| IntroductionEndoscopy is the accepted gold standard for the detection of premalignant conditions of the stomach. The current study aims to compare high definition endoscopy with i-Scan imaging vs. high definition endoscopy with white-light imaging for the diagnosis of premalignant gastric lesions such as chronic superficial gastritis, chronic atrophic gastritis, intestinal metaplasia and gastric intraepithelial neoplasia.MethodsA total of 62 patients undergoing upper i-Scan endoscopy were enrolled in the study. Modulated the EPK-i system from the common white-light imaging to the twin-mode during endoscopic observation and captured images if only lesions visible or suspected in any endoscopic imaging mode, taking targeted biopsies as possible as could in these imaged areas, and then the samples were performed histopathologic examinations. We selected the images obtained with endoscopy basing on overall imaging quality and discarded the ones pathological diagnosed as cancerous. We used the paint program of computer to separate the twin-mode digital image into white-light parts and i-Scan parts individually.2 groups of endoscopists (group A: experienced; group B:inexperienced) were enrolled to evaluate the images independently. The evaluation contains 3 aspects:the character, boundary and number of the lesions. SPSS 13.0 was applied for statistical analysis. Pathological diagnosis were used as gold standard. Chi-square test or Fisher test were used to calculate the accuracy, sensitivity and specificity of the character evaluation of the lesions. Parameters of the two groups were compared by chi-square test. The data of the boundary and number evaluation of the lesions was calculated as mean±standard deviation. Parameters of the two groups were compared by Mann-Whitney U test.ResultsEventually, a total of 242 images (white-light endoscopy images:121, i-Scan images:121) were used for the boundary and number evaluation of the lesions.186 of which who have histopathologic diagnosis (white-light endoscopy images:93, i-Scan images:93) were used for the character evaluation of the lesions. The evaluation outcomes were that:for the evaluation of the lesion character, i-Scan is superior to white-light endoscopy for inexperienced endoscopists in the accuracy of judging the intestinal metaplasia and gastric intraepithelial neoplasia (intestinal metaplasia:22.6% vs.11.3%, P<0.05; gastric intraepithelial neoplasia:33.3%vs.16.7%, P<0.05), except that, i-Scan is equal to white-light endoscopy in evaluating the character of the lesions; for the evaluation of the lesion boundary, i-Scan is superior to white-light endoscopy (group A:1.19±0.82 vs.0.71±0.74, P<0.05; group B:0.86±0.88 vs.0.58±0.67, P<0.05); for the evaluation of lesion numbers, i-Scan is equal to white-light endoscopy (group A:2.15±2.36 vs.2.02±2.45, P>0.05; group B:1.19±1.42 vs. 1.17±1.46,P>0.05).ConclusionsI-Scan is superior to white-light endoscopy for inexperienced endoscopists in the accuracy of judging the intestinal metaplasia and gastric intraepithelial neoplasia. I-Scan is superior to white-light endoscopy in detecting the boundary of the lesions, but equal to white-light endoscopy in detecting the number of lesions.SignificanceThis study verified the value of i-Scan diagnosing premalignant gastric lesions, which is significant for the clinical application of i-Scan. |