Poster Presentation 37th Lorne Cancer Conference 2025

Spatial resolution of the head and neck cancer tumour microenvironment to identify tumour and stromal features associated with therapy response (#108)

Naomi Berrell 1 2 , James Monkman 1 , Meg L Donovan 2 , Tony Blick 1 , Ken O'Byrne 3 , Rahul Ladwa 1 4 , Chin Wee Tan 1 5 6 , Arutha Kulasinghe 1 2
  1. Frazer Institute, Faculty of Medicine, The University of Queensland, Brisbane, QLD, Australia
  2. Queensland Spatial Biology Centre, Wesley Research Institute, Level 8 East Wing, The Wesley Hospital, Auchenflower, QLD, Australia
  3. Princess Alexandra Hospital, Brisbane, QLD, Australia
  4. Department of Head & Neck Oncology, Princess Alexandra Hospital, Brisbane, QLD, Australia
  5. Bioinformatics Division, Walter and Eliza Hall Institute of Medical Research, Melbourne, VIC, Australia
  6. Department of Medical Biology, Faculty of Medicine, Dentistry and Health Sciences, University of Melbourne, Parkville, VIC, Australia

Despite advancements in various cancer treatments, including chemoradiotherapy, surgery and immunotherapies, treatment for head and neck cancers (HNC) remains a challenge. For patients with HNC, relapse or recurrent disease will occur in half of all patients, and these patients have a median overall survival of 8.9 (HPV negative tumours) and 18 months (HPV positive tumours).  Immunotherapies are showing long term and durable responses in a subset of patients, but we lack robust, predictive biomarkers for response in HNC. Understanding the tumour microenvironment (TME) may reveal key biological differences between patient responses and could help us predict therapy response. To explore the TME we applied geospatial whole-transcriptome profiling (NanoString GeoMx Digital Spatial Profiler) and spatial proteomics profiling (Akoya PhenoCycler-Fusion) on a tumour-microarray. The cohort included 25 tumour cores of advanced head and neck tumours, 12 of which were collected from 2 immunotherapy responsive patients (5 cores) and 4 immunotherapy-resistant patients (8 cores). Through mapping the whole transcriptome of tumour and stromal regions, we identified dysregulated metabolic pathways within the tumour, including hypoxia and glycolysis, to be associated with patients who experienced a poor response to immunotherapy. Additionally, when we assessed the relationship of cellular interactions and locations within the TME using single-cell spatial proteomics, the infiltration of CD8 T cells into the tumour mass was higher in patients who had improved response to immunotherapy. Furthermore, we found that the presence of cellular neighbourhoods that were rich in natural killer cells, dendritic cells and macrophages were more common in responsive patients, whereas tumour- dense neighbourhoods were more frequent in non-responsive patients.  By harnessing the power of high-plex spatial transcriptomics and proteomics we identify novel alterations within the TME that could be contributing to immunotherapy resistance and response in HNC.