Single-cell spatial analysis of tumor immune architecture in diffuse large B cell lymphoma

Multiplexed immune cell profiling of the tumor microenvironment (TME) in cancer has improved our understanding of cancer immunology, but complex spatial analyses of tumor-immune interactions in lymphoma are lacking. Here we used imaging mass cytometry (IMC) on 33 cases of diffuse large B cell lymphoma (DLBCL) to characterize tumor and immune cell architecture and correlate it to clinicopathological features such as cell of origin, gene mutations, and responsiveness to chemotherapy. To understand the poor response of DLBCL to immune check point inhibitors (ICI), we compared our results to IMC data from Hodgkin lymphoma (HL), a cancer highly responsive to ICI, and observed differences in the expression of PD-L1, PD-1, and TIM-3. We created a spatial classification of tumor cells and identified sub-regions of immune activation, immune suppression, and immune exclusion within the topology of DLBCL. Finally, the spatial analysis allowed us to identify markers such as CXCR3, which are associated with penetration of immune cells into immune desert regions, with important implications for engineered cellular therapies. SIGNIFICANCEThis is the first study to integrate tumor mutational profiling, cell of origin classification, and multiplexed immuno-phenotyping of the TME into a spatial analysis of DLBCL at the single cell level. We demonstrate that, far from being histo-pathologically monotonous, DLBCL has a complex tumor architecture, and that changes in tumor topology can be correlated with clinically relevant features. This analysis identifies candidate biomarkers and therapeutic targets such as TIM-3, CCR4, and CXCR3 that are relevant for combination treatment strategies in immuno-oncology and cellular therapies such as CAR-T cells.