肺灌洗三维超声微泡灌注成像与动态传导模式评价方法及其应用

Aiming at the problems that the existing lung lavage technique can not provide timely on-site monitoring and evaluation, and effectively guide the lung lavage process and improve the efficiency of lung lavage, based upon the research achievements of three aspects including single and multiple microbubble dynamics in liquid/tissue, contrast detection and monitoring imaging and parametric perfusion imaging, this project proposes and investigates: (1) fast three-dimensional ultrasound microbubble contrast perfusion imaging method with high contrast and high resolution in small volume, which is based on two-dimensional compressive adaptive beamforming and pulse inversion bubble wavelet transform technique, and is used for real-time monitoring of lung lavage process; (2) dynamic conduction model between contrast agent perfusion model and single point in imaging region and dynamic conduction model among multiple points in imaging region, which are based on the precise extraction technique of single-voxel TIC curve with high signal-to-clutter ratio; (3) verification of feasibility and effectiveness of three-dimensional ultrasound microbubble contrast perfusion imaging and dynamic conduction model analysis and evaluation method in lung lavage, which is based on animal model of living pneumoconiosis. This project makes on-site ultrasound monitoring and evaluation, and effective feedback and guidance possible, and also promotes the application of ultrasound microbubble contrast imaging technique in clinical lung lavage and the development of occupational pneumoconiosis treatment to a higher level.

针对现有肺灌洗技术难以对灌洗过程做出及时的现场监控评价，无法有效指导肺灌洗过程并提高灌洗效率等问题，本项目在对液体/组织内微泡动力学、造影微泡检测与监控成像及灌注参量成像这三个方面取得的成果的基础上，提出和研究：（1）发展基于二维压缩自适应波束合成和脉冲逆转微泡小波变换技术的肺灌洗中高对比、高分辨的小区域三维超声微泡造影灌注参量快速成像方法，以实时监控肺灌洗过程；（2）以高信杂比的单体素TIC曲线提取技术为基础来发展肺灌洗过程中造影剂灌注模式和成像区域内单点之间以及成像区域内多点之间的动态传导模式，以定量评价灌洗程度；（3）基于活体尘肺犬动物模型来证实肺灌洗三维超声微泡造影灌注成像与动态传导模式分析评价方法的可行性和有效性。使肺灌洗现场超声监控评价及有效反馈与指导成为可能，也将推进超声微泡造影成像技术在临床肺灌洗中的应用，促进职业尘肺病治疗向更高层次发展。

肺部疾病诊断及灌洗过程中的超声监控成像、损伤定量评价及造影灌注评价技术是目前该研究领域的重点和难点，本项目主要研究内容和进展包括：（1）设计并制作了多种用于定性及定量估计的肺部仿体模型，实验证明海绵滴水仿体和玻璃珠-明胶仿体可以较好地用于肺部B线检测和损伤定量估计。（2）研究了基于超声特征参数提取的肺部B线检测方法，特征参数包括排列熵、信息熵、模式熵、Nakagami参量、峰度、偏度、能量、平均频率等，研究结果表明在B线和非B线区域这些特征参数均有显著性差异，信息熵参量具有最高的B线检测性能，检测精度达91.7%，AUC达0.90；研究了基于非线性SVM分类方法的双参量B线识别技术，发现排列熵+信息熵组合具有最佳识别效果，B线识别灵敏度达90.521%，特异性达98.106%，准确率达96.328%，AUC达0.95；同时，研究了基于补偿增益的MAD和MSD肺部B线和损伤成像技术，发现基于射频数据分析可以对肺部损伤进行成像，基于包络提取后数据分析可以对B线区域进行成像。（3）研究了回波峰值时间法和射频相关法来定量估计肺部损伤深度，研究结果表明估计的损伤深度与肺部损伤具有正相关性，该技术为肺部损伤定量评价提供了新思路。（4）研究了基于脉冲逆转的宽带次谐波微泡成像技术，将图像CTR最大提高了9.1 dB；研究了平面波条件下的基于奇异值分解空时滤波和基于MCMCDA微泡跟踪算法的造影微泡成像技术，得到了深部组织高灵敏度、高分辨率的造影成像结果；研究了基于对数Nakagami参量编码的超声造影成像及灌注评价技术，该技术可以补偿非均匀声场下血流动力学评价时产生的估计误差。（5）扩展研究了超声造影微泡的次谐波压力估计技术，证实了次谐波的产生与环境压力成负相关性，在环境压力为20-160 mmHg时，灵敏度达0.24 dB/mmHg，为后续肺部血管压力估计的研究提供了支持。