Rapid advances in artificial intelligence (AI) technology are deeply changing clinical diagnosis and treatment decision⁃making model in oncology field. Clinical diagnosis and treatment decision support systems built on big data and machine learning are being implemented across various links such as cancer prevention, diagnosis and treatment, creating new opportunities for oncology. In the process of tumor diagnosis and treatment, AI has demonstrated advantages throughout improving screening and diagnostic capabilities, optimizing therapeutic regimen, and reducing in repetitive clinician workload, which is expected to enhance the standardization and individualization levels of tumor diagnosis and treatment. Nevertheless, multiple challenges remain for AI clinical application in oncology, such as limited model interpretability undermining clinician⁃patient trust, difficulties integrating with existing hospital information systems, persistent “data silos” , and concerns about patient data privacy and security. These issues hinder large⁃scale application of AI in oncology field. This review summarizes the current status and challenges of AI⁃assisted clinical decision⁃making in oncology and proposes strategies to address existing gaps, including fine⁃tuning large language models with disease⁃specific knowledge and clinical guidelines, and building more efficient multimodal, data⁃fusion intelligent⁃assisted decision support systems.

无