Spatial omics technologies have transformed biomedical research by enabling unprecedented insight into cellular behaviour within its native tissue microenvironment at diverse spatial resolutions. These advancements have facilitated the identification of spatial communities—groups of neighbouring cells that interact and coordinate to drive essential biological processes. However, there is an increasing need for methods that can integrate multi-omics data or combine samples from different platforms to robustly identify these novel spatial communities. In this study, we conduct a systematic evaluation of existing spatial community detection tools and present a novel approach designed to reliably detect communities across multiple cancer samples. In skin cancer, our findings reveal communities of similar cell type compositions that are conserved across Visium, CosMx, and Xenium data. We also observe consistent trends in communities across basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and melanoma samples. Notably, the tumour-enriched community is significantly expanded in melanoma compared to BCC and SCC. Furthermore, within the CosMx data, we detected a substantial increase in both the number and strength of cell-cell interactions within the tumour-enriched community, particularly among fibroblasts, natural killer (NK) cells, Langerhans cells (LC), and melanoma cells. These interactions are driven by key ligand-receptor pairs, including collagen, vitronectin, and macrophage migration inhibitory factor (MIF) pathways, which are known to contribute to tumour progression. Our findings provide a deeper understanding of the spatial organization and microenvironmental dynamics within skin cancer, offering new insights into disease mechanisms and potential therapeutic targets. We anticipate broad applications of our methods beyond skin cancer.