Surface pattern classification by enhanced-magnification endoscopy for identifying early gastric cancers

Background The correlation between fine surface patterns of gastric mucosal lesions and early gastric cancer is not sufficiently clear. Objective To evaluate the efficacy of surface pattern classification by enhanced-magnification endoscopy (EME) for identifying early gastric cancers. Design Observational study. Setting All procedures were performed at Mie University Hospital. Patients A total of 380 consecutive patients who underwent EGD by using magnification endoscopy. Among these subjects were found 250 newly detected lesions suspected of being gastric cancer. Methods Conventional magnification endoscopy (CME), magnification chromoendoscopy (MCE), and EME were performed, and surface patterns of lesions were classified into 5 types: type I, small round pits of uniform size and shape; type II, slit-like pits; type III, gyrus and villous patterns; type IV, irregular arrangement and size; and type V, destructive pattern. Biopsy specimens were obtained from all lesions. Main Outcome Measurements Correlation between surface pattern classification by EME and histopathologic findings of early gastric cancer. Results Surface patterns were evident by CME/MCE in only 66.4% (166/250) of lesions but in 100% (250/250) of lesions by EME. Classification by EME was as follows: type I, 52 lesions; type II, 12; type III, 146; type IV, 32; and type V, 8. By histopathologic examination, 16 early gastric cancers were detected between type IV or V lesions. Thus, classification of types IV-V strongly correlated with the presence of gastric cancer (sensitivity 100%, specificity 89.7%). Limitations Single-center study. Conclusions Surface pattern classification by EME may be useful for identifying early gastric cancers.