利用拉格朗日追踪减少生理运动对电极位移弹性成像的影响
该研究开发了基于自由手电极位移弹性成像的拉格朗日变形追踪方法,在球状包涵体体模和七组在体数据中与欧拉方法比较应变张量图像。在体结果显示拉格朗日追踪图像一致性更高,分割消融区标准差降低最高达98%,Dice系数提升最高达25%,消融区可视化成功率从约50%提升至近100%。
要点
- 拉格朗日变形追踪可减少生理运动导致的散斑图案变异,使消融区显示更一致。
- 在体数据中,拉格朗日应变张量图像分割消融区的标准差较欧拉方法降低最高达98%。
- 拉格朗日方法使消融区可视化的Dice系数提升最高达25%,成功率从约50%升至近100%。
诊断标准与关键数值
- 在体数据中,拉格朗日应变张量图像分割消融区的标准差较欧拉应变张量图像降低最高达98%。
- 拉格朗日应变张量图像的Dice系数较欧拉方法提升最高达25%。
- 消融区可视化成功率从约50%提升至近100%。
局限性
- 原文未报告局限性。
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
Minimally invasive treatments such as microwave ablation (MWA) have been growing in popularity for extending liver cancer survival rates in patients, when surgery is not an option. As a non-ionizing, real-time alternative to contrast-enhanced computed tomography, electrode displacement elastography (EDE) has shown promise as an imaging modality for MWA. Despite imaging efficacy, motion artifacts caused by physiological motion result in unintended speckle pattern variance, thereby inhibiting consistent and accurate ablated region visualization. To combat these unavoidable motion artifacts, a Lagrangian deformation tracking (LDT) approach based on freehand EDE was developed to track tissue movement and better define tissue properties. For validating LDT efficacy, a spherical inclusion phantom as well as seven in vivo data sets were processed, and strain tensor images were compared with identical time sampled images estimated using a traditional Eulerian approach. In vivo results revealed greater consistency among visualized LDT strain tensor images, with segmented ablated regions exhibiting standard deviation reductions of up to 98% when compared with Eulerian strain tensor images. Additionally, Lagrangian strain tensor images provided Dice coefficient improvements up to 25%, and success rates improved from approximately 50% to nearly 100% for ablated region visualization.
引用
- Pohlman RM, Varghese T. Physiological Motion Reduction Using Lagrangian Tracking for Electrode Displacement Elastography. Ultrasound Med Biol. 2019. PMID: 31806499. doi:10.1016/j.ultrasmedbio.2019.11.001
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。