lymphoma Risk and Survival. PloS one[J]. 2015;10(10):e0139329. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e.http://dx.doi.org/10.1371/journal.pone.0139329\u003c/span\u003e\u003cspan address=\".10.1371/journal.pone.0139329\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHiguchi T, Hashida Y, Matsuo K, et al. EBV-positive pyothorax-associated lymphoma expresses CXCL9 and CXCL10 chemokines that attract cytotoxic lymphocytes via CXCR3. Cancer science[J]. 2023;114(6):2622\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1111/cas.15782\u003c/span\u003e\u003cspan address=\"10.1111/cas.15782\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRuiduo C, Ying D, Qiwei W. CXCL9 promotes the progression of diffuse large B-cell lymphoma through up-regulating β-catenin. Biomed pharmacotherapy = Biomedecine pharmacotherapie[J]. 2018;107:689\u0026ndash;95. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1016/j.biopha.2018.07.171\u003c/span\u003e\u003cspan address=\"10.1016/j.biopha.2018.07.171\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSrinagesh H, Jackson C, Shiraz P, et al. A phase 1 clinical trial of NKTR-255 with CD19-22 CAR T-cell therapy for refractory B-cell acute lymphoblastic leukemia. Blood[J]. 2024;144(16):1689\u0026ndash;98. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1182/blood.2024024952\u003c/span\u003e\u003cspan address=\"10.1182/blood.2024024952\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Q, Zheng T, Lan Q, et al. Single-nucleotide polymorphisms in genes encoding for CC chemokines were not associated with the risk of non-Hodgkin lymphoma, Cancer epidemiology, biomarkers \u0026amp; prevention: a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive. Oncology[J]. 2013;22(7):1332\u0026ndash;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1158/1055-9965.Epi-13-0328\u003c/span\u003e\u003cspan address=\"10.1158/1055-9965.Epi-13-0328\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou Q, Huang L, Gu Y et al. The expression of CCL18 in diffuse large B cell lymphoma and its mechanism research, Cancer biomarkers: section A of Disease markers[J] 2018;21(4):925\u0026ndash;34\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.3233/cbm-171097\u003c/span\u003e\u003cspan address=\"10.3233/cbm-171097\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKorbecki J, Grochans S, Gutowska I, et al. CC Chemokines in a Tumor: A Review of Pro-Cancer and Anti-Cancer Properties of Receptors CCR5, CCR6, CCR7, CCR8, CCR9, and CCR10 Ligands. Int J Mol sciences[J]. 2020;21(20). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.3390/ijms21207619\u003c/span\u003e\u003cspan address=\"10.3390/ijms21207619\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKorbecki J, Olbromski M, Dzięgiel P. CCL18 in the Progression of Cancer. Int J Mol sciences[J]. 2020;21(21). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.3390/ijms21217955\u003c/span\u003e\u003cspan address=\"10.3390/ijms21217955\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJin X, Ding H, Ding N, et al. Homozygous A polymorphism of the complement C1qA276 correlates with prolonged overall survival in patients with diffuse large B cell lymphoma treated with R-CHOP. J Hematol oncology[J]. 2012. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e5:51.http://dx.doi.org/10.1186/1756-8722-5-51\u003c/span\u003e\u003cspan address=\"5:51.10.1186/1756-8722-5-51\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi J, Zhu Z, Zhu Y, et al. METTL3-mediated m6A methylation of C1qA regulates the Rituximab resistance of diffuse large B-cell lymphoma cells. Cell death discovery[J]. 2023;9(1):405. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1038/s41420-023-01698-2\u003c/span\u003e\u003cspan address=\"10.1038/s41420-023-01698-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKorkmaz B, Caughey GH, Chapple I, et al. Therapeutic targeting of cathepsin C: from pathophysiology to treatment. Pharmacol therapeutics[J]. 2018;190:202\u0026ndash;36. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1016/j.pharmthera.2018.05.011\u003c/span\u003e\u003cspan address=\"10.1016/j.pharmthera.2018.05.011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatsumoto Y, Tsukamoto T, Chinen Y, et al. Detection of novel and recurrent conjoined genes in non-Hodgkin B-cell lymphoma. J Clin experimental hematopathology: JCEH[J]. 2021;61(2):71\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.3960/jslrt.20033\u003c/span\u003e\u003cspan address=\"10.3960/jslrt.20033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXiao Y, Cong M, Li J, et al. Cathepsin C promotes breast cancer lung metastasis by modulating neutrophil infiltration and neutrophil extracellular trap formation. Cancer cell[J]. 2021;39(3):423\u0026ndash;37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ee7.http://dx.doi.org/10.1016/j.ccell.2020.12.012\u003c/span\u003e\u003cspan address=\"e7.10.1016/j.ccell.2020.12.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu-Monette ZY, Xiao M, Au Q, et al. Immune Profiling and Quantitative Analysis Decipher the Clinical Role of Immune-Checkpoint Expression in the Tumor Immune Microenvironment of DLBCL. Cancer Immunol research[J]. 2019;7(4):644\u0026ndash;57. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1158/2326-6066.Cir-18-0439\u003c/span\u003e\u003cspan address=\"10.1158/2326-6066.Cir-18-0439\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYe X, Wang L, Nie M, et al. A single-cell atlas of diffuse large B cell lymphoma. Cell reports[J]. 2022;39(3):110713. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1016/j.celrep.2022.110713\u003c/span\u003e\u003cspan address=\"10.1016/j.celrep.2022.110713\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-immunology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"imno","sideBox":"Learn more about [BMC Immunology](http://bmcimmunol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/imno/default.aspx","title":"BMC Immunology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Diffuse Large B Cell Lymphoma (DLBCL), DEGs, WGCNA, Hub genes, Immune cell infiltration","lastPublishedDoi":"10.21203/rs.3.rs-5876864/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5876864/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e Diffuse large B cell lymphoma (DLBCL), one distinct origin of hematological malignancies, has caused a major public health problem. However, the molecular mechanisms was not been clearly elucidated. The aim is to explore disease-specific diagnostic biomarkers and mechanisms to improve this situation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eThree microarray datasets (GSE25638, GSE12195, GSE12453) were downloaded from the Gene Expression Omnibus (GEO) database. The key genes in DLBCL patients were screened by differential expression genes (DEGs) analysis and weighted gene co-expression network analysis (WGCNA). Functional enrichment analysis and protein-protein interaction (PPI) network construction were employed to reveal DLBCL-related pathogenic molecules and underlying mechanisms. Random forest analysis was adopted for screening candidate biomarkers, and Kaplan Meier survival analysis were constructed to predict the risk of patients. The single‐sample gene set enrichment analysis was used to explore immune cell infiltration in lymphoma. Validation of the hub genes expression was confirmed by RT-PCR and immunohistochemistry (IHC) tests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003e95 key genes were obtained from three datasets about DLBCL patients by DEGs and WGCNA. The four hub genes (CXCL9, CCL18, C1QA, CTSC) were screened by random forest analysis and machine learning algorithm. The ROC results showed that the AUC was 1.00 in the training set, and the bootstrap verification was performed for 1000 times in the external validation set, and the AUC size was 0.839. The several pathways were found by Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analysis. The four hub genes were identified to be excellent potential for the survival of DLBCL patients. Dysregulated immune cell infiltrations were observed in DLBCL, as well as positive correlations with the four hub genes, respectively. Validation of the hub genes with high expressions was also demonstrated in DLBCL patients.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003e This study identified four candidate hub genes (CXCL9, CCL18, C1QA, CTSC) that could predict the risk of DLBCL, and CXCL9 may be essential in developing the disease, which provided a new perspective for the molecular mechanism and therapeutic targets for DLBCL.\u003c/p\u003e","manuscriptTitle":"Exploration of biomarkers for predicting the prognosis of patients with diffuse large B-cell lymphoma by machine-learning analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-07 12:43:39","doi":"10.21203/rs.3.rs-5876864/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-05-05T05:05:33+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-28T03:30:23+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-15T12:32:23+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"256484484990257544700849947136567528048","date":"2025-04-14T16:54:50+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"221110237717605779694689447793533326634","date":"2025-04-04T13:20:20+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-04T13:06:45+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-04-02T14:33:33+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Immunology","date":"2025-03-30T10:29:27+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-immunology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"imno","sideBox":"Learn more about [BMC Immunology](http://bmcimmunol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/imno/default.aspx","title":"BMC Immunology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"467f0512-2f4c-44a9-b2fc-5a4c0b4f1091","owner":[],"postedDate":"April 7th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-08-25T16:31:18+00:00","versionOfRecord":{"articleIdentity":"rs-5876864","link":"https://doi.org/10.1186/s12865-025-00738-z","journal":{"identity":"bmc-immunology","isVorOnly":false,"title":"BMC Immunology"},"publishedOn":"2025-08-20 16:13:00","publishedOnDateReadable":"August 20th, 2025"},"versionCreatedAt":"2025-04-07 12:43:39","video":"","vorDoi":"10.1186/s12865-025-00738-z","vorDoiUrl":"https://doi.org/10.1186/s12865-025-00738-z","workflowStages":[]},"version":"v1","identity":"rs-5876864","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5876864","identity":"rs-5876864","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}