胃间质瘤风险分类的CT影像学研究进展

Advances in CT Imaging Research on Risk Stratification of Gastrointestinal Stromal Tumors

摘要: 胃间质瘤(gastrointestinal stromal tumor, GIST)具有明显的生物学行为异质性，术前准确进行危险度分层对于制定个体化治疗方案、改善患者预后具有重要意义，也是消化道间叶源性肿瘤临床诊疗与转化研究的重点之一。CT影像学能够无创、重复性地评估肿瘤的形态学特征并开展定量分析，在胃间质瘤术前危险度分级、鉴别诊断及临床决策中发挥重要作用。本文围绕传统CT形态学评估、影像组学建模、多模态数据融合及人工智能辅助决策等方面，对胃间质瘤风险分类相关影像学研究进展进行综述，系统总结当前研究在标准化、泛化能力及可解释性等方面存在的主要问题，并对多中心前瞻性研究、标准化流程构建以及影像–分子机制关联分析等发展方向进行展望，以期为胃间质瘤精准诊疗体系的完善及其临床转化应用提供参考。

Abstract: Gastrointestinal stromal tumor (GIST) exhibits marked biological heterogeneity. Accurate preoperative risk stratification is of great importance for developing individualized treatment strategies and improving patient prognosis, and it is also one of the major focuses in the clinical diagnosis, treatment, and translational research of gastrointestinal mesenchymal tumors. Computed tomography (CT) can noninvasively and reproducibly evaluate tumor morphological features and perform quantitative analysis, thereby playing an important role in preoperative risk grading, differential diagnosis, and clinical decision-making for GIST. This article reviews recent advances in imaging research related to risk stratification of GIST, with a focus on conventional CT morphologic assessment, radiomics modeling, multimodal data fusion, and artificial intelligence-assisted decision-making. It also systematically summarizes the main limitations of current studies, including issues related to standardization, generalizability, and interpretability, and discusses future directions such as multicenter prospective studies, standardized workflow establishment, and imaging-molecular mechanism association analysis, with the aim of providing a reference for improving the precision diagnosis and treatment system for GIST and promoting its clinical translation.