Ting Liu, Jiamao Lin, Yang Li, Yong Wang
Immune checkpoint inhibitors have transformed oncology, yet durable responses remain unevenly distributed, and existing stratification tools-single-site biopsy, PD-L1 immunohistochemistry, and [¹8F]FDG-PET-fail to capture the spatial heterogeneity and dynamic evolution of the tumor immune microenvironment. ImmunoPET addresses these limitations by pairing antibody-based molecular recognition with whole-body quantitative PET, enabling non-invasive mapping of checkpoint expression across the entire disease burden. This review examines the biological rationale and translational status of targets spanning inhibitory axes (PD-1/PD-L1, CTLA-4), next-generation co-inhibitory receptors (LAG-3, TIM-3, TIGIT, VISTA), and co-stimulatory targets (ICOS, 4-1BB, B7-H3), alongside probe engineering principles including scaffold selection, radionuclide pairing, and bioorthogonal pretargeting. Clinical evidence across thoracic, genitourinary, hematological, and neuro-oncological contexts demonstrates that whole-body PET metrics outperform concurrent IHC in predicting treatment outcomes. Theranostic extensions and radiomics applications are further discussed. Standardizing quantitative thresholds and harmonizing acquisition protocols remain the critical steps toward regulatory qualification of ImmunoPET as a companion diagnostic.