Objective:To develop a deep learning-based system for real-time recognition and classification of portal hypertensive gastropathy (PHG) and evaluate its ability to assist junior endoscopists.Methods:A total of 2 848 gastroscopy images from 832 patients with liver cirrhosis were selected from Digestive Endoscopy Center databases of Renmin Hospital of Wuhan University, Wuhan Hospital of Traditional Chinese and Western Medicine, and the Second Hospital of Jingzhou from January 2015 to October 2023. This system referred to 3 endoscopic features of Baveno Ⅱ scoring system. Three models were developed respectively for gastric antral vascular ectasia (GAVE), mosaic-like pattern (MLP), and red marks (RM). The specific classification references were as follows: (1) GAVE model: 0 no, 1 yes; (2) MLP model: 0 no, 1 mild, 2 severe; (3) RM model: 0 no, 1 isolated, 2 fused. The classification results for endoscopic characteristics of PHG of 3 endoscopy experts were taken as the gold standard. The yolov8-m model was used for training. The training dataset, validation dataset, and test dataset were allocated at a ratio of 8∶1∶1. The test dataset was used to evaluate the performance of models and their auxiliary effects on endoscopists. The accuracy, recall, precision, specificity and Kappa coefficient were calculated. Results:The accuracy, recall, specificity of GAVE model were 96.0% (48/50), 87.5% (7/8) and 97.6% (41/42). There was no significant difference between its accuracy and the gold standard ( χ2=316.226, P=1.000). The precision of GAVE1 and GAVE0 were 87.5% (7/8) and 97.6% (41/42) respectively. The accuracy of MLP model was 84.1% (132/157), and there was no significant difference compared with the gold standard ( χ2=3.286, P=0.193). The precision and recall of MLP2 were 88.2% (15/17) and 75.0% (15/20). The precision and recall of MLP1 were 77.9% (60/77) and 88.2% (60/68). The precision and recall of MLP0 were 90.5% (57/63) and 82.6% (57/69). The accuracy of RM model was 87.9% (123/140), and there was no significant difference compared with the gold standard ( χ2=2.891, P=0.409). The precision and recall of RM2 were 94.7% (18/19) and 78.3% (18/23). The precision and recall of RM1 were 72.2% (26/36) and 81.3% (26/32). The precision and recall of RM0 were 92.9% (79/85) and 92.9% (79/85). The mean accuracy of the three junior endoscopists, with and without the assistance of the GAVE model, MLP model, and RM model, respectively increased from 95.3% to 99.3%, from 83.9% to 91.9%, and from 81.9% to 83.1%. The overall consistency analysis of the 3 junior endoscopists with the gold standard indicated that the consistency of the GAVE model before and after assistance was extremely strong (both an overall Kappa of 1.000); the consistency before assistance of the MLP model was moderate (with an overall Kappa of 0.601), which increased to extremely strong after assistance (with an overall Kappa of 0.964); and the consistency of the RM model before and after assistance was also relatively strong (with an overall Kappa of 0.792 before and 0.798 after). Conclusion:The deep learning system accurately identifies and classifies PHG features and significantly enhances diagnostic performance of junior endoscopists.

Objective:To develop a deep learning-based system for real-time recognition and classification of portal hypertensive gastropathy (PHG) and evaluate its ability to assist junior endoscopists.Methods:A total of 2 848 gastroscopy images from 832 patients with liver cirrhosis were selected from Digestive Endoscopy Center databases of Renmin Hospital of Wuhan University, Wuhan Hospital of Traditional Chinese and Western Medicine, and the Second Hospital of Jingzhou from January 2015 to October 2023. This system referred to 3 endoscopic features of Baveno Ⅱ scoring system. Three models were developed respectively for gastric antral vascular ectasia (GAVE), mosaic-like pattern (MLP), and red marks (RM). The specific classification references were as follows: (1) GAVE model: 0 no, 1 yes; (2) MLP model: 0 no, 1 mild, 2 severe; (3) RM model: 0 no, 1 isolated, 2 fused. The classification results for endoscopic characteristics of PHG of 3 endoscopy experts were taken as the gold standard. The yolov8-m model was used for training. The training dataset, validation dataset, and test dataset were allocated at a ratio of 8∶1∶1. The test dataset was used to evaluate the performance of models and their auxiliary effects on endoscopists. The accuracy, recall, precision, specificity and Kappa coefficient were calculated. Results:The accuracy, recall, specificity of GAVE model were 96.0% (48/50), 87.5% (7/8) and 97.6% (41/42). There was no significant difference between its accuracy and the gold standard ( χ2=316.226, P=1.000). The precision of GAVE1 and GAVE0 were 87.5% (7/8) and 97.6% (41/42) respectively. The accuracy of MLP model was 84.1% (132/157), and there was no significant difference compared with the gold standard ( χ2=3.286, P=0.193). The precision and recall of MLP2 were 88.2% (15/17) and 75.0% (15/20). The precision and recall of MLP1 were 77.9% (60/77) and 88.2% (60/68). The precision and recall of MLP0 were 90.5% (57/63) and 82.6% (57/69). The accuracy of RM model was 87.9% (123/140), and there was no significant difference compared with the gold standard ( χ2=2.891, P=0.409). The precision and recall of RM2 were 94.7% (18/19) and 78.3% (18/23). The precision and recall of RM1 were 72.2% (26/36) and 81.3% (26/32). The precision and recall of RM0 were 92.9% (79/85) and 92.9% (79/85). The mean accuracy of the three junior endoscopists, with and without the assistance of the GAVE model, MLP model, and RM model, respectively increased from 95.3% to 99.3%, from 83.9% to 91.9%, and from 81.9% to 83.1%. The overall consistency analysis of the 3 junior endoscopists with the gold standard indicated that the consistency of the GAVE model before and after assistance was extremely strong (both an overall Kappa of 1.000); the consistency before assistance of the MLP model was moderate (with an overall Kappa of 0.601), which increased to extremely strong after assistance (with an overall Kappa of 0.964); and the consistency of the RM model before and after assistance was also relatively strong (with an overall Kappa of 0.792 before and 0.798 after). Conclusion:The deep learning system accurately identifies and classifies PHG features and significantly enhances diagnostic performance of junior endoscopists.

Objective: To discuss the predictive value of liver and spleen stiffness detected by transientelasticity imaging technology on esophageal varices bleeding (EVB) of patients with hepatitis B cirrhosis. Methods: 100 patients with hepatitis B cirrhosis (18 cases were no EV, 30 cases were mild EV, 28 cases were moderate EV and 24 cases were severe EV) were selected. In these patients, 52 cases were no bleeding, 28 cases were single hemorrhage and 20 cases were multiple hemorrhage. As the Child-Pugh grading, there were 34 cases were A grade, 30 cases were B grade and 36 cases were C grade. All of the liver and spleen stiffness of these patients were detected by instantaneous elasticity imaging technology, and to compare different Child-Pugh grading of patients, different EV degree, and liver and spleen stiffness on different EVB situation. Results: The liver and spleen stiffness of patients in Child-Pugh C stage were significantly higher than that in Child-Pugh A stage and B stage, respectively (t=21.13, t=12.04, t=11.24, t=9.741; P<0.05). With the aggravating of the severe degree of EV, the liver and spleen stiffness were significant increasing, and there were significant differences between different EV degree patients for liver stiffness and spleen stiffness, respectively (F=7.494, F=8.129, P<0.05). The liver and spleen stiffness of multiple hemorrhage group were significant higher than that of no-hemorrhage group and single hemorrhage group, respectively (t=13.13, t=18.14, t=12.15, t=17.46; P<0.05). Conclusion: By using instantaneous elasticity imaging technology to detect liver and spleen stiffness can predict EVB of patient with hepatitis B cirrhosis and it has higher clinical value, therefore, it is a sophisticated noninvasive examination with simple operation and better repeatability.

OBJECTIVETo perform an analysis and comparative study of the clinical data for patients with cirrhosis and type 1 hepatorenal syndrome (HRS) who received treatment with terlipressin using high-or low-dose regimens.

METHODSA total of 56 patients with cirrhosis and type 1 HRS who presented for treatment to the Wuhan Medical Treatment Center and Taizhou Central Hospital between March 2010 and October 2012 were enrolled in the study. The patients were randomly assigned to the terlipressin treatment groups for receipt of the high-dose regimen (1 mg/6-8 h;n =27) or low-dose regimen (1 mg/12 h;n =29). All patients were assessed for 24-hour urine volume, serum blood urea nitrogen (BUN) and creatinine (Cr) levels, therapeutic effect and prognosis, and adverse reactions. Measurements were made before and after the treatment, and on post-treatment days 3, 7 and 14. Inter-group differences were assessed by statistical analyses.

RESULTSThe high-dose group showed an increase in 24-hour urine volumes from post-treatment day 3 (1112 ± 262 ml) to day 7 (1938 ± 312 ml), and the volumes on both days were significantly better than those of the low-dose group (day 3:986 ± 162 ml and day 7:1760 ± 300 ml, t =1.500, 1.830, P=0.038, 0.041). The high-dose group also showed a significantly better decreases in serum BUN levels (35.1 ± 8.6 to 30.2 ± 6.3 mmol/L vs.low-dose group: 43.2 ± 10.9 to 35.1 ± 7.6 mmol/L, t =3.200, 5.901, P =0.043, 0.047) and in serum Cr values (219.0 ± 35.1 to 128.2 ± 41.6 vs.low-dose group: 230.3 ± 82.1 to 151.5 ± 38.7, t =2.997, 5.765, P =0.036, 0.046).On post-treatment day 14 the 24-hour urine volume of patients in the high-dose group decreased (to 720+/-136 ml), but the difference from that of the low-dose group was not significant (vs. 620 ± 164 ml, t =1.855, P =0.069). The serum BUN level increased in the high-dose group (to 54.4 ± 15.0 mmol/L), which was statistically different from that in the low-dose group (vs .57.7 ± 17.3 mmol/L, t=5.166, P =0.022); the same trend was seen for the serum Cr value (397.8 ± 127.4 mumol/L vs. 480.3 ± 179.8 mumol/L, t =5.638, P =0.047). No statistically significant differences were observed for the groups in regard to significant efficiency, efficiency or 2-week survival rate (x2 =2.314, 1.767, 0.678, P =0.128, 0.128, 0.410 respectively), but the total efficiency was significantly different between the two groups (x² =5.793, P =0.016). In addition, no serious adverse reactions (including precordial pain, myocardial infarction or intestinal necrosis) were observed in either group.

CONCLUSIONTerlipressin therapy at both high and low dosages can lead to significant beneficial effects within as little as 3 days after the treatment; however, the high-dose appears to produce a better lasting efficacy (at day 14 after the treatment). The difference in doses does not appear to markedly affect significant efficiency, efficiency, nor the 2-week survival rate.

Adolescent ; Adult ; Aged ; Dose-Response Relationship, Drug ; Female ; Hepatorenal Syndrome ; drug therapy ; etiology ; Humans ; Liver Cirrhosis ; complications ; drug therapy ; Lypressin ; administration & dosage ; analogs & derivatives ; therapeutic use ; Male ; Middle Aged ; Retrospective Studies ; Treatment Outcome ; Young Adult