|| Checking for direct PDF access through Ovid
The 2001 WHO classification distinguishes five variants (centroblastic, immunoblastic, plasmablastic, anaplastic and T-cell rich) and three subtypes (primary mediastinal, intravascular and primary effusion large B-cell lymphoma) of diffuse large B-cell lymphomas (DLBCLs).The recognition of the three subtypes as distinct disease entities can be considered as an advance in our understanding of these tumours. However, the variants of DLBCLs, which significantly outnumber the subtypes in frequency, represent an unresolved area. Gene expression profiling (GEP) of the variants led to a discrepancy in results and produced more questions than answers. The authors, therefore, initiated a multi-institutional collaborative research project in Germany aimed at a subtle morphologic, genomic and transcriptional characterisation of DLBCLs and Burkitt lymphoma (BL). We included BL in our study for two reasons: (1) it belongs to aggressive B-NHLs; and (2) at present, there are no reliable criteria that can be applied to distinguish BL from DLBCL. The GEPs derived from 200 patient samples were correlated with reviewed histology, the degree of genetic imbalances and clinical features. The results of this approach show that: (i) the DLBCL can be divided into more than four molecular groups; and (ii) the BL cases, identified by the consensus of five out of six lymphoma expert pathologists, displayed a genomic and gene expression profile that was clearly distinct from those of most DLBCLs. The group of DLBCLs that resembled BL in their GEP had a remarkably good prognosis, whereas those that differed in their GEP from the consensus BLs had unfavourable survival rates. In conclusion, combined application of genomic and gene expression profiling in conjunction with consensus reviewed histology and clinical features, appears to be a reliable approach that enables a reproducible and clinically meaningful characterisation of mature aggressive B-NHLs.