Objective:To explore the value of radiomics and deep learning in predicting the efficacy of initial transarterial chemoembolization (TACE) for hepatocellular carcinoma (HCC).Methods:This was a cohort study. The imaging and clinical information of HCC patients treated with TACE in the Second Affiliated Hospital of Harbin Medical University from January 2015 to January 2021 were collected retrospectively. A total of 265 patients were divided into response group (175 cases) and non-response group (90 cases) according to the modified solid tumor efficacy evaluation criteria (mRECIST) 1 to 2 months after initial TACE. According to the proportion of 8∶2, the patients were randomly divided into training group (212 cases, 140 responders and 72 non-responders) and test set (53 cases, 35 responders and 18 non-responders). Univariate and multivariate logistic regression was used to screen clinical variables and construct a clinical model. The radiomics features were extracted from the preoperative CT images, and radiomics model was constructed after feature dimensionality reduction. Using the deep learning method, three residual network (ResNet) models (ResNet18, ResNet50 and ResNet101) were established, and their effectiveness was compared and integrated to build a deep learning model with best performance. Univariate and multivariate logistic regression was used to combine pairwise three models to establish the combined model. The receiver operating characteristic curve was used to evaluate the performance of the model to distinguish between TACE response and non-response groups.Results:In the test set, the area under the curve (AUC) of the clinical model and the radiomics model in the differentiation between response and non-response after TACE were 0.730 (95% CI 0.569-0.891) and 0.775 (95% CI 0.642-0.907). The AUC of ResNet18, ResNet50 and ResNet101 were 0.719, 0.748 and 0.533, respectively. The AUC for deep learning model obtained by integrating ResNet18 and ResNet50 was 0.806 (95% CI 0.665-0.946). After pairwise fusion, the combined deep learning-radiomics model showed the highest performance, with an AUC of 0.843 (95% CI 0.730-0.956), which was better than those of the deep learning-clinical model (AUC of 0.838, 95% CI 0.719-0.957) and the radiomics-clinical model (AUC of 0.786, 95% CI 0.648-0.898). Conclusions:The combined model of radiomics and deep learning has high performance in predicting the curative effect of TACE in patients with HCC before operation.

BACKGROUND: Cholecystectomy is a standard surgery for patients suffering from gallbladder diseases, while the causal effects of cholecystectomy on colorectal cancer (CRC) and other complications are still unknown. METHODS: We obtained genetic variants associated with cholecystectomy at a genome-wide significant level ( P value <5 × 10 -8 ) as instrumental variables (IVs) and performed Mendelian randomization (MR) to identify the complications of cholecystectomy. Furthermore, the cholelithiasis was also treated as the exposure to compare its causal effects to those of cholecystectomy, and multivariable MR analysis was carried out to judge whether the effect of cholecystectomy was independent of cholelithiasis. The study was reported based on Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization guidelines. RESULTS: The selected IVs explained 1.76% variance of cholecystectomy. Our MR analysis suggested that cholecystectomy cannot elevate the risk of CRC (odds ratio [OR] =1.543, 95% confidence interval [CI]: 0.607-3.924). Also, it was not significant in either colon or rectum cancer. Intriguingly, cholecystectomy might decrease the risk of Crohn's disease (OR = 0.078, 95% CI: 0.016-0.368) and coronary heart disease (OR = 0.352, 95% CI: 0.164-0.756). However, it might increase the risk of irritable bowel syndrome (IBS) (OR = 7.573, 95% CI: 1.096-52.318). Cholelithiasis could increase the risk of CRC in the largest population (OR = 1.041, 95% CI: 1.010-1.073). The multivariable MR analysis suggested that genetic liability to cholelithiasis could increase the risk of CRC in the largest population (OR = 1.061, 95% CI: 1.002-1.125) after adjustment of cholecystectomy. CONCLUSIONS The study indicated that cholecystectomy might not increase the risk of CRC, but such a conclusion needs further proving by clinical equivalence. Additionally, it might increase the risk of IBS, which should be paid attention to in clinical practice.
Humans ; Mendelian Randomization Analysis ; Irritable Bowel Syndrome ; Colorectal Neoplasms/genetics* ; Cholelithiasis/complications* ; Cholecystectomy/adverse effects* ; Genome-Wide Association Study ; Polymorphism, Single Nucleotide

BACKGROUND: Management of gastric leak after sleeve gastrectomy (SG) is challenging due to its unpredictable outcomes. We aimed to summarize the characteristics of SG leaks and analyze interventions and corresponding outcomes in a real-world setting. METHODS: To retrospectively review of 15,721 SG procedures from 2010 to 2020 based on a national registry. A cumulative sum analysis was used to identify a fitting curve of gastric leak rate. The Kaplan-Meier method and log-rank tests were performed to calculate and compare the probabilities of relevant outcomes. The logistic regression analysis was conducted to determine the predictors of acute leaks. RESULTS: A total of 78 cases of SG leaks were collected with an incidence of 0.5% (78/15,721) from this registry (6 patients who had the primary SG in non-participating centers). After accumulating 260 cases in a bariatric surgery center, the leak rate decreased to a stably low value of under 1.17%. The significant differences presented in sex, waist circumference, and the proportion of hypoproteinemia and type 2 diabetes at baseline between patients with SG leak and the whole registry population ( P = 0.005, = 0.026, <0.001, and = 0.001, respectively). Moreover, 83.1% (59/71) of the leakage was near the esophagogastric junction region. Leakage healed in 64 (88.9%, 64/72) patients. The median healing time of acute and non-acute leaks was 5.93 months and 8.12 months, respectively. Acute leak (38/72, 52.8%) was the predominant type with a cumulative reoperation rate >50%, whereas the cumulative healing probability in the patients who required surgical treatment was significantly lower than those requring non-surgical treatment ( P = 0.013). Precise dissection in the His angle area was independently associated with a lower acute leak rate, whereas preservation ≥2 cm distance from the His angle area was an independent risk factor. CONCLUSIONS Male sex, elevated waist circumference, hypoproteinaemia, and type 2 diabetes are risk factors of gastric leaks after SG. Optimizing surgical techniques, including precise dissection of His angle area and preservation of smaller gastric fundus, should be suggested to prevent acute leaks.
Humans ; Male ; Retrospective Studies ; Diabetes Mellitus, Type 2/complications* ; Obesity, Morbid ; Anastomotic Leak/epidemiology* ; Gastrectomy/methods* ; Reoperation/methods* ; Registries ; Laparoscopy/methods* ; Treatment Outcome

Internal hernia is the protrusions of the viscera through the peritoneum or mesentery in the abdominal cavity. It lacks specific clinical symptoms and has a poor prognosis. According to the anatomical classification, the intra-abdominal hernia can be divided into Duodenal paralysis, Winslow hernia, mesenteric hernia and mesenteric hiatal hernia. Among various imaging methods, multi-slice spiral CT has a higher value for preoperative diagnosis of internal hernia. This article will focus on the imaging features of several intra-abdominal hernias under multi-slice spiral CT, and summarize the general diagnostic strategies of internal hernia.

The purpose of this study is to develop glipizide push-pull osmotic pump (PPOP) tablets by using a formulation design expert system and an artificial neural network (ANN). Firstly, the expert system for the formulation design of osmotic pump of poor water-soluble drug was employed to design the formulation of glipizide PPOP, taking the dissolution test results of Glucotrol XL as the goal. Then glipizide PPOP was prepared according to the designed formulations and the in vitro dissolution was carried out. And in vivo evaluation was carried out between the samples which were similar to Glucotrol XL and the Glucotrol XL in Beagle dogs. The range of the factors of formulation and procedure, which could influence the drug release, was optimized using artificial neural network. Finally, the design space was found. It was found that the target formulation which was similar to Glucotrol XL in dissolution test could be obtained in a short period by using the expert system. The samples which were similar to Glucotrol XL were bio-equivalent to the Glucotrol XL in Beagle dogs. The design space of the key parameter coating weight gain was 9.5%-12.0%. It could be concluded that a well controlled product of glipizide PPOP was developed since the dissolution test standard of our product was more strict than that of Glucotrol XL.
Animals ; Area Under Curve ; Delayed-Action Preparations ; Dogs ; Drug Compounding ; methods ; Drug Delivery Systems ; Drug Design ; Expert Systems ; Female ; Glipizide ; administration & dosage ; chemistry ; pharmacokinetics ; Hypoglycemic Agents ; administration & dosage ; chemistry ; pharmacokinetics ; Male ; Neural Networks (Computer) ; Osmosis ; Polyethylene Glycols ; chemistry ; Random Allocation ; Sodium Chloride ; chemistry ; Solubility ; Tablets

BACKGROUND: 5-azacytidine (5-Aza) has been frequently used to induce bone marrow mesenchymal stem cells (BMSCs)differentiation into cardiomyocyte.OBJECTIVE: To observe expression of cardiomyocyte-related receptors in cardiomyogenic differentiation of rat BMSCs.METHODS: BMSCs of passage three were assigned to four groups: group Ⅰ: L-DMEM solution alone was replaced; Ⅱ:L-DMEM solution was replaced after induction of 100 mg/L AST+5 μmol/L 5-Aza for 24 hours; group Ⅲ: L-DMEM solution wasreplaced after induction of 10 μmol/L 5-Aza for 24 hours; and group Ⅳ: L-DMEM solution was replaced after induction of 5 μmol/L5-Aza for 24 hours. Culture medium was replaced every 3 days in each group. Differentiated cells were identified after 30 days ofinduction.RESULTS AND CONCLUSION: Expression of cardiomyocyte specific proteins Nkx2.5, cTnT and Desmin was detected in groupsⅢ, Ⅳ and Ⅱ after induction compared with group Ⅰ , with significant differences (P < 0.01). The amount of cTnT and Desminexpression expression was significantly higher in groups Ⅱ and Ⅲ compared with group Ⅳ (P < 0.01). The level of Nkx2.5expression was significantly higher in groups Ⅱ (P < 0.01) and Ⅲ (P < 0.05) compared with group Ⅳ. No Nkx2.5, cTnT andDesmin espression was detected in group Ⅰ. After induction for 2 weeks, cells with spontaneous contractility were observed ingroups Ⅱ and Ⅲ, indicating differentiation towards cardiomyocyte after induction. Results demonstrated that induction effectswere similar between 100 mg/L AST+5 μmol/L 5-Aza and 10 μmol/L 5-Aza. This may contribute to cytoprotective effects of AST,which can promote vascular endothelial cell proliferation, enhance celss tolerance to 5-Aza-induced cytotoxicity and upregulatecardiac-specific protein expression.