Fig. 3. Molecular and genetic analysis in DLBCL, not otherwise specified (NOS) stratified by MYC/BCL2 expression status and high grade B-cell lymphoma (HGBCL) with MYC/BCL2 double-hit (DH).

(A) Distribution plot for DLBCL-NOS with MYC/BCL2 double-high expression (DhE) and HGBCL-MYC/BCL2-DH cases. (B) Distribution plots in DLBCL-NOS cases with MYC-intermediate expression (MYC^inter) and BCL2-high expression (MYC/BCL2 double-expression, DE) and nonDE-DLBCL-NOS cases. Each column in the oncoplots represents one patient. Each row shows distribution of a genetic alteration with positive cases highlighted and prevalence shown on the right. Genetic alterations shown in the plots include non-synonymous mutations and copy number alterations detected by targeted next-generation sequencing and MYC and BCL2 alterations detected by fluorescence in situ hybridization occurring in ≥3 patients. Enriched genes in DhE with significantly higher mutation frequencies in DhE than in MYC^inter-DE cases or nonDE cases by Fisher’s exact test are marked by red asterisks. *: P ≤ 0.05, **: P ≤ 0.01, ***: P ≤ 0.001. Up arrows next to the ABC/GCB molecular subtypes and MCD/EZB genetic subtypes indicate their increased frequencies in the subgroup (significance was determined by Fisher’s exact test). KMT2D, ARID1A, TET2, SMARCA4, ETV6, and KDR are highlighted in red to indicate that their mutations had significant adverse prognostic effects in DE-DLBCL cases. The distribution of BCL2 gene rearrangement in MYC^inter-DE-DLBCL-NOS cases is highlighted in green, to indicate its non-significant association with a better overall survival (OS, P = 0.069). Abbreviations: rearrange., rearrangement; amp, amplification. (C) In DhE patients, EZB subtype and KMT2D mutation were associated with poorer OS with a statistically non-significant and significant P value, respectively. DhE patients with both EZB subtype and KMT2D mutations had significantly poorer OS compared with DhE patients with non-EZB subtypes.