Abstract
Patients with Barrett's esophagus (BE) undergo periodic endoscopie surveillance with random biopsies in an effort to detect dysplastic or early cancerous lesions. Surveillance may be enhanced by near-infrared Raman spectroscopy (NIRS), which has the potential to identify endoscopically-occult dysplastic lesions within the Barrett's segment and allow for targeted biopsies. The aim of this study was to assess the diagnostic performance of NIRS for identifying dysplastic lesions in BE in vivo. Raman spectra (P exc=70 mW; t=5 s) were collected from Barrett's mucosa at endoscopy using a custom-built NIRS system (λ exc=785 nm) equipped with a filtered fiber-optic probe. Each probed site was biopsied for matching histological diagnosis as assessed by an expert pathologist. Diagnostic algorithms were developed using genetic algorithm-based feature selection and linear discriminant analysis, and classification was performed on all spectra with a bootstrap-based cross-validation scheme. The analysis comprised 192 samples (112 non-dysplastic, 54 low-grade dysplasia and 26 high-grade dysplasia/early adenocarcinoma) from 65 patients. Compared with histology, NIRS differentiated dysplastic from non-dysplastic Barrett's samples with 86% sensitivity, 88% specificity and 87% accuracy. NIRS identified 'high-risk' lesions (high-grade dysplasia/early adenocarcinoma) with 88% sensitivity, 89% specificity and 89% accuracy. In the present study, NIRS classified Barrett's epithelia with high and clinically-useful diagnostic accuracy.
Original language | English (US) |
---|---|
Title of host publication | Progress in Biomedical Optics and Imaging - Proceedings of SPIE |
Editors | T. Vo-Dinh, W.S. Grundfest, D.A. Benaron, G.E. Cohn |
Pages | 140-146 |
Number of pages | 7 |
Volume | 5692 |
DOIs | |
State | Published - 2005 |
Event | Advanced Biomedical and Clinical Diagnostic Systems III - San Jose, CA, United States Duration: Jan 23 2005 → Jan 26 2005 |
Other
Other | Advanced Biomedical and Clinical Diagnostic Systems III |
---|---|
Country | United States |
City | San Jose, CA |
Period | 1/23/05 → 1/26/05 |
Fingerprint
Keywords
- Barrett's esophagus
- Dysplasia
- Endoscopy
- Raman spectroscopy
ASJC Scopus subject areas
- Engineering(all)
Cite this
Diagnostic potential of Raman spectroscopy in Barrett's esophagus. / Wong Kee Song, Louis Michel; Molckovsky, Andrea; Wang, Kenneth Ke Ning; Burgart, Lawrence J.; Dolenko, Brion; Somorjai, Rajmund L.; Wilson, Brian C.
Progress in Biomedical Optics and Imaging - Proceedings of SPIE. ed. / T. Vo-Dinh; W.S. Grundfest; D.A. Benaron; G.E. Cohn. Vol. 5692 2005. p. 140-146 23.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Diagnostic potential of Raman spectroscopy in Barrett's esophagus
AU - Wong Kee Song, Louis Michel
AU - Molckovsky, Andrea
AU - Wang, Kenneth Ke Ning
AU - Burgart, Lawrence J.
AU - Dolenko, Brion
AU - Somorjai, Rajmund L.
AU - Wilson, Brian C.
PY - 2005
Y1 - 2005
N2 - Patients with Barrett's esophagus (BE) undergo periodic endoscopie surveillance with random biopsies in an effort to detect dysplastic or early cancerous lesions. Surveillance may be enhanced by near-infrared Raman spectroscopy (NIRS), which has the potential to identify endoscopically-occult dysplastic lesions within the Barrett's segment and allow for targeted biopsies. The aim of this study was to assess the diagnostic performance of NIRS for identifying dysplastic lesions in BE in vivo. Raman spectra (P exc=70 mW; t=5 s) were collected from Barrett's mucosa at endoscopy using a custom-built NIRS system (λ exc=785 nm) equipped with a filtered fiber-optic probe. Each probed site was biopsied for matching histological diagnosis as assessed by an expert pathologist. Diagnostic algorithms were developed using genetic algorithm-based feature selection and linear discriminant analysis, and classification was performed on all spectra with a bootstrap-based cross-validation scheme. The analysis comprised 192 samples (112 non-dysplastic, 54 low-grade dysplasia and 26 high-grade dysplasia/early adenocarcinoma) from 65 patients. Compared with histology, NIRS differentiated dysplastic from non-dysplastic Barrett's samples with 86% sensitivity, 88% specificity and 87% accuracy. NIRS identified 'high-risk' lesions (high-grade dysplasia/early adenocarcinoma) with 88% sensitivity, 89% specificity and 89% accuracy. In the present study, NIRS classified Barrett's epithelia with high and clinically-useful diagnostic accuracy.
AB - Patients with Barrett's esophagus (BE) undergo periodic endoscopie surveillance with random biopsies in an effort to detect dysplastic or early cancerous lesions. Surveillance may be enhanced by near-infrared Raman spectroscopy (NIRS), which has the potential to identify endoscopically-occult dysplastic lesions within the Barrett's segment and allow for targeted biopsies. The aim of this study was to assess the diagnostic performance of NIRS for identifying dysplastic lesions in BE in vivo. Raman spectra (P exc=70 mW; t=5 s) were collected from Barrett's mucosa at endoscopy using a custom-built NIRS system (λ exc=785 nm) equipped with a filtered fiber-optic probe. Each probed site was biopsied for matching histological diagnosis as assessed by an expert pathologist. Diagnostic algorithms were developed using genetic algorithm-based feature selection and linear discriminant analysis, and classification was performed on all spectra with a bootstrap-based cross-validation scheme. The analysis comprised 192 samples (112 non-dysplastic, 54 low-grade dysplasia and 26 high-grade dysplasia/early adenocarcinoma) from 65 patients. Compared with histology, NIRS differentiated dysplastic from non-dysplastic Barrett's samples with 86% sensitivity, 88% specificity and 87% accuracy. NIRS identified 'high-risk' lesions (high-grade dysplasia/early adenocarcinoma) with 88% sensitivity, 89% specificity and 89% accuracy. In the present study, NIRS classified Barrett's epithelia with high and clinically-useful diagnostic accuracy.
KW - Barrett's esophagus
KW - Dysplasia
KW - Endoscopy
KW - Raman spectroscopy
UR - http://www.scopus.com/inward/record.url?scp=21844468957&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=21844468957&partnerID=8YFLogxK
U2 - 10.1117/12.584986
DO - 10.1117/12.584986
M3 - Conference contribution
AN - SCOPUS:21844468957
VL - 5692
SP - 140
EP - 146
BT - Progress in Biomedical Optics and Imaging - Proceedings of SPIE
A2 - Vo-Dinh, T.
A2 - Grundfest, W.S.
A2 - Benaron, D.A.
A2 - Cohn, G.E.
ER -