Comparison of the Lymph2Cx Assay and Hans Algorithm in Determining the Cell-of-Origin of Diffuse Large B-Cell Lymphomas, Not Otherwise Specified
Inju Cho, Nara Yoon, Jiyeon Hyeon, Jongmin Sim, Hae Yong Yoo, Seok Jin Kim, Won Seog Kim, Young Hyeh Ko
Abstract
In the era of precision medicine, accurate and reproducible assignment of cell-of-origin (COO) in diffuse large B-cell lymphoma patients has become important. The Lymph2Cx assay is accurately determining COO by analyzing RNA expression of 20 selected genes while the Hans algorithm based on immunohistochemistry is the most popular method for routine daily diagnosis. However, there are discrepancies between the 2 methods, which need to be evaluated for better correlation. We prospectively analyzed 156 cases of diffuse large B-cell lymphoma, not otherwise specified to analyze the characteristics of discrepancy groups of COO determined by Lymph2Cx and Hans algorithm. We investigated the pattern and cause of discrepancy of COO assigned by the 2 methods. Hans algorithm classified 50 cases (32%) as germinal-center B-cell-like (GCB) type and 106 cases (68%) as non-GCB type. Lymph2Cx assay assigned 43 cases (28%) as GCB type, 94 cases (60%) as activated B-cell-like type, and 19 cases (12%) as intermediate/unclassified type. The agreement rate was 86% after exclusion of unclassified type. With regard to the clinicopathologic factors related with discrepancy between Hans algorithm and Lymph2Cx assay, endoscopic biopsy of the gastrointestinal tract (4/11, 36%) showed higher discrepancy rate (P=0.052). Immunophenotypically, CD10/BCL6/MUM1 GCB type and CD10/BCL6//MUM1 (=30%, low level expression) non-GCB type exhibited a significantly higher discrepancy rate (6/13, 46%; 4/13, 31%) (P=0.0001). Activated B-cell-like subgroup via Lymph2Cx assay predicted poor progression-free survival (mean survival duration 28.6 mo, P=0.049) compared with the GCB and unclassified type. Hans algorithm revealed no significant difference in progression-free survival and overall survival (P=0.122 and 0.121). These results suggest that when assigning COO via Hans algorithm, CD10/BCL6/MUM1 GCB type and CD10/BCL6/MUM1 (=30%, low level) non-GCB type require careful interpretation, especially if the MUM1 staining is weak and heterogeneous in the biopsied specimen.