【作者】 肖嘉莹；

【导师】 汤井田； 蒋华北；

【作者基本信息】 中南大学 ， 生物医学工程， 2011， 博士

【摘要】 骨关节炎(Osteoarthritis,简称OA),又称退行性关节炎,是一种常见的关节炎症。据世界卫生组织统计,50岁以上人群中,骨关节炎的发病率为50%。目前,虽无可以根治该病的方法,但临床研究表明,如果在骨关节炎早期及早发现并采用手术或者药物方法进行医治,就可以使患病关节的病情得到改善。本文尝试采用一种新的生物医学成像方法,即光声成像(PAT,Photoacoustic Tomography),来进行骨关节炎的早期诊断。它不仅具有光学成像各组织间高对比度的特点,而且能够提供超声成像所具有的高分辨率,更重要的是,相比于其它现有的生物医学成像方法,它对关节内的早期病变更加敏感,从而使得骨关节炎早期的诊断效率得到极大提高。光声成像的要点在于,它采用一套有效的算法,使得人体关节内不同组织按光学特性得到很好地分辨。本文提出并实现了一套以有限元和牛顿迭代法为基础的单波长和多波长非线性定量重建算法,来重建人体关节组织内的光学参数,以及如血红蛋白浓度、水含量、声速等生理或功能性参数。相比于现有的其他重建算法,这套算法在于将光声传播方程和光扩散模型相结合,所以能够提供别的重建算法所没有的光吸收系数的重建,从而给出定量的结果,这是这套算法的独特优点。此外,这套算法还消除了普通光声算法中声速均匀的假设,因此除了可以得到和组织内各种光吸收物质有关的生理或功能性参数,还可以得到声速的分布,为临床诊断多提供了一个额外的依据。本文从光声成像的原理出发,详细地推导了单波长和多波长定量重建算法,对多波长算法的最佳成像条件进行了讨论,并分别对单波长和多波长算法进行了一系列的数值模拟及实验模拟,从理论和实验上验证了这套算法的有效性和精确性。最后,将单波长和多波长算法应用于临床中,对骨关节炎患者和正常人的临床数据进行了图像重建,得到了关节中的光吸收系数、血氧浓度、水含量、声速分布等各种生理或功能性图像。临床数据表明,本文所实现的单波长和多波长算法可以有效的区分患有骨关节炎的关节和正常关节。本文首次提出了基于有限元的单波长和多波长定量光声重建算法,并且将该算法应用于骨关节炎的临床诊断中,填补了光声成像领域在人体硬组织成像的空白。临床数据重建表明,这种方法很有可能成为早期手指骨关节炎诊断的强有力工具,从而为骨关节炎诊断提供科学、定量的参考依据。更多还原

【Abstract】 Osteoarthritis (OA) is a common degenerative, slowly progressive joint disease. According to the World Heath Organization (WHO), approximately 50% of people over 50 year-old suffer from OA today. While there is currently no cure for this disease, numerous clinical studies have shown that the progression of the diseased joint may be modified by medical or surgical intervention if the disease is detected early.The goal of this thesis research is to develop a novel biomedical imaging modality called photoacoustic tomography (PAT), which promises to significantly advance our ability to detect such diseased joints at an early stage, and thereby offer the potential for curative OA treatment for this devastating disease. PAT is a newly developed in vivo imaging method for biomedical tissues, which uniquely combines high optical contrast and high ultrasound resolution in a single modality. In addition, it offers better sensitivity to the tissue changes in an OA joint at an early stage over all the cuurent medical imaging modalities.The key to the development of successful PAT for joint imaging is an effective image reconstruction algorithm that allows one to characterize the joint tissues. In this thesis we have developed and implemented finite element based reconstruction algorithms for recovering tissue optical property, functional/ physiological parameters such as hemoglobin concentration and water content and acoustic property using both single-and multi-spectral photoacoustic measurements. A unique advantage of our reconstruction method is that it allows us to obtain the distribution of absolute optical absorption coefficient by incorporating a light transport model into our PAT reconstruction framework to realize quantitative PAT. This reconstruction method also has eliminated the assumption of homogenous acoustic velocity in the image domain so that it allows direct recovery of tissue chromophore concentrations and acoustic velocity using tomographic multi-spectral photoacoustic measurements.Extensive simulations and laboratory experiments using tissue-like phantoms have been conducted using both single- and multi-spectral photoacoustic data in order to assess the overall imaging capabilities of the reconstruction algorithms developed. Both the single- and multi-spectral reconstruction algorithms have also been tested and evaluated using pilot clinical data from patients with hand OA and healthy volunteers.To our knowledge, the quantitative reconstruction algorithms developed represent the first of their kind. Their applications to osteoarthritis imaging as well as hard tissue imaging are also for the first time. The successful pilot clinical experience indicates that the single- and multi-spectral quantitative PAT methods developed in this thesis have great potential to serve as a clinical tool for early detection of osteoarthritis in the hand.更多还原