Application of AI-Based Endoscopic Ultrasound in Gastrointestinal Stromal Tumors

随着消化内镜的普遍应用胃肠黏膜下隆起时常被发现，病变类型多种多样。由于疾病特点、检查手段本身的局限性及医师间的差异性，诊断及鉴别诊断难度较大。特别对于胃肠道间质瘤，这是一种具有恶性潜能的肿瘤，正确预测其生物学行为和危险度，且以此为依据，为患者制定具有针对性的个体化治疗方案。而超声内镜及人工智能的发展使得准确诊断成为可能，故本综述阐述了超声内镜及其与人工智能的结合在评估胃肠道间质瘤中发展和应用，最后提出未来展望和挑战。With the widespread application of digestive endoscopy, submucosal elevations in the gastrointestinal tract are frequently detected. These lesions vary widely in type, and due to the characteristics of the diseases, the inherent limitations of the examination techniques, and differences among physicians, the identification and differentiation present significant challenges. Especially for gastrointestinal stromal tumors (GISTs), which are neoplasms with malignant potential, clinical practice necessitates accurate prediction of their biological behavior and risk stratification. This serves as the foundation for formulating individualized treatment plans tailored to each patient. The advancements in endoscopic ultrasound (EUS) and artificial intelligence (AI) have enabled accurate diagnosis. This review systematically examines the evolution and clinical utilization of EUS, particularly its synergistic integration with AI in the diagnostic evaluation of GISTs, and critically explores both emerging opportunities and persistent challenges in advancing this interdisciplinary approach.