BACKGROUND AND AIM: White globe appearance (WGA) is a small white lesion with a globular shape that can be identified by magnifying endoscopy with narrow-band imaging (M-NBI). WGA was recently reported as a novel endoscopic marker that can differentiate between gastric cancer (GC) and low-grade adenoma. However, the usefulness of WGA for differentiating GC from non-cancerous lesions (NC), including those of gastritis, is unknown. METHODS: To compare the prevalence of WGA in GC and NC, we carried out a prospective study of 994 patients undergoing gastroscopy. All patients were examined for target lesions that were suspected to be GC. When a target lesion was detected, the presence or absence of WGA in the lesion was evaluated using M-NBI, and all target lesions were biopsied or resected for histopathological diagnosis. Primary endpoint was a comparison of WGA prevalence in GC and NC. Secondary endpoints included WGA diagnostic performance for diagnosing GC. RESULTS: A total of 188 target lesions from 156 patients were analyzed for WGA, and histopathological diagnoses included 70 cases of GC and 118 cases of NC. WGA prevalence in GC and NC was 21.4% (15/70) and 2.5% (3/118), respectively (P < 0.001). WGA diagnostic accuracy, sensitivity, and specificity for detecting GC were 69.1%, 21.4%, and 97.5%, respectively. CONCLUSIONS: WGA prevalence in GC is significantly higher than that in NC. Because WGA is highly specific for GC, the presence of WGA is useful to diagnose GC.
BACKGROUND AND AIM: White globe appearance (WGA) is a small white lesion with a globular shape that can be identified by magnifying endoscopy with narrow-band imaging (M-NBI). WGA was recently reported as a novel endoscopic marker that can differentiate between gastric cancer (GC) and low-grade adenoma. However, the usefulness of WGA for differentiating GC from non-cancerous lesions (NC), including those of gastritis, is unknown. METHODS: To compare the prevalence of WGA in GC and NC, we carried out a prospective study of 994 patients undergoing gastroscopy. All patients were examined for target lesions that were suspected to be GC. When a target lesion was detected, the presence or absence of WGA in the lesion was evaluated using M-NBI, and all target lesions were biopsied or resected for histopathological diagnosis. Primary endpoint was a comparison of WGA prevalence in GC and NC. Secondary endpoints included WGA diagnostic performance for diagnosing GC. RESULTS: A total of 188 target lesions from 156 patients were analyzed for WGA, and histopathological diagnoses included 70 cases of GC and 118 cases of NC. WGA prevalence in GC and NC was 21.4% (15/70) and 2.5% (3/118), respectively (P < 0.001). WGA diagnostic accuracy, sensitivity, and specificity for detecting GC were 69.1%, 21.4%, and 97.5%, respectively. CONCLUSIONS: WGA prevalence in GC is significantly higher than that in NC. Because WGA is highly specific for GC, the presence of WGA is useful to diagnose GC.