Background and aim The ability to differentiate between mucosal (M) or microinvasive submucosal (SM1: depth of less than 500 lm) and invasive submucosal (SM2: depth of 500 lm or more) cancer is paramount when choosing the method of treatment for early gastric cancer (EGC). The ''non-extension sign'' relates to a localized increase in thickness and rigidity due to massive submucosal invasion by a cancer. The present study sought to assess the ability of conventional endoscopy (CE) to correctly identify SM2 cancer using only the non-extension sign. Methods This is a retrospective study based on a prospectively collected database. EGCs had been diagnosed according to invasion depth as M-SM1 or SM2. In terms of the endoscopic diagnostic criterion, lesions positive for the non-extension sign were classified as SM2 cancers, while those negative for the non-extension sign were classified as M-SM1 cancers. Histopathological findings were used as the gold standard. Results We examined a total of 863 lesions from 704 patients, comprising 104 true-positive, 733 true-negative, 9 false-positive, and 17 false-negative lesions. This yielded a sensitivity of 92.0 % (95 % confidence interval (CI), 87.0-97.0 %), a specificity of 97.7 % (95 % CI, 96.7-98.8 %), a positive predictive value of 85.9 % (95 % CI, 79.7-92.1 %), a negative predictive value of 98.8 % (95 % CI, 98.0-99.6 %), and a diagnostic accuracy of 96.9 % (95 % CI, 95.8-98.1 %). Conclusion The non-extension sign may be useful for accurately determining the suitability of minimally invasive endoscopic treatment. Nevertheless, considering the limitations of retrospective analysis, a further prospective study is warranted to confirm the diagnostic reliability of the non-extension sign.