Spatial single-cell landscaping van de immuun micro-omgeving en melanoom cel heterogeniteit in metastaserende en niet-metastaserende vroeg stadium melanomen

The rising global incidence of melanoma, particularly early-stage melanomas (ESMs), highlights the urgent need for precision oncology approaches. ESMs, defined by a Breslow thickness of <1 mm, are typically managed through surgical excision alone due to their perceived low metastasis rates. However, approximately 10.23\% of ESM cases progress to metastasis, underscoring the need for reliable biomarkers to predict disease progression. Current diagnostic limitations arise from the small size of ESM lesions, reliance on formalin-fixed, paraffin-embedded (FFPE) tissue, and the inability of bulk transcriptomics to isolate invasive tumor components responsible for metastasis. Advances in spatial omics technologies now enable high-resolution, single-cell analyses from FFPE samples, offering unprecedented insights into the tumor microenvironment (TME). Nevertheless, adopting spatial multiomics technologies presents unique challenges, particularly ensuring data reliability. Particularly, for imaging based spatial omics technologies, While not all spatial-omics approaches are imaging-based, the majority rely on imaging data, where high-quality, artifact-free outputs are essential for robust and accurate analyses. Similar to other immunofluorescence assays, spatial proteomics is prone to imaging artifacts, including air bubbles, out-of-focus areas, tissue folds, antibody aggregates, and dust particles. To address these challenges, this thesis introduces QUALIFAI, a deep learning-based automated artifact detection and removal tool. QUALIFAI demonstrates high classification and segmentation accuracy, providing a reliable solution to remove immunofluorescence imaging artifacts and enhancing the integrity of downstream analyses. In the context of ESMs, we applied our in-house spatial proteomics method - Multiplexed Immunohistochemistry Labelling by Antibody Neodeposition (MILAN) - to investigate the immune landscape and tumor heterogeneity. Our analysis identified —AQP1+ melanoma and intermediate melanoma cells—as key melanoma cell subtypes associated with metastasis. Furthermore, metastasizing ESMs exhibited a reduction in antigen-presenting cDC1 cells, underscoring the complex interplay between tumor cells and the immune microenvironment. These findings suggest that communities of specific melanoma phenotypes and inflammatory cells could serve as predictive biomarkers for ESM progression, thereby refining clinical stratification and informing therapeutic decision-making. Overall, this thesis offers an explorative, spatial overview of cellular communities with prognostic potential in ESM.