Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group
Journal Home About Issues in Progress Current Issue All Issues Feature Issues

Effects of axial resolution improvement on optical coherence tomography (OCT) imaging of gastrointestinal tissues

Yu Chen, Aaron D. Aguirre, Pei-Lin Hsiung, Shu-Wei Huang, Hiroshi Mashimo, Joseph M. Schmitt, and James G. Fujimoto

Open Access Open Access

Abstract

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 µm 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 ~5 µm. The goal in this manuscript is to compare standard resolution OCT and ultrahigh resolution OCT (UHR-OCT) for image quality and computeraided detection using normal and Barrett’s esophagus. OCT images of gastrointestinal (GI) tissues were obtained using UHR-OCT (5.5 µm) and standard resolution OCT (13 µm). 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.

©2008 Optical Society of America

Full Article  |  PDF Article
More Like This
Morphological analysis of optical coherence tomography images for automated classification of gastrointestinal tissues

P. Beatriz Garcia-Allende, Iakovos Amygdalos, Hiruni Dhanapala, Robert D. Goldin, George B. Hanna, and Daniel S. Elson
Biomed. Opt. Express 2(10) 2821-2836 (2011)

Ultrahigh resolution optical biopsy with endoscopic optical coherence tomography

Paul R. Herz, Yu Chen, Aaron D. Aguirre, James G. Fujimoto, Hiroshi Mashimo, Joseph Schmitt, Amanda Koski, John Goodnow, and Chris Petersen
Opt. Express 12(15) 3532-3542 (2004)

Image analysis for classification of dysplasia in Barrett’s esophagus using endoscopic optical coherence tomography

Xin Qi, Yinsheng Pan, Michael V. Sivak, Joseph E. Willis, Gerard Isenberg, and Andrew M. Rollins
Biomed. Opt. Express 1(3) 825-847 (2010)

View More...

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (12)
Fig. 1.
Fig. 1. The schematic of the OCT system. (LCS: low-coherence source; BS: 90/10 fiber based beam splitter; PC: polarization controller; AGC: air-gap coupling; DCG: dispersion compensating glass; PBS: polarization beam splitter; D: detector; BPF: band-pass filter; DEMOD: amplitude demodulator; Σ: vector summation).

Download Full Size | PDF

Fig. 2.
Fig. 2. The spectrum (A) and point spread function (B) of both superluminescent diode (SLD) for standard resolution imaging and Cr:Forsterite laser source for ultrahigh-resolution imaging.

Download Full Size | PDF

Fig. 3.
Fig. 3. (A) Diagram illustrating the partition of NxN 2D DFT of OCT image into 4 regions for texture analysis (Fx and Fz are the spatial frequencies in the lateral and axial dimensions, respectively). (B) Representative region of interest (ROI) from an EOCT image of normal esophagus (500 µm×500 µm, 256×256 pixels) and (C) its 2D DFT. (D) Representative region of interest (ROI) from an EOCT image of Barrett’s esophagus (500 µm×500 µm, 256×256 pixels) and (E) its 2D DFT.

Download Full Size | PDF

Fig. 4.
Fig. 4. (A) Diagram illustrating the 3×3 neighborhood of pixels used to compute CSAC measures. (B) Representative region of interest (ROI) from an EOCT image of normal esophagus (500 µm×500 µm, 256×256 pixels) and (C) Barrett’s esophagus (500 µm×500 µm, 256×256 pixels). (D)-(I) Histograms of the distribution of local CSAC measurements.

Download Full Size | PDF

Fig. 5.
Fig. 5. Ex vivo imaging of a gastric nodule using UHR- and standard resolution OCT. (A) UHR-OCT image with 4x enlargement in (B); (C) Standard resolution OCT image with 4x enlargement in (D). UHR-OCT image shows reduced speckle size and resolves finer morphological features. Ex vivo imaging enables UHR and standard resolution OCT image the same tissue region exactly.

Download Full Size | PDF

Fig. 6.
Fig. 6. Representative axial scans from UHR- and standard resolution OCT image (indicated by the red-dotted lines in Fig. 5). (A) Axial scan from UHR-OCT image with zoomed region in (B). (C) Axial scan from standard resolution OCT image with zoomed region in (D). A-scan of UHR-OCT reveals finer patterns in tissue structure as shown by sharper modulation pattern compared to standard resolution OCT (red oval).

Download Full Size | PDF

Fig. 7.
Fig. 7. Comparison of UHR- and standard resolution OCT images of a biopsy specimen of Barrett’s esophagus ex vivo. (A) Standard resolution OCT image with 4x enlargement in (D); (B) UHR-OCT image with 4x enlargement in (E); (C) Corresponding histology (H&E, original mag. x40) with 4x enlargement in (F). UHR-OCT image shows reduced speckle size and visualizes fine morphological features more clearly than standard resolution.

Download Full Size | PDF

Fig. 8.
Fig. 8. In vivo endoscopic OCT imaging of normal esophagus using UHR- and standard resolution OCT. (A) UHR-OCT image with 4x enlargement (B); (C) Standard resolution OCT image with 4x enlargement (D). UHR-OCT image shows reduced speckle size and resolves finer morphological features (indicated by arrows).

Download Full Size | PDF

Fig. 9.
Fig. 9. Representative axial scans from OCT images (indicated by the red-dotted lines in Fig. 8). (A) Axial scan from UHR-OCT image with zoomed region in (B). (C) Axial scan from standard resolution OCT image with zoomed region in (D). A-scan of UHR-OCT reveals finer patterns in tissue structure as shown by sharper modulation pattern compared to standard resolution OCT (red oval).

Download Full Size | PDF

Fig. 10.
Fig. 10. In vivo endoscopic OCT imaging of Barrett’s esophagus using UHR- and standard resolution OCT. (A) UHR-OCT image with 4x enlargement (B); (C) Standard resolution OCT image with 4x enlargement (D). UHR-OCT image shows reduced speckle size and resolves finer morphological features (indicated by arrows).

Download Full Size | PDF

Fig. 11.
Fig. 11. Scatter plots of first two linear discrimination functions (LD1 and LD2) from PCA-LDA for normal and Barrett’s esophagus using 2D-DFT features from (A) standard resolution OCT images and (B) UHR-OCT images. (C) and (D) are the corresponding box plots for the distribution of LD1, which has the maximum separation between groups (“+” denotes outlier). The normal and Barrett’s groups are further separated using UHR-OCT.

Download Full Size | PDF

Fig. 12.
Fig. 12. Scatter plots of first two linear discrimination functions (LD1 and LD2) from PCA-LDA for normal and Barrett’s esophagus using CSAC features from (A) standard resolution OCT images and (B) UHR-OCT images. (C) and (D) are the corresponding box plots for the distribution of LD1, which has the maximum separation between groups (“+” denotes outlier). The normal and Barrett’s groups are further separated using UHR-OCT.

Download Full Size | PDF

Tables (1)

Tables Icon

Table 1. One-way ANOVA significance comparisons for UHR- and standard resolution (SDR) OCT classifications of normal and Barrett’s esophagus.

View Table

Equations (9)

Equations on this page are rendered with MathJax. Learn more.

Δ z = 2 ln 2 π • λ 0 2 Δ λ
F ( x , z ) = ∑ m = 0 N − 1 ∑ n = 0 N − 1 I ( m , n ) e − j ( 2 Π N ) xm e − j ( 2 Π N ) zn , x , z = 0 , 1 , … , N − 1 .
SCOV = 1 4 ∑ ( g i − u ) ( g i ′ − u )
VAR = 1 8 ∑ ( g i 2 + g i ′ 2 ) − u 2
BVAR = 1 16 ∑ ( g i + g i ′ ) 2 − u 2
WVAR = 1 16 ∑ ( g i − g i ′ ) 2
SVR = WVAR BVAR
SAC = SCOV VAR
u = 1 8 ∑ ( g i + g i ′ )
Select as filters
Select Topics Cancel
Top
       
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.
Privacy | Terms of Use