期刊
OPTICS EXPRESS
卷 16, 期 4, 页码 2469-2485出版社
OPTICAL SOC AMER
DOI: 10.1364/OE.16.002469
关键词
-
类别
资金
- NCI NIH HHS [R01-CA75289-11] Funding Source: Medline
- NEI NIH HHS [R01-EY11289-20] Funding Source: Medline
Optical coherence tomography (OCT) is an emerging medical imaging technology which generates high resolution, cross-sectional images in situ, without the need for excisional biopsy. Previous clinical studies using endoscopic OCT with standard 10-15 mu m axial resolution have demonstrated its capability in diagnosing Barrett's esophagus (BE) and high-grade dysplasia (HGD). Quantitative OCT image analysis has shown promise for detecting HGD in Barrett's esophagus patients. We recently developed an endoscopic OCT system with an improved axial resolution of similar to 5 mu m. The goal in this manuscript is to compare standard resolution OCT and ultrahigh resolution OCT (UHR-OCT) for image quality and computer-aided detection using normal and Barrett's esophagus. OCT images of gastrointestinal (GI) tissues were obtained using UHR-OCT (5.5 mu m) and standard resolution OCT (13 mu m). Image quality including the speckle size and sharpness was compared. Texture features of endoscopic OCT images from normal and Barrett's esophagus were extracted using quantitative metrics including spatial frequency analysis and statistical texture analysis. These features were analyzed using principal component analysis (PCA) to reduce the vector dimension and increase the discriminative power, followed by linear discrimination analysis (LDA). UHR-OCT images of GI tissues improved visualization of fine architectural features compared to standard resolution OCT. In addition, the quantitative image feature analysis showed enhanced discrimination of normal and Barrett's esophagus with UHR-OCT. The ability of UHR-OCT to resolve tissue morphology at improved resolution enables visualization of subtle features in OCT images, which may be useful in disease diagnosis. Enhanced classification of image features using UHR-OCT promises to help in the computer-aided diagnosis of GI diseases. (C) 2008 Optical Society of America.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据