Development And Evaluation Of Immune Cell Signature-based Prognostic Model Of Diffuse Large B-cell Lymphoma

Diffuse large B-cell lymphoma(DLBCL)is a clinically and genetically heterogeneous disease with complex tumor microenvironment(TME)alterations.CHOP(cyclophosphamide,doxorubicin,vincristine,and prednisone)chemotherapy is a classic treatment for DLBCL,particularly for the late-stage DLBCL patients.However,the CHOP chemotherapy could produce serious adverse reactions for patients with weak constitution.Currently,majority of the DLBCL patients are treated with R-CHOP(Rituximab,cyclophosphamide,doxorubicin,vincristine,and prednisone)first-line therapy(i.e.monoclonal antibody combined with chemotherapy),and the treatment efficiency and overall survival rate of patients treated by R-CHOP are much higher compared with traditional chemotherapy drugs(e.g.CHOP).Different from CHOP,R-CHOP exhibits no obvious toxicity,which makes cure possible.However,if early recurrence or treatment failure occurs after treatment with R-CHOP,the survival rate of patients will be very poor.Therefore,it is of great significance to evaluate the prognosis of DLBCL patients as soon as possible to formulate the treatment strategy of DLBCL.The methods for predicting prognosis of DLBCL mainly include IPI,R-IPI,NCCN-IPI,and genetic subtypes.However,these prognostic methods mainly focus only on the status of patients or the heterogeneity of the tumor,while none has considered the tumor microenvironmental factors,which probably makes the prognostic effect is not ideal.Till now,the prognostic value of immune cell signatures of TME in DLBCL remain largely unclear.The occurrence and progression of tumor may be closely related to the changes of immune cells in tumor microenvironment.To better understand the role of tumor microenvironment and to determine the function of immune cells in tumor progression and metastasis of DLBCL,we established a model by analyzing the relationship between immune cell characteristics and clinical prognosis of patients with DLBCL who received first-line chemotherapy or immune-chemotherapy.We aimed to identify high-risk DLBCL with specific immune cell signatures in TME.Clinical and gene expression data of DLBCL patients were obtained from previously reported retrospective datasets(n=408)in Gene Expression Omnibus(GSE10846 and GSE53786)and a multi-center prospective clinical trial NHL001(n=68,NCT01852435).Patients treated with CHOP regimen(n=159)from GSE10846 were referred as the training cohort for CHOP regimen,while patients treated with R-CHOP regimen(n=192)from GSE10846 were referred as the training cohort for R-CHOP regimen.Patients from NHL001(n=68)and GSE53786(n=57)were referred as the validation cohorts for R-CHOP regimen.We then used CIBERSORT software to estimate the relative proportions of 22 subtype of immune cells in TME of DLBCL.We established a prognostic model for CHOP regimen including age,Ann-Arbor stage,serum lactate dehydrogenase(LDH),and T cells CD4 memory activated,defining a low-risk group(n=80)with 2-yr overall survival(OS)of 82.2%(95%CI 73.7-90.7%)and a high-risk group(n=79)with 2-yr OS of 40.5%(95%CI 29.6-51.3)(p<0.0001).Moreover,prognostic model for R-CHOP regimen included Age,ECOG performance status,LDH,T cells follicular helper and macrophages M0,defining a low-risk group(n=103)with 2-yr OS of 92.9%(95%CI 87.9-98.0%)and a high-risk group(n=89)with 2-yr OS of 52.5%(95%CI 41.1-63.8)(p<0.0001).Taken together,the results showed that memory CD4 T activated cell could serve as a poor prognostic factor,while mast cell could serve as a good prognostic factor for CHOP regimen.Macrophage M0 was indicated to be a good prognostic factor for R-CHOP regimen.In summary,immune cell signatures could be used as prognostic markers and provided further insights for individualized immunotherapeutic strategies in DLBCL.