Objective To investigate the value of digital image analysis (DIA) of endoscopic ultrasonography (EUS) in the differential diagnosis of benign and malignant pancreatic neuroendocrine neoplasm (PNEN).Methods The relation between various parameters of EUS-imaging and benign and malignant lesions in 47 patients clinically diagnosed PNEN were retrospectively analyzed.Photoshop CS5 software was performed for digital image processing,and lesions related parameters were collected,including area,perimeter,length,circularity,gray,gray ratio,and gray standard deviation.The statistical method of t test was performed for comparison between two groups.Receiver operating characteristic (ROC) curve was analyzed in length,circularity,average gray scale ratio and gray standard deviation.Results Among the 47 patients,35 cases and 12 cases were in benign group and malignant group,respectively.The mean gray scale ratio and the circularity of benign group were significantly higher than those of malignant group (0.80±0,05 vs 0.74±0.07,0.63±0.17 vs 0.40±0.09),and the differences were statistically significant (t=2.659 and 5.787,both P＜0.05).The gray standard deviation of benign group were lower than that of malignant group (9.90 ± 1.24 vs 12.55± 3.27),and the difference was statistically significant (t=-2.733,P=0.018).The area under the curve (AUC) of circularity was 0.724 (95％ confidence interval(CI):0.546 to 0.901),the cut-off value was 0.767,the sensitivity and specificity were 71.43 ％ and 66.67％,respectively.The AUC of average gray ratio was 0.888 (95％CI:0.785 to 0.991),the cut-off value was 0.412,the sensitivity and specificity were 94.29％ and 75.00％,respectively.The AUC of gray standard deviation was 0.811 (95％CI:0.647 to 0.974),the cut-off value was 11.02,the sensitivity and specificity were 66.67％ and 85.71％,respectively.When combined with the three parameters of circularity,average gray scale ratio and gray standard deviation,the sensitivity and specificity were 97.14％ and 83.33％,and the accuracy was 93.61％.Conclusions EUS with DIA technology can improve the detection of EUS images to PNEN,and which may be complementary to EUS guided fine needle aspiration.It also privided a noninvasive,objective,convenient,and effective option for the differential diagnosis of benign and malignant PNEN.

Objective:To analyze the incidence and risk factors of non-acute symptomatic portal vein thrombosis (PVT) after endoscopic gastric variceal injection (GVI) in the treatment of liver cirrhosis with gastric variceal bleeding (GVB).Methods:66 patients with GVB who were treated with GVI for the first time from July 2017 to October 2019 in Minhang Hospital Affiliated to Fudan University were retrospectively analyzed. The data of gender, age, preoperative Child-Pugh grade, preoperative platelet count, preoperative plasma D-dimer concentration, preoperative splenic length, preoperative portal vein velocity, preoperative splenic vein velocity, preoperative portal vein diameter, preoperative splenic vein diameter, treatment times, total number of injection points, total dose of sclerosing agent and tissue adhesive agent were collected. The patients were divided into PVT group and non-PVT group according to the occurrence of non-acute symptomatic PVT within one year after GVI. Univariate analysis was performed first, and then the factors with P<0.2 were included in the binary logistic regression model to screen the risk factors of PVT after GVI. Results:There were 25 cases (37.88%) in PVT group and 41 cases (62.12%) in non-PVT group. There were significant differences in D-dimer concentration, spleen length, Child-Pugh grade and total dose of sclerosing agent between the two groups ( P<0.05). The D-dimer concentration ( OR=2.319, 95% CI:1.359-3.956), spleen length ( OR=1.044, 95% CI:1.007-1.081) and total dose of sclerosing agent ( OR=1.075, 95% CI:1.004-1.152) were independent risk factors for PVT ( P<0.05). Conclusions:Preoperative D-dimer concentration, spleen length and total dose of sclerosing agent can predict the risk of PVT after GVI. In order to reduce the risk of PVT after GVI, the dose of sclerosing agent should be reduced as much as possible.