Article ID Journal Published Year Pages File Type
3307894 Gastrointestinal Endoscopy 2009 10 Pages PDF
Abstract

BackgroundThe endoscopic diagnosis of an elevated-type early gastric cancer is often difficult. The optimal band imaging (OBI) system can reconstruct the best spectral images decomposed from ordinary endoscopic images with free selection of 3 wavelengths and provide unmagnified images with high light intensity as well as magnified images.ObjectiveTo evaluate whether the OBI system facilitates detection of the demarcation lines between an elevated-type early gastric cancer and surrounding tissue and thus is more helpful for performing endoscopic therapy.DesignA prospective study.SettingJichi Medical University, Japan.PatientsSeventy-five patients, 81 lesions with an elevated-type early gastric cancer.Main Outcome MeasurementsA comparison between OBI images and conventional endoscopic images in the identification of the demarcation lines of an elevated-type early gastric cancer without magnification and the rate of success in identifying the abnormal surface structure of cancer by using low-magnified OBI images.ResultsDemarcation lines were easily identified in OBI images, even without magnification, because such cancers could be clearly distinguished from the surrounding whitish atrophic mucosa. Inexperienced endoscopists could determine demarcation lines with significantly greater accuracy with unmagnified OBI images than with conventional images. With 40-fold magnification, irregular microstructural or nonstructural patterns were also found within cancer lesions in all cases studied but in none of the cases in the surrounding noncancerous mucosa.ConclusionsThe new contrast images obtained with the OBI system enable better determination of the demarcation lines of elevated-type early gastric cancers, and this system may be useful for performing endoscopic therapy of this type of cancer as a new endoscopic modality.

Related Topics
Health Sciences Medicine and Dentistry Gastroenterology
Authors
, , , , , , , , ,