Yan Si-Yan, Fu Xin-Yu, Yang Yan, Jia Liu-Yi, Liang Jia-Wei, Li Ying-Hui, Yan Ling-Ling, Zhou Ying, Zhou Xian-Bin, Li Shao-Wei, Mao Xin-Li
Taizhou Hospital of Zhejiang Province, Zhejiang University, Linhai, Zhejiang, China.
Department of Gastroenterology, Taizhou Hospital of Zhejiang Province Affiliated to Wenzhou Medical University, Linhai, Zhejiang, China.
Best Pract Res Clin Gastroenterol. 2025 Mar;75:102004. doi: 10.1016/j.bpg.2025.102004. Epub 2025 Mar 28.
Esophageal squamous cell carcinoma (ESCC) remains a significant global health burden with high incidence and mortality rates, particularly in developing regions. Early detection is crucial for improving patient survival, yet conventional screening methods such as endoscopy and non-endoscopic techniques face limitations in accuracy, cost, and dependency on clinician expertise. This review explores the transformative role of artificial intelligence (AI) in ESCC screening. AI technologies, including machine learning, deep learning, and transfer learning, demonstrate remarkable potential for early ESCC screening by targeting high-risk populations, optimizing screening modalities, refining screening intervals, and enhancing cost-effectiveness. AI-driven systems improve lesion detection, vascular pattern recognition, and risk prediction by integrating imaging, genomic, and clinical data. Additionally, AI applications in liquid biopsy analysis enable non-invasive detection of circulating tumor cells and DNA, further advancing early diagnosis. Despite these advancements, challenges such as dataset variability, model generalizability, algorithm transparency, and ethical and legal concerns require resolution to fully harness AI's capabilities. This paper highlights the current applications, persistent challenges, and future directions for AI in revolutionizing ESCC screening.
食管鳞状细胞癌（ESCC）仍然是一个重大的全球健康负担，其发病率和死亡率都很高，尤其是在发展中地区。早期检测对于提高患者生存率至关重要，然而，诸如内窥镜检查和非内窥镜技术等传统筛查方法在准确性、成本以及对临床医生专业知识的依赖方面都存在局限性。本综述探讨了人工智能（AI）在ESCC筛查中的变革性作用。包括机器学习、深度学习和迁移学习在内的人工智能技术，通过针对高危人群、优化筛查方式、细化筛查间隔以及提高成本效益，在早期ESCC筛查中显示出巨大潜力。人工智能驱动的系统通过整合成像、基因组和临床数据，改善病变检测、血管模式识别和风险预测。此外，人工智能在液体活检分析中的应用能够对循环肿瘤细胞和DNA进行非侵入性检测，进一步推动早期诊断。尽管取得了这些进展，但诸如数据集变异性、模型通用性、算法透明度以及伦理和法律问题等挑战仍需要解决，以充分发挥人工智能的能力。本文重点介绍了人工智能在革新ESCC筛查方面的当前应用、持续挑战和未来方向。
