Cancer diagnostic imaging is undergoing a fundamental transition from conventional anatomical visualization to molecular-targeted and functionally driven profiling of tumor biology. Next-generation imaging no longer merely detects morphological lesions, but decodes the dynamic biological states underlying tumor initiation, progression, immune remodeling, metabolic reprogramming, and therapeutic response. Imaging targets with high specificity, biological accessibility, and functional relevance are primary drivers of this technological transition. Current cancer imaging targets can be categorized into three spatial layers: extracellular biomarkers, membrane-associated targets, and intracellular biomarkers. Progress in nanotechnology, molecular engineering, bioorthogonal chemistry, and artificial intelligence (AI) has greatly facilitated the construction of multimodal, multi-target, and intelligent imaging systems. Nevertheless, clinical translation still faces prominent obstacles, including tumor heterogeneity, temporal biomarker fluctuation, limited tissue penetration, nonspecific background signals from off-target activation, and translational gaps. This review systematically classifies imaging targets based on their spatial distribution and biological function, summarizes the latest advances in cancer diagnosis and imaging, and discusses emerging research directions and future prospects for next-generation precision oncology.