CHARACTERIZATION OF THE GENETIC LANDSCAPE OF HIGH‐GRADE B‐CELL LYMPHOMA, NOS – AN LLMPP PROJECT
Brett Collinge, Laura K. Hilton, Jasper Wong, Susana Ben‐Neriah, Christopher Rushton, Graham W. Slack, Pedro Farinha, James R. Cook, German Ott, Andreas Rosenwald, Elı́as Campo, C. Amador, Timothy C. Greiner, Philipp W. Raess, Joo Y. Song, Giorgio Inghirami, Elaine S. Jaffe, Dennis D. Weisenburger, Wing C. Chan, Harald Holte, Klaus Beiske, Kai Fu, Jan Delabie, Stefania Pittaluga, Andrew L. Feldman, Kerry J. Savage, Andrew J. Mungall, Louis M. Staudt, Christian Steidl, Lisa M. Rimsza, Ryan D. Morin, David W. Scott
Abstract
Introduction: High-grade B-cell lymphoma encompasses (1) tumours with MYC and BCL2 and/or BCL6 rearrangement (HGBL-DH/TH), and (2) tumours lacking this combination of rearrangements but displaying “high-grade” morphology (HGBL-NOS) – blastoid or morphology intermediate between diffuse large B-cell lymphoma (DLBCL) and Burkitt lymphoma (BL). Limited by the rarity of HGBL-NOS, it remains unclear whether these tumours share a unifying biology and/or bear molecular similarity to HGBL-DH/TH. This uncertainty is further confounded by interobserver variability in morphological assessment and artifacts introduced during tissue fixation. Here, we aim to characterize the molecular features of HGBL-NOS, as confirmed by central pathology review (CPR) by a panel of expert lymphoma pathologists, to determine if it represents a biologically distinct subset of tumours. Methods: Archival diagnostic tissue biopsies from 93 HGBL-NOS tumours were submitted from LLMPP sites, with fluorescent in situ hybridization (FISH) testing for MYC, BCL2 and BCL6 to exclude HGBL-DH/TH. CPR has been completed for 61 tumours. Cell-of-origin (COO) and double hit signature (DHITsig) were determined by gene expression profiling. Recurrent somatic mutations (SNVs/Indels) were identified from the intersection of three out of four predictions by Strelka2, LoFreq, Mutect2 and SAGE on whole genome or exome sequencing of 46 tumours. Genetic subtypes were assigned using the LymphGen data portal. Results: MYC rearrangement was detected in approximately half of all identified tumours, while a similar but not fully overlapping proportion were DHITsig+. Cases spanned the GCB (56%) and ABC (28%) COO subgroups, with 16% unclassified. Following CPR, nearly half of tumours were reclassified as DLBCL (44%) or BL (5%). The frequency of MYC rearrangement and the distribution of COO subgroups was similar between tumours confirmed as HGBL-NOS and those reclassified as DLBCL, while 55% of confirmed HGBL-NOS were DHITsig+ compared with 39% of reassigned tumours. Few recurrently mutated genes were shared among confirmed HGBL-NOS tumours and no differentially mutated genes were identified through comparison to tumours reclassified as DLBCL. While most tumours were unclassified by the LymphGen algorithm, the remainder were spread across EZB (13%), MCD (9%), ST2 (7%), and BN2 (4%). Conclusion: The high reclassification rate highlights the difficulty in distinguishing high-grade morphology, while the lack of shared molecular features among CPR confirmed HGBL-NOS suggests this morphological distinction identifies a heterogeneous group of tumours. Furthermore, these tumours appear to be largely unrelated to HGBL-DH/TH-BCL2, which are almost all uniformly EZB, GCB and DHITsig+. Further molecular characterization and comparisons to related entities is required to determine how these tumours are best classified. Keywords: Genomics, Epigenomics, and Other -Omics, Aggressive B-cell non-Hodgkin lymphoma, Pathology and Classification of Lymphomas No conflicts of interest pertinent to the abstract.