CD138在弥漫大B细胞淋巴瘤免疫分型构建及预后评估中的意义

Significance of CD138 in immunohistochemical profiles and its correlation with prognosis in diffuselarge B-cell lymphoma

目的探讨CD138的表达及不同免疫分型方法与弥漫大B细胞淋巴瘤（DLBCL）患者预后的关系。方法收集106例DLBCL组织标本，应用免疫组化EnVision法检测CD138、CD10、MUM1及bcl-6的表达。基于免疫标记结果，根据4种不同免疫分型方法对106例患者进行分型，比较各亚型患者的预后，分析不同免疫分型方法对DLBCL患者预后的影响。结果106例组织标本中，CD20均为弥漫阳性，CD3均为阴性，CD10的阳性率为21．7％，bcl-6的阳性率为26．4％，MUM1的阳性率为56．6％，CD138的阳性率为15．1％。其中CD10和bcl-6阳性者总生存期长于阴性者（P=0．001，P=0．041），而CD138阳性者总生存期短于阴性者（P=0．003）。多因素Cox回归分析显示，ECOG评分、Hans分型、Colomo分型分别为DLBCL患者生存期（OS）和无进展生存期（PFS）的独立影响因素，两种免疫分型方法相对于OS的OR值（0．259和0．255）和PFS的OR值（0．248和0．244）接近。结论Hans分型和Colomo分型均与DLBCL患者的预后有关，且对预后的影响力相近，CDl38的表达对DLBCL患者预后的预测意义不大。

Objective The purpose of this study was to classify the diffuse large B-cell lymphoma （DLBCL） into different prognostic subgroups according to four different detection methods of the expression of CD138, CD10, bcl-6, and MUM1. In particular to investigate the significance of CD138 in immunohistochemical profiles and its correlation with prognosis in DLBCL. Methods Immunohistochemical EnVision method was used to detect the expression of CD138, CD10, bcl-6 and MUM1 in 106 cases of DLBCL and reconstructed into four different subtyping algorithms. Algorithm-1, according to the expression of CD10, bcl-6 and MUM1, the cases were assigned to GCB and non-GCB groups. Algorithm-2,according to the expression of CDI38, CD10, bel-6 and MUM1, the cases were assigned to A, B, C, D groups. Algorithm-3 ,according to the expression of CDIO and MUM1, the cases were assigned to GCB and non-GCB groups. Algorithm-4, according to the expression of CD138, CD10, bcl-6 and MUM1, the cases were assigned to GCB and non-GCB groups. 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 and bcl-6 were positive in 15.1% （16/106）, 56.6% （60/106）, 21.7 （23/106） and 26.4% （28/ 106）, respectively. The expression of CDIO and bcl-6 was associated with favorable OS （P = 0. 001 and O. 041, respectively）, whereas the expression of CD138 was associated with unfavorable OS （P =0.003 ）. Using multivariate Cox proportional hazards regression analysis, algorithm-1 and -4 were almost at the same level for prognosis of OS （ OR = 0. 259, 0. 255 ） and PFS （ OR = 0. 248, O. 244 ）. Conclusions Both Hans＇s algorithm and Colombo＇s algorithm including CD138 detection are associated with the prognosis of DLBCL patients. The two algorithms have similar OR value according to Cox analysis. However, positive expression of CD138 is of minor significance in prediction of the prognosis in DLBCL patients.