BACKGROUND AND OBJECTIVES: Temporal and spectral fluorescence spectroscopy can identify adenomatous colonic polyps accurately. In this study, these techniques were examined as a potential means of improving the surveillance of high grade dysplasia (HGD) in Barrett's esophagus (BE). STUDY DESIGN/ MATERIALS AND METHODS: Using excitation wavelengths of 337 and 400 nm, 148 fluorescence spectra, and 108 transient decay profiles (at 550 +/- 20 nm) were obtained endoscopically in 37 patients. Corresponding biopsies were collected and classified as carcinoma, HGD, or low risk tissue (LRT) [non-dysplastic BE, indefinite for dysplasia (IFD), and low grade dysplasia (LGD)]. Diagnostic algorithms were developed retrospectively using linear discriminant analysis (LDA) to separate LRT from HGD. RESULTS: LDA produced diagnostic algorithms based solely on spectral data. Moderate levels of sensitivity (Se) and specificity (Sp) were obtained for both 337 nm (Se = 74%, Sp = 67%) and 400 nm (Se = 74%, Sp = 85%) excitation. CONCLUSIONS: In the diagnosis of HGD in BE, steady-state fluorescence was more effective than time-resolved data, and excitation at 400 nm excitation was more effective than 337 nm. While fluorescence-targeted biopsy is approaching clinical usefulness, increased sensitivity to dysplastic changes-possibly through modification of system parameters-is needed to improve accuracy levels. Copyright 2003 Wiley-Liss, Inc.
BACKGROUND AND OBJECTIVES: Temporal and spectral fluorescence spectroscopy can identify adenomatous colonic polyps accurately. In this study, these techniques were examined as a potential means of improving the surveillance of high grade dysplasia (HGD) in Barrett's esophagus (BE). STUDY DESIGN/ MATERIALS AND METHODS: Using excitation wavelengths of 337 and 400 nm, 148 fluorescence spectra, and 108 transient decay profiles (at 550 +/- 20 nm) were obtained endoscopically in 37 patients. Corresponding biopsies were collected and classified as carcinoma, HGD, or low risk tissue (LRT) [non-dysplastic BE, indefinite for dysplasia (IFD), and low grade dysplasia (LGD)]. Diagnostic algorithms were developed retrospectively using linear discriminant analysis (LDA) to separate LRT from HGD. RESULTS: LDA produced diagnostic algorithms based solely on spectral data. Moderate levels of sensitivity (Se) and specificity (Sp) were obtained for both 337 nm (Se = 74%, Sp = 67%) and 400 nm (Se = 74%, Sp = 85%) excitation. CONCLUSIONS: In the diagnosis of HGD in BE, steady-state fluorescence was more effective than time-resolved data, and excitation at 400 nm excitation was more effective than 337 nm. While fluorescence-targeted biopsy is approaching clinical usefulness, increased sensitivity to dysplastic changes-possibly through modification of system parameters-is needed to improve accuracy levels. Copyright 2003 Wiley-Liss, Inc.
Authors: Diego R Yankelevich; Dinglong Ma; Jing Liu; Yang Sun; Yinghua Sun; Julien Bec; Daniel S Elson; Laura Marcu Journal: Rev Sci Instrum Date: 2014-03 Impact factor: 1.523
Authors: William R Lloyd; Robert H Wilson; Seung Yup Lee; Malavika Chandra; Barbara McKenna; Diane Simeone; James Scheiman; Mary-Ann Mycek Journal: Biomed Opt Express Date: 2013-12-02 Impact factor: 3.732