| Objective The purpose of this study is to subclassify the diffuse large B-cell lymphoma (DLBCL ) into different prognostic subgroups according to four different methords with the expression of, CD138,CD10,Bcl-6,MUM1,Bcl-2. Partically to investigate the significance of different subclassifing methords. This study is also to investigate the genetic abnormalities about t(14;18) translocation and MYC gene rearrangent and their correlation to prognosis in DLBCL. Methods Immunohistochemical Envision methord was invited to detect the expression of CD138,CD10,Bcl-6,MUM1 in 106 cases of DLBCL and reconstructed into four different subtyping algorithms. Biopsy specimens from 106 DLBCLs were also analyzed for t(14;18) translocation and MYC gene rearrangent using interphase fluorescent in situ hybridization (FISH). For t(14;18) translocation polymerase chain reaction (PCR) was also performed. Following up was included as well. Statistical analysis was performed using the spss 13.0 and differences were considered significant at p<0.05. Results CD138, MUM1, CD10,Bcl-6 and Bcl-2 were positive in 15.1%(16/106), 56.6%(60/106), 21.7(23/106), 26.4%(28/106) and 73.6%(78/106) respectively. The expression of CD10 and Bcl-6 was associated with favourable OS (p=0.001 and 0.041 respectively), whereas the expression of CD138 was associated with unfavourable OS (p=0.003). According to algorithm 1 and 4,GCB subgroup showed favourable 5 year OS and PFS than non-GCB subgroup(p<0.001). Using multivariate COX proportional hazards regression analysis, algorithm 1 and 4 were almost in the same level for prognosis of OS(RR=0.259,0.255) and PFS(RR=0.248,0.244). The t(14;18) translocation was identified in 27 of 106 cases (25.5%) by FISH and 16 of 106 cases (15.1%) by PCR. The translocation was significantly correlated with the expression of CD10 protein and immunophenotypic subtype(p<0.001). No association between the expression of bcl-2 protein and t(14;18) translocation was found. Multivariate analysis confirmed that both t(14;18) translocation and bcl-2 expression were significantly associated with DLBCLs patients'survival. However, compared with bcl-2 expression, t(14;18) translocation indicated a worse outcome to patients.(for OS hazard ratio 4.235; 95%CI, 2.153-8.329, p<0.001 and 2.743; 95%CI, 1.262-5.962, p=0.011). The MYC translocation was identified in 13 (12.3%) of 106 cases. All MYC~+ DLBCLs showed a non germinal center (nonGCB) type. MYC~+ DLBCLs showed significantly poorer overall survival (OS) and progression free time (PFS), with a median OS and PFS time of 4.7 and 3.2 months respectively (p<0.001). Multivariate analysis using a Cox proportional hazard model confirmed that MYC~+ ( fou OS, RR 5.254,; 95%CI, 2.354-11.723, p<0.001) was the strongest independent predictor.Conclusions The expression of CD138 was associated with unfavourable prognosis, whereas CD10 and Bcl6 showed favourable prognosis.Each algorithm played its role in subclassifying. With survival analysis algorithm 1 and 4 even identified some cases with high risk in death and thus seem to be more applicable in our series. FISH was superior to PCR to investigate t(14;18) translocation. The subgroup of GCB DLBCL with t(14;18) translocation has poorer outcome. Therefore, t(14;18) was a beneficial factor for evaluation of DLBCL immunophenotype and prediction prognosis. The outcome of GCB -type DLBCL patients should be interpreted in the context of the t(14;18) translocation, and detection of the t(14;18) translocation should be included as a routine diagnostic test in these cases. DLBCL with MYC rearrangement is associated with poor outcome. The outcome of patients with the two genetic abnormalities were even poorer. Because only a few patients received rituximab, and its usefulness could not be assessed. Future studies with larger numbers of patients are required. |