Study On The Diagnosis Model And Thrombolysis Treatment Of Portal Vein Thrombosis

Purpose:Portal vein thrombosis(PVT)is a pathophysiological process which was caused partial or complete obstruction of the blood flow in the portal vein and its branches due to abnormal clotting of the blood in the portal vein.It is a rare clinical disease with atypical clinical manifestations.It was resolved to the extensive portal vein obstruction as soon as possible,and intestinal perforation,peritonitis,shock,multiple organ failure,and even death may occur.Extrahepatic portal hypertension(EPH)may occur in PVT untreated patients with clinical manifestations such as esophageal varices,portal hypertension,portal biliopathy gastrointestinal disease,recurrent cholangitis,hematemesis and melena.Existing literature suggests that as soon as possible to start treatment may improve the prognosis of PVT patients.The purpose of this study is to construct the PVT diagnosis model,evaluation of the diagnosis model in the diagnosis of PVT,and compare the clinical efficacy of the systemic thrombolysis and indirectly transcatheter thrombolytic therapy of PVT follow the former two study,and then provide evidence for clinical diagnosis and treatment of PVT.Method:1.The clinical date such as history,clinical manifestations,laboratory examination,physical sign of PVT patient from January 2005 to 2016 June was analysis retrospectively?To build model using R software,the PVT as the dependent variable,43 clinical data as independent variables into the calculation model,by Lasso algorithm with 10 fold cross operation.Then the significance variables by Lasso selected were included into the SVM model.All samples were built the trained set and all the original samples as test set were used to validate the model classification ability.2.The patients with three phase enhanced CT scan were predicted through the Lasso-SVM model to the diagnosis of PVT,and these patients were diagnosis with CDUS.The capacity of CUDS and diagnostic model for the diagnosis of PVT was evaluated.The diagnostic value of three phase contrast enhanced CT and CDUS was compared according the anatomic sites and the indirect signs of PVT.3 The 41 atients were included according to inclusion exclusion criteria who diagnosed as PVT by the first part of the diagnostic model,and confirmed PVT by ultrasound,CT angiography,MRI angiography or DSA angiography.As the treatment methods were divided into two groups: Thrombolysis group(ST,n=20)and SMA indirect catheter thrombolysis group(ICT,n=21).ST group was given low molecular weight heparin calcium injection 100u/kg q12 h,while the peripheral venous infusion of Urokinase infusion,40-60 million u/d divided into two times to intravenous drip.SMA g Cobra catheter was indwell in ICT group under intervention method,while given urokinase 40-60 million u/d trough the catheter by microinjection pump to continuous infusion.Result:1The 13 significance variable was gained according to the Lasso algorithm,such as cirrhosis,splenomegaly,two D-dimer,splenectomy,thrombophilia tendency,ascites syndrome,recent history of abdominal surgery,abdominal distension,C-reactive protein which related positively to diagnosis PVT,on the contrary,albumin and abdominal tenderness which was negatively correlated in the diagnosis of PVT.The variables selected by Lasso were included in the SVM model,and the sensitivity of the model forecasting PVT was 91.5%,specificity was 100%,positive predictive value was 100%,negative predictive value was 95.9%.It can be seen that the classification accuracy of the model is 97.16% and the error rate is about 2.84%.2.The LASSO-SVM model predicts PVT with 43 true positives,4 false positives,0 false negatives,and 94 true negatives.CDUS check out the true positive patients 46 people,false negative 1 people,false positive 4 people,true negative 90 people in 94 PVT patients.The specificity of CDUS for PVT diagnosis was 0.957(95.7%),and the sensitivity was 0.979(97.9%).The specificity of the diagnostic model was 1(100%),and the sensitivity was 0.915(91.5%).The positive predictive value of CDUS was 0.920(92%),and the negative predictive value was 0.989(98.9%).The positive predictive value of the diagnostic model was 1(100%),and the sensitivity was 0.959(95.9%).It was showed that the negative predictive value of the two methods was not statistically significant(p=0.1864),and the positive predictive value was statistically significant(p=0.0369).The detection ability of intestinal ischemia was weak by CDUS examination with Kappa value 0.495.3.The catheter success rate was 100% in CT group.In group ST,the average time of thrombolysis was 5.65 ±2.01 days,CT group was 4.71 ±1.19 days,and the dosage of urokinase in the ST group was an average of 193.33 ±76.22 million units,in the CT group,with an average of 223.50±31.67 million units,and there was no statistical difference between the two groups.In group ST,there was 2 cases hematuria,and one case deep venous thrombosis in the lower limb,In Group ICT,there was 1 cases puncture site bleeding and 1 cases hematochezia,improved after thrombolysis were suspended,no needed blood transfusion therapy;1 cases failure then transmit to surgical treatment following multiple organ dysfunction,and death on the hospitalized 20 days.There were 4 cases failure in ST group and 1 case in ICT group in which there were no significant differences in the rate of failure between the two groups.THE length of stay was 8-41 days,with an average of 16.20 ± 8.08 days in ST group,8-32 days,with an average of 16.67±5.82 days in ICT group,there was no significant difference between the two groups.There was no significant difference in the mortality rate between the two groups.The average time of symptom remission in group ST was 4.80±1.70 days,and 3.55±1.15 days in group ICT and there was significant difference between the two groups(p=0.014).The comparison of portal venous patency score showed that the recanalization of portal vein was different in different time and different groups,and the trend of portal vein recanalization varied with time.With the passage of time,the recanalization of blood flow increased gradually,and reached a peak after treatment 3months,and then the tended to be mild after treatment 6months.Patients were followed up for an average of 28.60± 8.84 months in ST group,the average of 30.75 ±8.87 months in the ICT group,and the recurrence rate of the patients in the two groups no significant difference.9 cases with portal vein cavernous(CTPV)in ST group,4 cases in ICT group,there was no significant difference between the two groups.In group ST,there were 12 cases with portal hypertension and esophageal or gastric varices,while 6 cases in ICT group,no significant difference between the two groups was fund(p=0.043).At the end of follow-up,1 cases died of liver cancer in group ST after 29 months,1 cases died of cirrhosis after 21 months;after19 months,1 cause died of unknown;1 cases in group ICT after 31 months died of cirrhosis of the liver,after 23 months,1 cases died of liver cancer,and one patient died within 30 days.There was no significant difference in mortality rate between the two groups,and there was no significant difference in survival curve between the two groups.Conclusion:The Lasso algorithm is used to select the variables,and the diagnosis model is built into the SVM model.Compared CDUS with the PVT diagnosis model to judgment the real patient of PVT,CDUS lack anatomical information to provide,in order to obtain more intuitive and adequate PVT related anatomical information,reduce medical costs,save time for examination,diagnosed by PVT model for PVT patients can directly for three phase enhanced CT check.Patients with PVT validate the SVM model were given by the SMA catheter thrombolysis and system thrombolysis therapy,and used ultrasound as a follow-up tool to assess the hemodynamics of the portal vein.It was found that SMA catheter thrombolysis can quickly relieve clinical symptoms,accelerate the blood flow of portal vein recanalization,reduce the incidence of portal hypertension,gastrointestinal diseases and esophageal varices the.Recommended color Doppler ultrasound can be used as a follow-up tool for PVT.