95 Characterizing the tumor microenvironment of a colorectal adenocarcinoma sample with the Orion spatial biology platform and HALO quantitative image analysis

Background

Understanding the tumor microenvironment (TME) can help drive improved treatments. High-grade colorectal adenocarcinoma (CRC) is poorly differentiated and spreads more quickly than others. Thus, with poor patient survival prognostic markers can help improve treatment. Multiplex immunofluorescence and digital pathology can provide a spatial understanding of cellular relationships within the TME. With the ability to not only identify but also quantify many biomarkers, this can provide insight into the interactions tumorigenic and inflammatory cells. Advances in Artificial intelligence (AI) will allow for separation of tumor and stroma and thus identify the biological interactions and immunoprofiling between/of the two regions. Here we identify the inflammatory state of the tumor biopsy and explore the potential for targeted biomarker treatments such as the role of PD-L1/anti-PD1 therapy for CRC.

Methods

A formalin fixed paraffin embedded (FFPE) colorectal tissue section was stained using a multiplex immunofluorescence (IF) panel in one staining round, followed by whole-slide imaging with the Orion™ instrument. All the markers were captured in one imaging round, along with tissue autofluorescence, which was isolated as an additional channel to provide additional tissue morphology information. The panel consisted of Hoechst/nucleus, autofluorescence, Ki-67, CD3e, CD4, CD163, CD8a, PD-L1, PD-1, FOXP3, CD68, CD20, Pan-CK, and Vimentin. The resulting digital slide was analyzed using the HALO® image analysis platform with the Highplex FL analysis module from Indica Labs to better understand the TME. HALO AI networks were leveraged to segment tumor and stromal tissue compartments and improve nuclear and membrane segmentation across the slide. Once tissue and cells were segmented cell phenotyping, proximity analysis, and density heat mapping was performed.

Results

Nearly 2 million cells were quantified on this high-grade colon adenocarcinoma tumor specimen. Cell counts were obtained to identify multiple phenotypes like cytotoxic T cells, tumor associated macrophages, and activated tumor cells. Prominent clusters of PD-L1 expression were identified with the density heatmaps. Spatial analysis was employed to map the degree of T-cell infiltration in the tumor microenvironment.

Conclusions

A combined workflow using the Orion platform for multiplex IF staining and imaging with HALO quantitative image analysis delivers a comprehensive solution for exploring the immune landscape of the TME.