论文部分内容阅读
Objective Primary breast diffuse large B-cell lymphoma(PB-DLBCL),the most common histologic subtype of lymphoid malignancy in the breast,is a clinically and genetically heterogeneous disease that has insufficient systematic studies on the pathological and molecular features,optimal treatment scheme,as well as the prognostic factors.Generally,PB-DLBCL mainly shows a phenotype of ABC(60-90%)according to the COO classification.However,as an uncommon site of involvement in extranodal DLBCL,PB-DLBCL has its own particularity with heterogenic biological and clinical characters.B-cell receptor(BCR)-pathway components and networks are complex and variable,which may provide the basis of underlying the biological diversity of the PB-DLBCL.The aim of our study was to identify the distinct subtypes of PB-DLBCLs and then evaluate the prognosis of this rare malignant lymphoma.Methods We retrospectively studied 68 cases of female patients with PB-DLBCLs,whose histologic type was classified according to the 2016 World Health Organization(WHO)classification.Formalin-fixed,paraffin-embedded tissue specimens were collected from 68 patients diagnosed with PB-DLBCL patients for immunohistochemistry(IHC)analysis.We carried out hierarchical clustering analysis to evaluate protein expression detected by IHC staining of samples from 68 PB-DLBCL patients.The gene expression data from TCGA database was obtained to validate the identified clusters.Furthermore,we analyzed the correlation between clusters and clinicopathological parameters of PBDLBCL patients.Survival outcomes were analyzed by Kaplan-Meier methods.Results We firstly screened several general clinical and biological markers of lymphoma cells,and then detected the protein expressions of these markers by IHC studies.The expressions of CD10,BCL6,MUM1,PI3K,AKT2,JAK2,STAT3,MAPK,BCL10,NF-κ B p50,Myc,BCL2,MCL1,BCL-xL,Ki67 and P53 were detected.According to hierarchical clustering analysis,we identified three robust clusters based on the BCR signaling pathway,including two recognized NF-κ B-dependent and PI3K-dependent clusters,and a distinct subset of PB-DLBCL with NF-κ B-independent anti-apoptotic overexpression plus PI3K signaling,which exhibited an evolving definition and distinctive characters of a cluster group.Our cluster classification based on the BCR signaling pathway was coincidence with the DLBCL datasets from the TCGA databases,and identification of the clusters was associated with different COO subtypes in PBDLBCL.In addition,the clusters were significantly associated with the clinical stages(P = 0.018)and the Ki67 expressions(P = 0.032).Furthermore,survival analysis results showed an inferior outcome in NF-κ B-dependent cluster patients and favorable survival in the PI3K-dependent cluster patients,suggesting an important predictive value of the three clusters.Conclusions Our study provided a new perspective for understanding clinical complexity of PB-DLBCLs,and gave evidence for finding targeted treatment strategies.Our clustering studies to identify three robust PB-DLBCL clusters distinctively according to BCR signaling components presented a novel signature for assessing previously unrecognized protein expression subsets.These findings in our study provided two important implications in PB-DLBCL patients.First,the distinct protein expression subtypes driven by different BCR signaling pathways might contribute to the subsequent biological behavior of the lymphoma cells.Second,each cluster had distinct survival outcomes after therapy and probably guided to the selection of targeted therapies owing to their distinct signaling abnormalities.