Ennishi D, Jiang A, Boyle M, et al.
Double-hit expression signature defines a distinct subgroup of germinal center B-cell-like diffuse large B-cell lymphoma.
J Clin Oncol.
2018; doi: 10.1200/JCO.18.01583. [Epub ahead of print].

Approximately 10 percent of diffuse large B-cell lymphoma (DLBCL) cases carry translocations that involve the gene for MYC and the gene for BCL2 on chromosomes 8 and 14, respectively. These double-hit lymphomas (DHLs) have inferior outcomes with standard R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone) chemotherapy. They have been reclassified as high-grade B-cell lymphomas with double-hit cytogenetics (HGBL-DH) and are almost exclusively of germinal center derivation. Although we are aware of their poor prognosis, the genetic and molecular events that define these lymphomas have not been investigated, and  alternative treatment options are therefore lacking.

Dr. Daisuke Ennishi and colleagues report a comprehensive genetic evaluation of a cohort of 157 germinal center B-cell (GCB) –derived DLBCL tumors, 25 (16%) of which had double-hit cytogenetics. Through RNAseq, they identified 104 genes that were most significantly up- or down-regulated between DHL and non-DH GCB DLBCL. This gene signature, termed DHITsig, was positive in 42 tumors (27%), including 22 (88%) of the 25 DHLs. It only missed three of the DHLs in this cohort but included an additional 20 lymphomas that did not have double-hit cytogenetics. Despite this, this gene signature was associated with inferior treatment outcomes with R-CHOP, including shorter time to progression, decreased disease-specific survival, and decreased overall survival, compared with the DHITsig-negative cohort. This inferior prognosis held for the non-DHL patients as well. The DHITsig remained significant in multivariable analyses and was not associated with other poor prognostic factors such as the International Prognostic Index or tumor volume. The genetic signature was validated in an independent cohort of GCB DLBCL. The signature was enriched for the upregulation of genes associated with the intermediate and dark, rather than light, zones of the germinal center, as well as for MYC and E2F targets, genes associated with oxidative phosphorylation, and MTORC1 signaling. Additionally, these tumors had reduced expression of inflammatory and/or immune signatures, reduced numbers of CD4+ tumor infiltrating lymphocytes, and loss of surface major histocompatibility complex molecules. Mutations in genes involved in chromatin modification such as CREBBP, EZH2, and KMT2D, as well as in TP53 and DDX3X were more frequent in the DHITsig-positive group. 

Importantly, because RNAseq is not readily available as a clinical test for patients with newly diagnosed DLBCL, this group developed a clinically applicable, NanoString-based assay using a 30-gene module to identify these patients in real time in a nonresearch setting. Using this test on a cohort of 347 patients, less than 5 percent of tumors were misclassified, and only approximately 10 percent of tumors were indeterminant. Most significantly, the prognostic significance of the DHITsig was maintained with DHITsig-positive and DHITsig-indeterminant tumors having significantly inferior outcomes.

This study signifies the first investigation into the genetic and molecular factors that associate with the inferior outcomes seen in DHL. It defines a genetic signature that identifies not only 88 percent of DHL but an additional population of high-risk non-DHL patients that constitute 10 to 15 percent of GCB DLBCL. This work identifies new targets in these diseases and thus new treatment options, including histone deacetylase inhibitors and drugs that target oxidative phosphorylation and the proteasome. Finally, the findings of this study will not remain of solely academic interest as they are accompanied by the development and validation of a clinically available test that can be used in a real-world setting.

Competing Interests

Dr. Jacobson indicated no relevant conflicts of interest.