PF530 CORRELATION OF CD10 EXPRESSION DETERMINED BY FLOW CYTOMETRIC IMMUNOPHENOTYPING AND CELL-OF-ORIGIN DETERMINATION BY GENE EXPRESSION PROFILING IN DIFFUSE LARGE B-CELL LYMPHOMA

Abstract

Background: Diffuse large B-cell lymphomas (DLBCL) represent a clinically heterogeneous group that is classified together based on similarities in morphology and immunophenotype. However, it was discovered that gene expression profiling (GEP) could classify DLBCL into distinct molecular subgroups based on cell-of-origin (COO), including germinal center B-cell type (GCB), activated B-cell type (ABC), and unclassified (UNC) type. COO assignment of DLBCL has important biological and prognostic significance, as well as potential therapeutic implications, with the ongoing development of selective agents for treatment of specific DLBCL subtypes. Indeed, the current WHO Classification for lymphoma requires COO determination for all cases of DLBCL at diagnosis using either GEP or immunohistochemistry methods. Flow cytometric immunophenotyping is also commonly used in the diagnostic workup of DLBCL, and we were interested to determine how well CD10 expression determined by flow cytometry correlated with GCB type DLBCL using a clinical Lymph2Cx GEP COO assay. Here, we summarize the results of 74 clinical DLBCL cases analyzed by both flow cytometric immunophenotyping and Lymph2Cx COO testing. Aims: We wished to determine the predictive value of CD10 expression determined by flow cytometric immunophenotyping for the presence or absence of GCB type DLBCL established by GEP. Methods: 74 clinical DLBCL cases were subjected to Lymph2Cx GEP COO analysis performed on the NanoString nCounter® Digital Analyzer and flow cytometric immunophenotyping with CD10 expression using a BD FACSCanto II flow cytometer. Hans algorithm immunohistochemical stains (CD10, BCL6, and MUM1) were also performed on 65 of the 74 cases. Results: Lymph2Cx COO testing detected 40 GCB, 25 ABC, and 9 UNC cases. 36 cases (49%) were positive for CD10 expression by flow cytometric immunophenotyping, and 38 cases were negative for CD10 expression (51%). 32 of the 36 CD10-positive cases were determined to GCB type by Lymph2Cx testing. 30 of the 38 CD10-negative cases were determined to be either ABC or UNC type by Lymph2Cx analysis. The predictive value of CD10 expression by flow cytometry was 89% for GCB type DLBCL by GEP COO analysis. In CD10-negative cases, the predictive value for the presence of ABC or UNC type DLBCL was 79%. Interestingly, CD10-positive cases by flow cytometry also correlated with GCB type DLBCL determined by Hans algorithm immunostains in 34 of 35 cases (97%), but only 31 of these cases (89%) were true GCB-DLBCL by GEP analysis. However, lack of CD10 expression by flow cytometry correlated with non-GCB type DLBLC by the Hans algorithm in only 21 of 30 cases (70%), with nine cases exhibiting a GCB type immunohistochemical profile. Summary/Conclusion: In summary, we have demonstrated that the presence of CD10 expression in DLBCL determined by flow cytometric immunophenotyping shows a very strong correlation with GCB type by Lymph2Cx GEP COO testing. CD10 expression by flow cytometric immunophenotyping also correlates well with GCB designation by the Hans algorithm immunohistochemical staining. Lack of CD10 expression by flow cytometric immunophenotyping also showed good correlation with ABC or UNC type DLBCL determined by GEP. These findings suggest that when clinical samples are limiting and only limited immunophenotyping data is available, the presence or absence of CD10 expression determined by flow cytometric immunophenotyping can be helpful for predicting COO in DLBCL.