摘要胃肠肿瘤是全球面临的重大公共卫生挑战，其高发病率与高死亡率反映当前诊疗策略亟待优化。在这一背景下，人工智能（Al）技术凭借其强大的计算与数据驱动学习能力，在胃肠肿瘤外科领域的应用迅速扩展，覆盖诊断分期、术中导航、疗效及预后预测等关键环节，有望为临床提供精准诊疗新模式。首先在诊断分期上，AI通过内镜、影像与病理的大规模数据训练，提高了早期病变检出率和浸润深度、肿瘤转移判断的准确性；在术中阶段，AI有望自动识别关键解剖结构（如血管、神经、转移病灶）及操作规范，为外科医生提供智能“手术地图”及质量标准；在疗效及预后预测方面，AI可整合多模态异构信息搭建大数据模型，实现患者疗效与复发生存的预测并推进新型药物研发。尽管目前AI模型的泛化性、可解释性及伦理监管等方面仍面临挑战，但随着AI技术更广泛地应用及与临床病例的逐步增量相互交叉验证，AI或将重塑胃肠外科肿瘤的诊疗新模式。

abstractsGastrointestinal tumors represent a significant global public health challenge, characterized by high incidence and mortality, which reflect the urgent need for optimizing current diagnostic and therapeutic strategies. Based on that, artificial intelligence (AI), empowered by advanced computational capacity and data-driven learning, has rapidly extended its application in gastrointestinal surgical oncology. Its role includes diagnosis and staging, intraoperative navigation, and prediction of treatment outcomes and prognosis, drawing the outline of precision-oriented clinical paradigm. For diagnosis and staging, AI models trained on large-scale endoscopic, radiologic, and histopathologic datasets have improved early lesion detection, estimation of tumor invasion depth, and accuracy of metastasis assessment. Intraoperatively, AI is expected to automatically identify critical anatomical structures (e.g., vessels, nerves, metastatic foci) and standardize surgical procedures, thereby equipping surgeons with intelligent "surgical maps" to enhance surgical quality. For outcome prediction, AI can integrate heterogeneous multimodal data to construct large-scale predictive models, thereby enabling individualized forecasts of therapeutic efficacy, recurrence, and survival, while also supporting drug development. Despite challenges in model generalizability, interpretability and ethical oversight, AI is still expected to reshape diagnostic and therapeutic strategies in gastrointestinal surgical oncology, with broader clinical implementation and accumulation of validated cases.