Endoscopic Diagnosis of Chronic Atrophic Gastritis and Early Gastric Cancer: From Basics to Advanced Imaging

Chronic atrophic gastritis (CAG) is the principal precursor lesion for gastric adenocarcinoma and represents a key target for endoscopic surveillance and early intervention. Although the global age-standardised incidence of gastric cancer has declined over recent decades, the absolute number of cases continues to rise because of population ageing and increasing incidence in younger individuals. The prognosis remains poor in advanced disease, whereas early gastric cancer (EGC) detected at a mucosal stage is associated with excellent long-term survival and may be curable with endoscopic resection. Consequently, high-quality endoscopic detection of premalignant gastric lesions is essential to reduce gastric cancer mortality. This review summarises current concepts in the endoscopic diagnosis of CAG, gastric intestinal metaplasia (GIM), and EGC, from conventional white-light endoscopy through to advanced imaging and artificial intelligence (AI)-assisted systems. Fundamental principles of high-quality oesophagogastroduodenoscopy are discussed, including adequate inspection time, systematic mucosal assessment, mucosal cleansing, and standardised photo-documentation. Characteristic endoscopic appearances of normal gastric mucosa, atrophy, and intestinal metaplasia are reviewed, alongside established staging systems including the Kimura–Takemoto and EGGIM classifications. The role of image-enhanced endoscopy is examined in detail, including narrow-band imaging, linked colour imaging, texture and colour enhancement imaging, and magnification optical enhancement. These modalities improve visualisation of pit patterns, microvascular architecture, and hallmark features of intestinal metaplasia such as the light blue crest sign, substantially increasing diagnostic sensitivity and specificity compared with conventional white light imaging alone. Advanced imaging combined with magnification also enhances the detection and characterisation of EGC. Emerging evidence regarding AI-assisted endoscopy demonstrates promising diagnostic accuracy for CAG, GIM, and early neoplasia, with improved lesion detection and reduced miss rates in several studies. However, limitations relating to external validation, generalisability, and integration into routine practice remain. Continued advances in optical imaging, structured training, and AI-supported diagnostics are likely to play an increasingly important role in improving early gastric cancer detection and surveillance outcomes worldwide.