Abstract P080

Feasibility of tumor microenvironment analysis in pediatric and adolescent classical Hodgkin lymphoma by automated cyclic immunofluorescence staining

Background: The tumor microenvironment (TME) plays a central role in classical Hodgkin lymphoma (cHL) biology. While the TME has been extensively studied in adult cHL, spatial and single-cell data of pediatric and adolescent patients remain limited. In this study, we implemented cyclic immunofluorescence to characterize the cellular composition of the TME in pediatric and adolescent cHL.

Methods: Diagnostic biopsy material from pediatric and adolescent patients treated within the EuroNet-PHL-C1 trial was analyzed (Mauz-Körholz et al., 2023). Early response after two cycles of OEPA was defined as adequate response (AR) with partial morphological remission or greater and PET negativity and inadequate response (IR) when criteria for AR were not met (Mauz-Körholz et al., 2023). Overall, n=39 specimens were assembled in a tissue microarray (TMA). Multiplex phenotyping was performed on TMAs using the MACSimaTM imaging platform and a 28-antibody panel (including CD30, CD3, CD4, CD8a, CD20, CD11c, CD163, CD68, CD56, CD66b, CD45RA, CD45RO, LAG3, TIM3, PD1, PDL1, HLA-DR, HLA-ABC). Single-cell segmentation and phenotyping were performed using StarDist and Nimbus, followed by downstream quantification in QuPath.

Results: By standard protocols of multiplex immunophenotyping, Hodgkin-Reed-Sternberg cells (HRSCs) were insufficiently detected and frequently oversegmented (one HRSC erroneously considered as multiple cells due to multinuclearity). We established and optimized a workflow allowing us to reliably identify and localize HRSCs as a prerequisite for spatial analysis in cHL. Moreover, we developed an analysis pipeline that allows us to assign phenotypes to the densely packed cells of the cHL microenvironment, including reliable macrophage detection. Overall, our assay assigned a phenotype to 78.5% of all cells in the tissue. As expected, HRSCs comprised a minority of cells (1.1%) and CD4+ T-cells (44.7%) dominated the TME. The content of B-cells was variable, and NK-cells were rarely detected. M1 macrophages were more abundant than M2 macrophages.

Conclusion: We demonstrate feasibility of multiplex immunophenotyping of the TME in cHL using the MACSimaTM imaging platform. We managed to overcome limitations in detecting HRSCs and implemented a data analysis pipeline for the TME of cHL. Data acquisition is currently ongoing and a correlation of TME features and treatment response (AR vs. IR) will be presented at the meeting.