[Current Imaging Approaches for Pediatric Brain Tumors].

After leukemia, pediatric brain tumors are the second most common childhood malignancies and are associated with significant neurologic and developmental sequelae, rendering accurate and early diagnosis critical. The 2021 World Health Organization classification for central nervous system tumors emphasizes the utility of molecular pathology in distinguishing between pediatric and adult brain tumors. Imaging studies, including magnetic resonance imaging (T1-, T2-, diffusion-, and perfusion-weighted imaging, diffusion tensor imaging, and magnetic resonance spectroscopy) and positron emission tomography using fluorodeoxyglucose or amino acid tracers, integrate structural assessment with functional and quantitative techniques, enabling the evaluation of cellularity, perfusion, and metabolism. Characteristic imaging patterns support diagnosis and prognosis across tumor types, including diffuse gliomas (adult- and pediatric-type, low- and high-grade subtypes), circumscribed astrocytic gliomas, ependymomas, glioneuronal tumors, choroid plexus tumors, embryonal tumors (e.g., medulloblastoma), pineal region tumors, craniopharyngiomas, nerve sheath tumors, germ cell tumors, meningiomas, Langerhans cell histiocytosis, hamartomas, and cavernous malformations. Molecular features increasingly guide treatment strategies. Emerging technologies, such as radiomics and artificial intelligence (AI), are improving tumor classification, segmentation, and recurrence prediction, with advances such as federated learning and explainable AI supporting privacy-preserving and interpretable models. Imaging also plays roles beyond detection, including surgical planning, treatment monitoring, and prognostication. Future integration of multimodal imaging and AI is expected to improve precision, standardization, and individualized pediatric neuro-oncology care through rapid, noninvasive diagnostics.