Background and aims: The pathogenesis of gastric cancer involves premalignant changes of the gastric mucosa. An accurate estimation of the topography and severity of these lesions represents an important step in detecting premalignant lesions, thereby classifying patients into low or high risk of developing gastric cancer. We prospectively analyzed the diagnostic performance of narrow-band imaging with magnification endoscopy (NBI-ME) for assessing premalignant gastric lesions during real-time examination.
Patients, materials and methods: A total number of 59 patients were examined by NBI-ME and target biopsies of the antrum, corporeal, and incisura angularis levels. Modified endoscopic patterns were classified into three groups: type A [tubulo-villous mucosal pattern with regular microvessels, or the light blue crest (LBC) sign], type B [disappearance of normal subepithelial capillary network (SECN) pattern], and type C [irregular mucosal pattern (IMP) and∕or irregular vascular pattern (IVP)]. The endoscopic diagnosis was compared to histological findings (the gold standard). The NBI-ME results were assessed for accuracy, sensitivity, specificity, and negative and positive predictive values in detecting intestinal metaplasia, atrophic gastritis and dysplasia.
Results: Analysis of endoscopic patterns showed a good correlation with premalignant lesions (p<0.05). Type A pattern showed 80.2% accuracy, 80.43% sensitivity and 80% specificity [area under receiver operating characteristic (AUROC) of 0.8] in detecting intestinal metaplasia. Diagnostic performance for assessment of atrophic gastritis was not ideal (69.5% accuracy, 83.72% sensitivity, 56.04% specificity, AUROC 0.69). Pattern C represents a reliable endoscopic marker for the diagnosis of dysplasia (91.1% accuracy, 83.3% sensitivity, 91.81% specificity, AUROC 0.87). The extension of precancerous lesions was estimated during endoscopic examination.
Conclusions: NBI-ME represents a valuable tool in the assessment of premalignant gastric lesions, thereby categorizing patients into low and high risks of developing gastric cancer. The applicability of the method in routine practice is promising, as it helps shape the follow up protocol of patients with premalignant lesions of the stomach. It is worth mentioning that, this method requires standardization, additional training, and expertise.