WFUMB肝脏超声融合成像技术综述与立场声明
本文为世界超声医学与生物学联合会(WFUMB)两部分立场声明的技术部分,系统综述2012年1月至2025年1月发表的CT/MRI与超声融合成像研究。通过PICO框架评估电磁追踪融合的目标配准误差及对常规B型超声不显影肝脏病灶的可见性改善。结论为超声融合成像可显著提高肝脏介入操作的空间准确性,安全性良好,并发症率低。
要点
- 对B型超声不显影病灶,融合成像可提高检出率并指导消融,操作成功率可达90%–95%。
- 临床研究中目标配准误差约为4–14 mm,报告时应意识到与体模理想值的差距。
- 自动配准与手动配准准确性相当且更快,可减少操作者依赖性。
核心结论
- 超声融合成像通过将实时超声与CT/MRI数据对齐,显著提高肝脏介入操作的空间准确性。
- 与单纯常规超声相比,超声融合成像可改善介入引导效果,并维持较低的并发症概况。
研究设计与样本
- 本文为WFUMB两部分立场声明的技术部分,采用系统综述结合专家共识的方法。
- 检索PubMed、Scopus、Embase和IEEE Xplore数据库,时间范围为2012年1月至2025年1月,辅以手动引文追踪和AI辅助查询生成。
- 采用PICO框架,围绕电磁追踪融合目标配准误差和融合对困难显示肝脏病灶可见性改善两个核心问题筛选文献。
关键影像表现/征象
- 融合成像可改善常规B型超声上不显影肝脏病灶的可检测性,使病灶得以显示并用于引导介入操作。
- 超声融合成像对困难部位肿瘤治疗可能尤其有益,如尾状叶病灶和胆管周围病灶。
- 原中文中未报告具体病灶的超声回声特征或CT/MR信号密度表现,无法归纳更多影像征象。
诊断标准与关键数值
- 体模理想条件下目标配准误差约为1–3 mm。
- 临床研究中目标配准误差约为4–14 mm。
- 融合成像引导介入操作在高达90%–95%的病例中获得成功,对应常规B型超声病灶不显影或显示困难的情况。
- 多研究安全性概况良好,主要并发症发生率普遍低于2%。
- 自动配准方法速度更快,准确性与手动配准相当,可能减少操作者依赖性。
临床应用与阅片要点
- 当常规B型超声无法充分显示肝脏病灶时,可考虑采用基于CT或MRI的超声融合成像辅助定位和引导。
- 对位于尾状叶或胆管周围等困难部位的肿瘤,融合成像可能具有额外临床价值。
- 阅片和报告中应关注配准误差,临床条件下的目标配准误差可明显大于体模理想值。
局限性
- 原文为技术综述与立场声明,纳入研究的异质性、研究设计和偏倚风险未在摘要中量化报告。
- 临床研究的目标配准误差明显高于体模理想条件,实际应用中可能受患者呼吸、体位和设备因素影响。
- 人工智能和增强现实优化配准工作流的结论属于预期方向,并非基于已完成临床研究的确定性结论。
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
OBJECTIVE: Ultrasound fusion imaging is a hybrid technique that combines real-time ultrasonography (US) with pre-acquired computed tomography (CT) or magnetic resonance imaging (MRI), using electromagnetic (EM) tracking to enable precise spatial correlation between modalities. This technology is increasingly used for liver imaging and interventions, especially when conventional B-mode US fails to provide adequate lesion visualization. The aim of this technical review and position statement is to evaluate the technical accuracy (target registration errors) and lesion visibility of ultrasound fusion imaging based on published evidence and expert consensus.
METHODS: This manuscript was designed as the technical component of a two-part World Federation for Ultrasound in Medicine and Biology (WFUMB) position statement. A systematic review was conducted using a PICO framework focused on two core questions: (i) to evaluate EM-tracked fusion target registration errors (PICO T1), and (ii) whether fusion improves visibility of lesions in difficult-to-image liver lesions (PICO T2). Literature from January 2012 to January 2025 was searched across PubMed, Scopus, Embase, and IEEE Xplore, with manual citation tracking and AI-assisted query generation. Eligible studies included research on US/CEUS fusion with CT/MRI, reporting technical accuracy and lesion conspicuity.
RESULTS: The technical accuracy of fusion imaging was consistently high, with target registration errors (TRE) of ∼1-3 mm in ideal phantom settings and ∼4-14 mm in clinical studies. Automatic registration methods were faster and similarly accurate as manual registration, possibly reducing operator dependence. Fusion imaging improved the detectability of lesions not visible (occult) on conventional B-mode US, increasing the diagnostic yield and enabling successful interventions (e.g., ablation) in up to 90%-95% of cases. Safety profiles across studies were favorable, with major complication rates generally below 2%. Furthermore, fusion imaging might prove especially beneficial for treating tumors in difficult locations (e.g., caudate lobe, peribiliary lesions).
CONCLUSION: Ultrasound fusion imaging significantly enhances the spatial accuracy of liver interventions by aligning real-time US with CT/MRI datasets. It improves interventional procedures guidance and maintains a low complication profile as compared to conventional US alone. Advancements in artificial intelligence (AI) and augmented reality (AR) are expected to further optimize image co-registration workflows and clinical outcomes. This technical review supports the broader adoption of fusion imaging as a key tool in liver imaging and intervention.
引用
- Săftoiu, Ewertsen, Popescu, et al. WFUMB Liver Ultrasound Fusion Imaging Technical Review and Position Statement: Focus on CT/MRI-Based Fusion. Ultrasound Med Biol. 2026. PMID: 42471281. doi:10.1016/j.ultrasmedbio.2026.05.026
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。