Suppr超能文献

[基于容积重建技术的虚拟内窥镜检查:技术层面]

[Virtual endoscopy with a volumetric reconstruction technic: the technical aspects].

作者信息

Pavone P, Laghi A, Panebianco V, Catalano C, Giura R, Passariello R

机构信息

Università degli Studi La Sapienza, Roma.

出版信息

Radiol Med. 1998 Jun;95(6):618-23.

PMID:9717545
Abstract

OBJECTIVE

We analyze the peculiar technical features of virtual endoscopy obtained with volume rendering.

MATERIAL AND METHODS

Our preliminary experience is based on virtual endoscopy images from volumetric data acquired with spiral CT (Siemens, Somatom Plus 4) using acquisition protocols standardized for different anatomic areas. Images are reformatted at the CT console, to obtain 1 mm thick contiguous slices, and transferred in DICOM format to an O2 workstation (Silicon Graphics, Mountain View CA, USA) with processor speed of 180 Mhz, 256 Mbyte RAM memory and 4.1 Gbyte hard disk. The software is Vitrea 1.0 (Vital Images, Fairfield, Iowa), running on a Unix platform. Image output is obtained through the Ethernet network to a Macintosh computer and a thermic printer (Kodak 8600 XLS).

RESULTS

Diagnostic quality images were obtained in all the cases. Fly-through in the airways allowed correct evaluation of the main bronchi and of the origin of segmentary bronchi. In the vascular district, both carotid strictures and abdominal aortic aneurysms were depicted, with the same accuracy as with conventional reconstruction techniques. In the colon studies, polypoid lesions were correctly depicted in all the cases, with good correlation with endoscopic and double-contrast barium enema findings. In a case of lipoma of the ascending colon, virtual endoscopy allowed to study the colon both cranially and caudally to the lesion. The simultaneous evaluation of axial CT images permitted to characterize the lesion correctly on the basis of its density values.

DISCUSSION

The peculiar feature of volume rendering is the use of the whole information inside the imaging volume to reconstruct three-dimensional images; no threshold values are used and no data are lost as opposite to conventional image reconstruction techniques. The different anatomic structures are visualized modifying the reciprocal opacities, showing the structures of no interest as translucent. The modulation of different opacities is obtained modifying the shape of the opacity curve, either using pre-set curves or in a completely independent way. Other technical features of volume rendering are the perspective evaluation of the objects, color and lighting. In conclusion, volume rendering is a promising technique to elaborate three-dimensional images, offering very realistic endoscopic views. At present, the main limitation is represented by the need of powerful and high-cost workstations.

摘要

目的

我们分析通过容积再现获得的虚拟内窥镜检查的独特技术特征。

材料与方法

我们的初步经验基于使用针对不同解剖区域标准化的采集协议,从螺旋CT(西门子,Somatom Plus 4)获取的容积数据生成的虚拟内窥镜图像。图像在CT控制台进行重新格式化,以获得1毫米厚的连续切片,并以DICOM格式传输到一台处理器速度为180兆赫兹、内存为256兆字节且硬盘为4.1千兆字节的O2工作站(美国加利福尼亚州山景城的硅图公司)。软件为Vitrea 1.0(美国爱荷华州费尔菲尔德的Vital Images公司),运行在Unix平台上。图像输出通过以太网网络传输到一台苹果电脑和一台热敏打印机(柯达8600 XLS)。

结果

所有病例均获得了诊断质量的图像。气道内的飞行浏览能够正确评估主支气管和段支气管的起源。在血管区域,颈动脉狭窄和腹主动脉瘤均被清晰显示,其准确性与传统重建技术相同。在结肠研究中,所有病例中的息肉样病变均被正确显示,与内镜检查和双对比钡灌肠检查结果具有良好的相关性。在一例升结肠脂肪瘤病例中,虚拟内窥镜检查能够对病变上下方的结肠进行研究。同时评估轴向CT图像有助于根据病变的密度值正确地对其进行特征描述。

讨论

容积再现的独特特征是利用成像容积内的全部信息来重建三维图像;与传统图像重建技术不同,不使用阈值,也不会丢失数据。通过改变相互间的不透明度来显示不同的解剖结构,将不感兴趣的结构显示为半透明。通过使用预设曲线或以完全独立的方式改变不透明度曲线的形状来实现对不同不透明度的调节。容积再现的其他技术特征包括对物体的透视评估、颜色和光照。总之,容积再现是一种用于生成三维图像的有前景的技术,能提供非常逼真的内窥镜视图。目前,主要的局限性在于需要功能强大且成本高昂的工作站。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验