Abstract
Cancer prognosis relies not only on genetic and molecular biomarkers but also on the spatial organization of tumor and immune cells within the tumor microenvironment. Recent advances in spatial biology, particularly hyperplex immunofluorescence (IMF) imaging, have enabled high-dimensional, quantitative assessment of cell-cell interactions at the protein level. Nearest neighbor analysis (NNA) and proximity analysis have emerged as crucial computational methods for quantifying spatial distributions of tumor, stromal, and immune cells in hyperplex IMF datasets, providing insights into tumor heterogeneity, immune infiltration, and treatment response. This review explores the current state of nearest neighbor and proximity analysis in cancer research, focusing on their applications in prognosis using single-cell spatial proteomics data generated by hyperplex IMF imaging. We summarize key computational approaches, including nearest neighbor distance metrics, Ripley's K-function, Voronoi tessellation, and graph-based models, that characterize spatial architecture within the tumor microenvironment. We highlight recent applications of hyperplex IMF in cancers showcasing how spatial proteomic signatures improve prognostic models. Furthermore, we discuss the integration of machine learning and AI-driven methods to leverage these spatial features for predictive modeling. Despite significant progress, challenges remain, including standardization of methodologies, variability in imaging technologies, and the need for large-scale, high-quality datasets. Addressing these challenges could lead to more accurate risk stratification and personalized treatment strategies. By providing a comprehensive overview of nearest neighbor and proximity analysis in the context of hyperplex IMF-based spatial proteomics, this review aims to bridge the gap between computational methodologies and clinical applications, offering new perspectives on how spatial organization at the protein level influences cancer prognosis.