基于MRI的人工智能模型术前预测肝细胞癌微血管侵犯的系统评价与Meta分析
该系统评价与Meta分析纳入29项研究,汇总2838例内部验证和1161例外部验证病例,评估基于MRI的人工智能模型术前预测肝细胞癌微血管侵犯的性能。内部验证汇总敏感度0.81、特异度0.82、AUC 0.88,外部验证AUC 0.85;传统机器学习敏感度高于深度学习,影像组学联合临床特征优于单纯影像组学。但存在显著异质性且GRADE证据等级低。
要点
- 基于MRI的AI模型术前预测肝细胞癌微血管侵犯内部验证AUC 0.88、外部验证AUC 0.85。
- 传统机器学习敏感度在内部和外部验证中均高于深度学习(均P<0.05)。
- 影像组学联合临床特征模型的敏感度和特异度均优于单纯影像组学模型(P<0.01)。
核心结论
- 基于MRI的AI在肝细胞癌微血管侵犯术前预测中表现出较高性能,尤其是结合多模态影像和临床变量的模型,但显著异质性和低GRADE证据等级可能影响证据强度。
研究设计与样本
- 系统评价与Meta分析,遵循PRISMA指南检索PubMed、Embase和Web of Science至2025年5月,纳入29项使用MRI-based AI模型且经组织病理学确诊微血管侵犯的研究,采用QUADAS-2工具和GRADE框架评估质量,使用双变量随机效应模型进行统计合成。
诊断标准与关键数值
- 内部验证汇总敏感度0.81(95%CI 0.76-0.85)、特异度0.82(95%CI 0.78-0.85)、诊断比值比19.33(95%CI 13.15-28.42)、AUC 0.88(95%CI 0.85-0.91);外部验证AUC 0.85。
临床应用与阅片要点
- 原文未据此给出推荐。
局限性
- 存在显著异质性。
- GRADE证据等级低,可能影响证据强度。
- 需要方法学标准化和多中心前瞻性验证以确保临床适用性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
PURPOSE: To evaluate the performance of magnetic resonance imaging (MRI)-based artificial intelligence (AI) in the preoperative prediction of microvascular invasion (MVI) in patients with hepatocellular carcinoma (HCC).
METHODS: A systematic search of PubMed, Embase, and Web of Science was conducted up to May 2025, following PRISMA guidelines. Studies using MRI-based AI models with histopathologically confirmed MVI were included. Study quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool and the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. Statistical synthesis used bivariate random-effects models.
RESULTS: Twenty-nine studies were included, totaling 2838 internal and 1161 external validation cases. Pooled internal validation showed a sensitivity of 0.81 (95% CI: 0.76-0.85), specificity of 0.82 (95% CI: 0.78-0.85), diagnostic odds ratio (DOR) of 19.33 (95% CI: 13.15-28.42), and area under the curve (AUC) of 0.88 (95% CI: 0.85-0.91). External validation yielded a comparable AUC of 0.85. Traditional machine learning methods achieved higher sensitivity than deep learning approaches in both internal and external validation cohorts (both P < 0.05). Studies incorporating both radiomics and clinical features demonstrated superior sensitivity and specificity compared to radiomics-only models (P < 0.01).
CONCLUSIONS: MRI-based AI demonstrates high performance for preoperative prediction of MVI in HCC, particularly for MRI-based models that combine multimodal imaging and clinical variables. However, substantial heterogeneity and low GRADE levels may affect the strength of the evidence, highlighting the need for methodological standardization and multicenter prospective validation to ensure clinical applicability.
引用
- Han X, Shan L, Xu R, et al. Assessing MRI-based Artificial Intelligence Models for Preoperative Prediction of Microvascular Invasion in Hepatocellular Carcinoma: A Systematic Review and Meta-analysis. Acad Radiol. 2025. PMID: 40670226. doi:10.1016/j.acra.2025.06.030
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。