CASIA OpenIR  > 毕业生  > 博士学位论文
近红外荧光成像术中导航技术研究与多类临床应用探索
何坤山
2019-05-30
页数127
学位类型博士
中文摘要

X线成像、磁共振成像、超声成像等医学影像技术已在术前精确诊断和术后疗效评估等方面发挥了重要的作用。但是在手术过程中,尚缺乏客观有效的成像技术,外科医生主要还是依靠眼睛和手对组织间的色泽、质地、形态进行判别以切除病灶。亟需一种客观有效的成像技术在手术过程中,实时显示病灶的位置和形态,辅助医生完成精准切除。此外,避免神经、血管、尿路和胆管等重要的组织结构受到损伤,有助于实现更好的手术效果和预后。如何在手术过程中精确定位病灶边界,发现微小病灶,识别重要的组织结构,是目前临床实践中存在的挑战性问题。
新兴的光学分子影像技术有望解决这一临床挑战。光学分子影像是一种在体、实时成像新技术,能够真实反映生物体组织内某种基因的表达或大分子的生物学特性,并连续记录和展示疾病相关的分子事件和动力学过程。近红外荧光成像技术利用特定波长的光源激发生物体内的细胞或报告基因或荧光染料等物质,以获取目标区域的荧光信息。该技术具有灵敏度高、特异性强、背景噪声低、穿透深度大等优势,已成为国际研究的热点。本文围绕近红外荧光成像术中导航技术,在成像方法、设备研发和临床应用三个方面开展了全链条、系列化的工作。论文的主要研究内容和创新点如下:
1. 在成像方法方面,针对近红外荧光图像的噪声分布特征,提出了基于小波阈值和模糊C均值聚类的降噪算法,有效地去除了干扰信号,降低了固有噪声,提高了导航设备的分辨率;为了更好地展示荧光图像中所蕴含的信息,提出了基于图像融合的手术导航实时可视化方法,在最大化荧光信息传递效率的同时,最小化了原始解剖结构信息的损失,实现了双通道荧光图像和彩色图像的实时融合和高清显示。
2. 在设备研发方面,针对荧光信号弱且环境光干扰强的问题,建立了基于微弱荧光信号的高灵敏高精度探测数理模型,并设计了大视场大相对孔径高均匀度的光学系统,提高了信号采集能力和抗干扰能力,实现了高信背比和低噪声成像。针对双通道彩色图像和荧光图像同时快速成像的需求,通过创新光路和光学元件以及模块化的系统设计,研制了适用于临床开放和微创手术的高灵敏高分辨术中近红外荧光成像术中导航设备,实现了双通道信号的高帧频同时成像。
3. 在临床应用方面,针对乳腺癌前哨淋巴结活检中传统蓝染法检出率较低的问题,与中国人民解放军总医院、弋矶山医院和汕头大学医学院附属肿瘤医院联合开展了多中心临床试验,共入组了99例乳腺癌病人,将前哨淋巴结的检出率从69.9%提高到了95.5%;针对胸外科手术中交感神经无法准确定位的问题,与北大人民医院联合开展了单中心临床试验,共入组了15例肺结节病人,神经节显影率100%,取得了显著的临床效果。

 

英文摘要

Medical imaging techniques, such as X-ray imaging, magnetic resonance imaging and ultrasound imaging, have played an important role in preoperative accurate diagnosis and postoperative efficacy evaluation. However, in the course of surgery, there is still no objective and effective imaging technology. Surgeons mainly rely on the eyes and hands to distinguish the color, texture and morphology between tissues to remove the lesions. There is an urgent need for an objective and effective imaging technique to display the location and shape of the lesion in real-time during the operation to assist the doctor in performing precise resection. In addition, avoiding damage to important tissue structures such as nerves, blood vessels, urinary tracts, and bile ducts can help achieve better surgical outcomes and prognosis. How to accurately locate the boundary of the lesion during the operation, find small lesions, and identify important tissue structures is a challenging problem in current clinical practice.
The emerging optical molecular imaging technology is expected to solve this clinical challenge. Optical molecular imaging is a new in-vivo and real-time imaging technology, which can truly reflect the expression of a gene or the biological characteristics of a macromolecular, and dynamically record and display molecular events and their dynamic processes. Near infrared fluorescence imaging technology uses a light source of specific wavelength to stimulate cells or reporter genes or fluorescent dyes in the organism to obtain fluorescence information of the target region. The technology has the advantages of high sensitivity, high specificity, low background noise and large penetration depth, and has become a hotspot of international research. This paper focuses on the navigation technology in near-infrared fluorescence imaging, and carries out the whole-chain and series work in three aspects: imaging method, equipment development and clinical application. The main research content and innovation points of this paper are as follows:
1. In the aspect of imaging method, aiming at the noise distribution characteristics of near-infrared fluorescence images, a noise reduction algorithm based on wavelet threshold and fuzzy c-means clustering is proposed, which effectively removes interference signals, reduces the inherent noise and improves the resolution of the device. In order to better display the information contained in fluorescent images, a real-time visualization method of surgical navigation based on image fusion is proposed, which maximizes the efficiency of fluorescence information transmission while minimizing the loss of original anatomical structure information, and realizes real-time fusion and high-definition display of dual-channel color images and fluorescent images.
2. In the aspect of equipment development, aiming at the problem of weak fluorescence signal and strong ambient light interference, we have established a high-sensitivity high-precision detection mathematical model based on weak fluorescence signal, and designed an optical system with large field of view and large relative aperture and high uniformity. The signal acquisition capability and anti-interference ability of the device enable high signal-to-background ratio and low noise imaging. For the simultaneous rapid imaging of color images and fluorescent images, a high-sensitivity high-resolution near-infrared fluorescence image-guided device for clinical open and minimally invasive surgery have been developed through innovative light path and optical components and modular system design. Simultaneous imaging of high frame rate of color images and fluorescent images is realized.
3. In the aspect of clinical application, in response to the low detection rate of traditional blue staining in breast cancer sentinel lymph node biopsy, a multi-center clinical trial was conducted in conjunction with Chinese PLA General Hospital, Yijishan Hospital and the affiliated Cancer Hospital of Shantou University Medical College. A total of 99 breast cancer patients were enrolled, and the detection rate of sentinel lymph nodes was increased from 69.9% to 95.5%. For the problem that sympathetic nerves could not be accurately located in thoracic surgery, a single-center clinical trial was conducted jointly with Peking University People's Hospital. A total of 15 patients with pulmonary nodules were enrolled, and the ganglion detection rate was 100%, which achieved significant clinical results.

 

关键词近红外荧光成像 术中导航 临床应用 前哨淋巴结 胸交感神经
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/25762
专题毕业生_博士学位论文
通讯作者何坤山
推荐引用方式
GB/T 7714
何坤山. 近红外荧光成像术中导航技术研究与多类临床应用探索[D]. 中国科学院自动化研究所. 中国科学院大学,2019.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
近红外荧光成像术中导航技术研究与多类临床(5601KB)学位论文 限制开放CC BY-NC-SA
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[何坤山]的文章
百度学术
百度学术中相似的文章
[何坤山]的文章
必应学术
必应学术中相似的文章
[何坤山]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。