Predicting outcome in dogs with diffuse large B-cell lymphoma with a novel immune landscape signature.

Canine diffuse large B-cell lymphoma (cDLBCL) is characterized by high mortality and clinical heterogeneity. Although chemo-immunotherapy improves outcome, treatment response remains mainly unpredictable. To identify a set of immune-related genes aberrantly regulated and impacting the prognosis, we explored the immune landscape of cDLBCL by NanoString. The immune gene expression profile of 48 fully clinically characterized cDLBCLs treated with chemo-immunotherapy was analyzed with the NanoString nCounter Canine IO Panel using RNA extracted from tumor tissue paraffin blocks. A Cox proportional-hazards model was used to design a prognostic gene signature. The Cox model identified a 6-gene signature (, , , , ) strongly associated with lymphoma-specific survival, from which a risk score was calculated. Dogs were assigned to high-risk or low-risk groups according to the median score. Thirty-nine genes were differentially expressed between the 2 groups. Gene set analysis highlighted an upregulation of genes involved in complement activation, cytotoxicity, and antigen processing in low-risk dogs compared with high-risk dogs, whereas genes associated with cell cycle were downregulated in dogs with a lower risk. In line with these results, cell type profiling suggested the abundance of natural killer and CD8+ cells in low-risk dogs compared with high-risk dogs. Furthermore, the prognostic power of the risk score was validated in an independent cohort of cDLBCL. In conclusion, the 6-gene-derived risk score represents a robust biomarker in predicting the prognosis in cDLBCL. Moreover, our results suggest that enhanced tumor antigen recognition and cytotoxic activity are crucial in achieving a more effective response to chemo-immunotherapy.