Spatial genomics to reveal an immunosuppressive tumor microenvironment after chemoradiation in head and neck cancer patients.

6059 Background: Although chemoradiotherapy (CRT) remains the cornerstone of management for locoregionally advanced head and neck squamous cell carcinomas (HNSCC), recurrence occurs in ~50% of patients. We sought to characterize the impact of CRT on the tumor microenvironment (TME) and identify factors associated with progression. Emerging spatial biology technologies, such as NanoString GeoMx Digital Spatial Profiler (DSP), enable spatially defined whole-transcriptomic profiling of cells along the tumor-stromal interface in tumor biopsies to resolve immune cell identities, activation states and signaling pathways in adjacent tumor cells. We used a multimodal approach including immunohistochemistry (IHC), bulk tumor gene expression and spatial genomics to characterize paired pre- and post-CRT tumor samples from HNSCC patients. Methods: 20 patients with locoregionally advanced HNSCC who had undergone curative intent CRT and developed subsequent recurrence requiring salvage surgery from June 14, 2010 to June 14, 2020 were retrospectively identified at Duke University Medical Center. Matched formalin-fixed paraffin embedded tumor specimens collected prior to CRT and at salvage surgery were compared via IHC for markers of immune infiltration and PDL1 expression, bulk tumor whole transcriptomic profiling, and spatial transcriptomic analysis of tumor and stroma using DSP were further examined in “early” progressors ( < 6 months after CRT) and “late” progressors (≥6 months after CRT). Results: IHC analyses demonstrated significant reductions in T-cell immune populations (total CD3 + , CD8 + , 4-1BB + cell density [ p < 0.001]) after CRT. PDL1 + expression was similar before and after CRT. Neither T-cell infiltrate nor tumor PDL1 levels at surgery were associated with progression. Bulk tumor gene expression analysis demonstrated that macrophage and fibroblast markers were increased post-CRT. In early progressors, molecular dissection of the tumor-stromal interface using DSP expression profiling revealed macrophage associated genes ( CD68, MRC1/CD206, CD163, F13A1) at salvage surgery had a more pronounced increase in macrophage markers in stroma with enriched M2 genes compared to late progressors. Conclusions: Integrated spatial analysis with high resolution transcriptomics and IHC suggest CRT induces an immunosuppressive environment characterized by decreased T-cell infiltration and an influx of putative immunosuppressive macrophages. We hypothesize that tumor associated macrophages and/or fibroblasts may contribute to immunosuppression due to “reprogramming” of the TME after CRT. This exploratory study suggests that patients with progressive disease after chemoradiation may benefit from additional immunoregulatory therapies that ameliorate the immunosuppressive TME. These findings will be further validated in a larger cohort of patients.