Purpose To explore the diagnostic efficacy of texture analysis in grading multiple ligament injuries of the knee Joint.Materials and Methods MRI images of 203 patients with knee injuries in the Second Affiliated Hospital of Shantou University School of Medicine from June 2018 to September 2020 were retrospectively collected,and then graded the injuries of multiple ligaments.Extraction of texture features and discriminant analysis were accomplished by MaZda software.Spearman correlation analysis and LASSO penalized Logistic regression analysis were used to further explore the differential diagnostic efficacy of each texture parameter for grading knee ligament injuries.Results(1)MaZda software was found to have the highest diagnostic efficacy for patellar ligament injury grading by using different texture feature screening methods in combination with discriminant analysis.Non-linear discriminant analysis had the lowest misclassification rate of 4.43%.(2)The texture parameters that were ultimately included in the analysis of the subjects'work characteristic curves appeared more frequently in the gray level co-occurrence matrix and the run length matrix.(3)The area under the curve of the texture parameters finally included in the analysis were 0.870-0.975,0.772-0.912,0.836-0.929 and 0.766-0.920,respectively in patellar ligament,anterior cruciate ligament,posterior cruciate ligament and lateral collateral ligament for the diagnosis of knee ligament injury grading.The diagnostic efficacy for patellar ligament injury grading was the highest,which was consistent with the lowest misclassification rate for patellar ligament in the discriminant analysis of texture features.And the texture parameters ultimately included in patellar ligament were all highly correlated with its injury grading.Conclusion The texture analysis technique has high diagnostic efficacy for grading knee ligament injuries,and can provide a more accurate diagnostic imaging basis for the clinic.

Purpose To explore the diagnostic efficacy of texture analysis in grading multiple ligament injuries of the knee Joint.Materials and Methods MRI images of 203 patients with knee injuries in the Second Affiliated Hospital of Shantou University School of Medicine from June 2018 to September 2020 were retrospectively collected,and then graded the injuries of multiple ligaments.Extraction of texture features and discriminant analysis were accomplished by MaZda software.Spearman correlation analysis and LASSO penalized Logistic regression analysis were used to further explore the differential diagnostic efficacy of each texture parameter for grading knee ligament injuries.Results(1)MaZda software was found to have the highest diagnostic efficacy for patellar ligament injury grading by using different texture feature screening methods in combination with discriminant analysis.Non-linear discriminant analysis had the lowest misclassification rate of 4.43%.(2)The texture parameters that were ultimately included in the analysis of the subjects'work characteristic curves appeared more frequently in the gray level co-occurrence matrix and the run length matrix.(3)The area under the curve of the texture parameters finally included in the analysis were 0.870-0.975,0.772-0.912,0.836-0.929 and 0.766-0.920,respectively in patellar ligament,anterior cruciate ligament,posterior cruciate ligament and lateral collateral ligament for the diagnosis of knee ligament injury grading.The diagnostic efficacy for patellar ligament injury grading was the highest,which was consistent with the lowest misclassification rate for patellar ligament in the discriminant analysis of texture features.And the texture parameters ultimately included in patellar ligament were all highly correlated with its injury grading.Conclusion The texture analysis technique has high diagnostic efficacy for grading knee ligament injuries,and can provide a more accurate diagnostic imaging basis for the clinic.

Objective:To construct a risk prediction model for the inter-hospital transfer of critically ill children using machine learning methods, identify key medical features affecting transfer outcomes, and improve the success rate of transfers.Methods:A prospective study was conducted on critically ill children admitted to the pediatric transfer center of Hunan Children's Hospital from January 2020 to January 2021. Medical data on critical care features and relevant data from the Pediatric Risk of Mortality (PRISMⅢ) scoring system were collected and processed. Three machine learning models, including logistic regression, decision tree, and Relief algorithm, were used to construct the risk prediction model. A back propagation neural network was employed to build a referral outcome prediction model to verify and analyze the selected medical features from the risk prediction model, exploring the key medical features influencing inter-hospital transfer risk.Results:Among the 549 transferred children included in the study, 222 were neonates (40.44%) and 327 were non-neonates (59.56%). There were 50 children in-hospital deaths, resulting in a mortality rate of 9.11%. After processing 151 critical care medical feature data points, each model selected the top 15 important features influencing transfer outcomes, with a total of 34 selected features. The decision tree model had an overlap of 72.7% with PRISMⅢ indicators, higher than logistic regression (36.4%) and Relief algorithm (27.3%). The training prediction accuracy of the decision tree model was 0.94, higher than the accuracy of 0.90 when including all features, indicating its clinical utility. Among the top 15 important features selected by the decision tree model, the impact on transfer outcomes was ranked as follows based on quantitative feature violin plots: base excess, total bilirubin, ionized calcium, total time, arterial oxygen pressure, blood parameters (including white blood cells, platelets, prothrombin time/activated partial thromboplastin time), carbon dioxide pressure, blood glucose, systolic blood pressure, heart rate, organ failure, lactate, capillary refill time, temperature, and cyanosis. Eight of these important features overlapped with PRISMⅢ indicators, including systolic blood pressure, heart rate, temperature, pupillary reflex, consciousness, acidosis, arterial oxygen pressure, carbon dioxide pressure, blood parameters, and blood glucose. The decision tree was used to select the top 15 medical features with high impact on the neonatal and non-neonatal datasets, respectively. A total of 19 features were selected, among which there were 8 differences and 11 overlap terms between the important features of the neonatal and non-neonatal.Conclusions:Machine learning models could serve as reliable tools for predicting the risk of inter-hospital transfer of critically ill children. The decision tree model exhibits superior performance and helps identify key medical features affecting inter-hospital transfer risk, thereby improving the success rate of inter-hospital transfers for critically ill children.

Objective To explore the value of non-contrast CT findings of acute ischemic stroke(AIS)for predicting early prognosis after mechanical thrombectomy.Methods Data of 161 AIS patients from clinical center 1 who underwent mechanical thrombectomy were retrospectively analyzed.The patients were divided into training set(n=113)and internal test set(n=48)at the ratio of 7∶3,while 79 AIS patients who underwent mechanical thrombectomy from clinical center 2 were retrospectively enrolled as external test set.According to the National Institutes of Health stroke scale(NIHSS)scores 7 days after thrombectomy,patients'prognosis were classified as good(＜15 points)or poor(≥15 points).Pre-treatment non-contrast CT images of patients were reviewed,and CT findings were comparatively analyzed.Independent predictors of patients'early prognosis after mechanical thrombectomy were obtained with sequential univariate and multivariate logistic regressions,and a predicting model was established and visualized as a nomogram.The receiver operating characteristic curve was drawn,and the distinction was assessed with the area under the curve(AUC),then calibration was assessed with Hosmer-Lemeshow goodness of fit test,and the net benefit was evaluated with decision curve analysis(DCA).Results Alberta stroke program early CT score(ASPECTS),hyperdense middle cerebral artery sign(HMCAS)and basal ganglia calcification were all independent predictors of early prognosis of AIS after mechanical thrombectomy(all P＜0.05).The predictive model was established combining the above 3 variables and then visualized as a nomogram to predict prognosis of AIS after mechanical thrombectomy,with AUC of 0.776 in internal test set(χ2=6.052,P=0.417)and 0.800 in external test set(χ2=2.269,P=0.811).DCA showed that the nomogram might provide clinical net benefit within certain threshold probability ranges.Conclusion ASPECTS,HMCAS and basal ganglia calcification were all independent predictors of early prognosis of AIS after mechanical thrombectomy.The nomogram originated from predicting model combining the three could be used to somewhat accurately predict poor early prognosis after mechanical thrombectomy.

Objective:To evaluate the diagnostic efficiency of contrast-enhanced magnetic resonance angiography (CE-MRA), T 1WI and three-dimensional thin-layer T 1WI sequences (3D-T 1WI) in evaluating the blood supply of brain tumors, so as to improve the enhanced magnetic resonance imaging (MRI) scanning scheme of brain tumors. Methods:After the contrast agent was injected, 29 patients with brain tumors were scanned with CE-MRA, T 1WI and 3D-T 1WI. The imaging manifestations of " observing the enhancement of tumor substance" , " observing the small vessels in the tumor body" and " observing the small vessels around the tumor" were qualitatively classified, and the diagnostic differences of the three sequences were analyzed by McNemar test. Results:⑴ In the aspect of " observing the contrast medium in the tumor body" , CE-MRA, T 1WI and 3D-T 1WI sequences diagnosed as grade A1 were 0 cases, 8 cases (27.59%) and 19 cases (65.52%) respectively. The diagnostic efficiency of 3D-T 1WI sequences was better than that of T 1WI sequence ( P<0.05), and T 1WI sequence was better than that of CE-MRA sequence ( P<0.05). ⑵ In the aspect of " observing the small vessels in the tumor" , CE-MRA, T 1WI and 3D-T 1WI sequences diagnosed as grade B1 were 8 cases (27.59%), 5 cases (17.24%) and 16 cases (55.17%), respectively. The diagnostic efficiency of 3D-T 1WI was better than that of T 1WI and CE-MRA ( P<0.05). ⑶ In the aspect of " observing the small vessels around the tumor" , CE-MRA, T 1WI and 3D-T 1WI sequences diagnosed as grade C1 were 18 cases(62.07%), 5 cases (17.24%) and 14 cases (48.28%) respectively. The diagnostic efficiency of CE-MRA and 3D-T 1WI wre better than that of T 1WI ( P<0.05). Conclusions:The combination of 3D-T 1WI and CE-MRA sequence can comprehensively evaluate the blood supply of the tumor, which is of great significance for the localization, characterization and treatment of brain tumors.

Objective:To study the change of bifrontal metabolite concentration in patients with mild cognitive impairment (MCI) and its relationship with substantia alba demyelination using magnetic resonance spectroscopy (MRS) combined with linear combination of model (LCModel) quantitative technique.Methods:From May 2016 to December 2018, 25 patients with MCI (group A; 12 males, 13 females, age (60.5±5.2) years) and 15 healthy control subjects (group B; 6 males, 9 females, age (59.5±3.5) years) in the Second Affiliated Hospital of Shantou University Medical College were prospectively enrolled. The MCI patients were classified into 2 subgroups according to MRI results: group A1 with substantia alba demyelination (7 males, 4 females, age (62.1±3.9) years) and group A2 without substantia alba demyelination (5 males, 9 females, age (59.2±5.8) years). Software LCModel was used to quantitatively analyze the MRS original data and measure the absolute concentration of N-acetylaspartate compound (NAA), creatine compound (Cr), choline-containing compound (Cho), myoinositol (mI) and ratios of NAA/Cr, Cho/Cr, mI/Cr, NAA/mI in bilateral frontal lobe. Independent-sample t test was used to analyze the inter-group differences of the above parameters, while Pearson correlation analysis was performed to analyze correlations between the above parameters and cognitive function scores. Results:Compared with group B, group A had higher mI of both left and right frontal lobes (left: (5.19±1.28) vs (4.32±0.83), right: (4.87±1.11) vs (3.85±0.98); t values: 2.34, 2.93, both P<0.05); the mI/Cr of right frontal lobe in group A was also higher (1.19±0.31 vs 0.98±0.25; t=2.21, P<0.05), while the NAA/mI of right frontal lobe was lower (1.37±0.34 vs 1.78±0.47; t=-3.16, P<0.01). Differences of other parameters between group A and group B, and those between group A1 and group A2 were not significantly different ( t values: -1.70 to 1.35, all P>0.05). The mI of right frontal lobe was negatively correlated with Montreal Cognitive Assessment (MoCA) score and Mini-Mental State Examination (MMSE) score( r values: -0.35, -0.38, both P<0.05), on the contrary, NAA/mI of right frontal lobe was positively correlated with the cognitive function scores ( r values: 0.43, 0.40, both P<0.05). Conclusion:MCI may be related to the loss or dysfunction of neurons in the right frontal lobe, and MRS can provide theoretical basis for early recognition of MCI to some extent.

Objective To analyze the cause of the misdiagnosis of hepatolithiasisGassociated intrahepatic cholangiocarcinoma (IHHCC)by CT/MRI,and to improve early imaging diagnostic rate of IHHCC.Methods Thirteen patients histopathologically confirmed IHHCC after operation were retrospectively analyzed.1 1 cases had CT scans with contrast enhancement.5 cases had MRI scans with contrast enhancement and magnetic resonance cholangiopancreatography(MRCP).Combining with the imaging features of IHHCC (including the soft tissue mass shadow in the bile duct,the dilation and stenosis of the bile duct,the shriveling of the liver capsule,and etc.),medical history and serum tumor markers,misdiagnosis causes were analyzed.Results The causes of misdiagnosed IHHCC:(1 )High density bile duct stones in CT imaging concealed the soft tissue of adjacent cholangiocarcinoma;(2)The inconsistence of the degree and location of dilated bile duct with the location and size of bile dult stone was ignored;(3)Soft tissue lesions of cholangiocarcinoma with delayed enhancement were not carefully observed,or prolonged delayed scanning was not performed;(4)MRI scans and MRCP could show the tendrilGlike dilated intrahepatic bile duct,the filling defect of the bile duct wall with focally eccentric thickening,and stiffness of the bile duct;(5)Elder patients,long course of disease,recurrent attacks and increased CA19G9/CEA level could be factors related to IHHCC.Conclusion The early stage of hepatolithiasis combined with cholangiocarcinoma is easy to be misdiagnosed for cholelithiasis and cholangitis.DelayGenhanced CT,MRI scanning and MRCP are helpful to improve preoperative diagnostic rate for the patients with elder age,longer course of disease,repeated onset,increased CA1 9G9/CEA level and inconsistent degree of peripheral biliary dilatation with the size and location of the calculi.

Objective To discuss the clinical efficacy of improved TIPS of percutaneous portal vein puncture in treatment of acute upper gastrointestinal bleeding induced by portal hypertension.Methods 28 patients with acute upper gastrointestinal bleeding underwent improved TIPS therapy in our hospital were enrolled.The clinical data,laboratory parameters and hemodynamic changes were collected and analyzed before and after operation.Results The success rate for the first time and hemostatic rate of postoperative 24 hours in all patients was 100%.2(7.14%)patients underwent mild hepatic encephalopathy.After TIPS operation,the concentration of serum albumin increased,whereas,concentration of total bilirubin and alanine aminotransferase decreased(P<0.01).Portal vein pressure (PVP)of pre-and post-operation was(41.48 ± 3.72)mmHg and(28.91 ± 2.59)mmHg,and the hepatic venous pressure gradient (HVPG)was(20.30 ± 2.76)mmHg and(8.81 ± 2.04)mmHg.PVP and HVPG were both decreased significantly after operation(P<0.01).Conclusion Improved TIPS therapy can obtain good clinical efficacy and safety for esophageal and gastric varicose bleeding in acute cirrhosis.

Objective To explore the metabolite profiles of mesenchymal stem cells(MSCs)underwent death using 9.4 T high resolution MR spectroscopy.Methods MSCs were cultured and treated for 6,12 and 24 hours in a stimulated condition which included hypoxia,serum deprivation and changes of microenvironment.Cell death and the mortality was detected by light microscopy,Hocchst staining and flow cytometry analyses.The morality of stem cells was analyzed using one-way analysis of variance (ANOVA).Cell metabolite extraction was prepared by methanol-chloroform(M/C) method and analyzed on a 9.4 T MR device.1H-MR spectroscopy was obtained and the metabolite concentration of each time point was calculated and compared using one way ANOVA,the difference between two groups was analyzed by SNK test.Results Necrosis was the major form of cell death in the built model.The morality of every time sets was 16±4(0 h),658±61 (6 h),1 571 ± 154(12 h) and 2 816± 178(24 h) respectively,and the difference between each groups were statistically significant (F=298.96,P＜0.01).After induced stem cells death for 6,12 and 24 h,the metabolite concentrations at 0.89 ppm was (1.48±0.69),(2.32±0.63)and (2.15±0.45)nmol/mg respectively,and increased compared to thc control[(1.41 ±0.25)nmol/mg]with statistical significance (F=329.57,P＜0.01).The metabolite concentrations at 1.28 ppm was (6.42±0.31),(7.26±0.32)and (7.01 ±0.61)nmol/mg,respectively,and increased compared to the control[(5.76 ±0.74)nmol/mg]with statistical significance (F=19.56,P＜0.01).The metabolite concentrations at 1.60 ppm was (2.36±0.31),(2.29±0.16)and (2.31 ± 0.24) nmol/mg respectively,and increased compared to the control[(1.96 ± 0.27)nmol/mg]with statistical significance (F=4.35,P＜0.05).After induced stem cells death for 12 hours,the metabolite concentrations at 0.89 ppm was increased compared to 6 hours with statistical significance (P＜0.05).The metabolite concentrations at 1.28 ppm was increased compared to 6 hours with statistical significance (P＜0.05).After induced stem cells death for 24 hours,the metabolite concentrations at 0.89 ppm was decreased compared to 12 hours with statistical significance (P＜0.05).Conclusions There are some specific characteristics on MRS of MSCs underwent death,and the fatty acid peak may serve as a biomarker for cell death.

Objective To evaluate the absolute quantification of brain metabolites concentrations using external standard MRS in acute hypoxia ischemia encephalopathy (HIE) piglet model. Method Eight 7-day-old healthy piglets were subjected to insult of hypoxia ischemia (HI). The animals and an external standard phantom containing detectable metabolites of known concentrations were studied on a 1.5 T GE Signa scanner. The single-voxel proton magnetic resonance spectroscopy (1H-MRS) data were processed using LCModel software, and the quantification of N-acetylaspartate ( NAA), creatine (Cr) and lactate (Lac) were accomplished. Multivariate analysis of variance was performed to compare the NAA, Cr, Lac concentration differences in the brains of piglets pre- and post-HI (0h). In addition, the dynamic changes of brain metabolites concentrations of 2 HIE piglets were observed at the time points of 0 h and 2 h. Results One piglet was excluded because it was over anesthetized to death. Seven piglets' data were analyzed. The concentrations of NAA pre- and post-HI were ( 6. 86 ± 0. 49 ) mmol/kg and ( 5.73 ± 0. 88 ) mmol/kg respectively, they were ( 4. 65 ± 0. 73 ) mmol/kg and ( 4. 40 ± 0. 80 ) mmol/kg for Cr; and were 0. 00 mmol/kg and (0. 43 ± 0. 39) mmol/kg for Lac. After HI, decreased NAA concentration immediately was observed, and it was of statistical significance ( F = 8. 608, P = 0. 013 ). The concentration of Cr was insignificantly decreased ( F = 0. 379, P = 0. 550). The concentration of Lac was increased, and the difference was of statistical significance ( F = 8. 600 ,P = 0. 013 ). Dynamic observation showed a Lac peak immediately after HI and it decreased after 2 h post-HI. Conclusions External standard MRS using LCModel has great value in the quantitative analysis of brain metabolites. The changes of NAA and Lac concentrations are sensitive to reflect the early metabolic change of acute HIE.