The clinical and molecular characteristics of DZsig positive (DZsigpos) tumors in relation to HGBCL-DH-BCL2. (A-B) FFP and OS of patients treated with R-CHOP harboring DZsigpos tumors in the presence or absence of a FISH-based diagnosis of HGBCL-DH-BCL2 (HGBCL-DH-BCL2 DZsigpos vs DLBCL DZsigpos respectively), including the small subset of HGBCL-DH-BCL2 tumors that lack DZsig expression (HGBCL-DH-BCL2 DZsigneg/ind). DLBCL DZsigneg/ind are the remaining majority of patients with DLBCL, NOS. Patients harboring tumors with evaluable FISH and GEP data were included. Patients with HGBCL with MYC and BCL6 rearrangements (n = 8) were included in the DLBCL groups. Two patients with ABC tumors that are DZsigpos were excluded from the analysis. Log-rank test was used to compare survival curves. ∗∗∗∗P < .0001, ∗∗∗P < .001, ∗P < .05. (C) Molecular characterization of HGBCL-DH-BCL2-DZsigpos, DLBCL DZsigpos, and the HGBCL-DH-BCL2 DZsigneg/ind in a cohort combining the study population and extension cohort (1112 evaluable biopsies). Tumors were profiled by: (1) Digital GEP by the DLBCL90 assay for COO, DZsig, and relative MYC and BCL2 mRNA expression (z scores) displayed as a heat map in the bottom 2 rows. (2) FISH for MYC and BCL2 rearrangements. (3) Targeted capture or whole genome sequencing to identify MYC rearrangement partner in MYC rearranged tumors. (4) IHC for MYC, BCL2 and the GC marker CD10. “ABC” is a separate track of rare ABC tumors (n = 3) that were DZsigpos (marked with ∗, as not applicable in ABC) and/or HGBCL-DH-BCL2, which were excluded from subsequent analysis. (D) Box plots showing relative mRNA expression of MYC (left) and BCL2 (right) in HGBCL-DH-BCL2-DZsigpos tumors vs HGBCL-DH-BCL2 DZsigneg/ind shown as z scores from log2 normalized counts with means compared by t test. (E) Proportion of tumors positive for CD10 by IHC in HGBCL-DH-BCL2-DZsigpos vs HGBCL-DH-BCL2 DZsigneg/ind compared by Fisher exact test. IG, immunoglobulin; ind, indeterminate DZsig; mRNA, messenger RNA; UNCLASS, unclassified COO.
