Objective        To establish a new endoscopic ultrasonography (EUS) scoring system based on the Breast imaging reporting and data system (BI-RADS)

and evaluate its diagnosis value to identify the benign and malignant lesions in gastrointestinal mucosa or submucosa or outside wall oppression.       Methods      A retrospective study was carried on 509 cases with pathologically-diagnosed gastrointestinal mucosa or submucosa or outside wall oppression admitted in our department from June 2006 to December 2012. We firstly established the EUS-BIRADS scoring system with reference to BI-RADS ultrasound diagnostic criteria

and then 2 experienced endoscopists scored each group of EUS images (randomly from the 509 patients) by reference to EUS-BIRADS scoring system with no idea on clinical data and pathological diagnosis of the patients. The obtained results were compared scoring with pathological results. Receiver operating curve (ROC) analysis showed the area under the curve (AUG) was 0.947. We chose EUS at score 6.5 as the critical value (corresponding to Youden index maximum)

and stipulated cases with the score ＞6 tended to be malignant lesions

and the others tended to be non-malignant ones. Pearson’s Chi-Square test

McNemar test

sensitivity

specificity

positive predictive value (PPV)

negative predictive value (NPV)

Youden index

diagnostic accordance rate

and Kappa value were carried out to assess the diagnostic value of EUS-BIRADS scoring system.       Results      For the benign and malignant lesions in gastrointestinal mucosa or submucosa or outside wall oppression

the sensitivity

specificity

PPV

NPV

Youden index

diagnostic accordance rate

and Kappa value were 85.1%

92.9%

71.2%

96.8%

0.779 5

91.6%

and 0.723 4

respectively in our EUS-BIRADS scoring system. Pearson’s Chi-square test (Chi-square=269.573

P0.001) showed the results of EUS-BIRADS scoring system were not independent to those from pathological examination

and   Pearson’s R (0.728) indicated that there was a positive correlation between the 2 kinds of results. But McNemar test (P=0.014) showed the EUS-BIRADS results did not substitute for pathological diagnosis.       Conclusion      EUS-BIRADS scoring system which is established based on BI-RADS standard is consistence with those from pathology for the lesions in digestive tract and outer wall oppression

and it can assist clinicians to predict the nature of lesions before further examination.