Dissertation > Excellent graduate degree dissertation topics show

Establishment of Mathematic Model of Laboratory Diagnosis for Liver Fibrosis and Clinical Application of the Model

Author: ZhangZuo
Tutor: WangNianYue
School: Jiangsu University
Course: Clinical Laboratory Science
Keywords: Liver Fibrosis Arithmetical Model Serological Markers Logistic Regression Analysis CTGF ELISA
CLC: R575.2
Type: Master's thesis
Year: 2009
Downloads: 79
Quote: 1
Read: Download Dissertation

Abstract

Aims:To establish an arithmetical formula for predicting the severity and progression of liver fibrosis by screening markers relating to producing and progression of liver fibrosis and comparing the levels of the markers with corresponding results of pathological diagnosis,search new pathway for noninvasive diagnosis of liver fibrosis by validating the arithmetical formula for exploring research.To decrease cost of the tests for universal application of the model,an enzyme linked immunoabserbent assay(ELISA) for accurately detecting serum connective tissue growth factor(CTGF) was developed.Methods:collecting the patient’s histories and serums of 322 patients performed liver biopsy and admitted to Nanjing second hospital from March 2005 to December 2009.Some markers related to liver fibrosis were measured by sandwich ELISA.The patients were divided into training group and validating group according to the data performing liver biopsy.Twenty variables and the results of pathological diagnosis of all patients were input to compute,different endpoints of liver fibrosis were set according to the stage of liver fibrosis(significant fibrosis:≥S2, cirrhosis:S4).The area under receiver operation characteristics(ROC) curve(AUC) and the correlations between markers and the fibrosis stage were computed in training group.To establish the formulation predicting liver fibrosis and validate the accuracy of the formula,all markers were univariate analysis,logistic regression analysis and Fisher’s discriminate analysis.With anti-CTGF monoantibodies and polyantibodies and biotin-avidin amplifing system,to develop a quantitate ELISA for detecting serum CTGF.Results:Significant differences were found in thirteen variables between patients with significant fibrosis and mild fibrosis and patients with cirrhosis and no cirrhosis.The AUC of thirteen variables discriminating significant fibrosis and cirrhosis exceeded 0.7.Fibrosis stage correlated with age(years) of patients and the levels of connective tissue growth factor(CTGF)(ng/mL),AST/ALT ratio, gamma-glutamyltransferase(GGT)(ng/mL),Hyaluronic acid (HA)(ng/mL),and platelet count(10~9/L).Liver cirrhosis and significant fibrosis could be accurately predicted by combining the five variables. The formula established by logistic regression were named fibrosis index, including CTGF,AST/ALT ratio,PLT,HA.AUC of the formula for predicting significant fibrosis was 0.83(0.73-0.96) and 0.92(0.83-0.98) for liver cirrhosis in validation set.The accuracies of Fish’s discriminate analysis separating significant fibrosis and liver cirrhosis were 88.5%and 93.2%by combining the five variables CTGF,AST/ALT ratio,PLT,HA, GGT in validation set.To simplify the formula and decrease cost,the logistic regression model was recommended for clinics.With good precision and stability,a biotin-avidin ELISA system for detecting serum CTGF was developed,the system had strong correlation with CTGF ELISA kit from oversea.Conclusions:Most of patients with significant fibrosis and liver cirrhosis could be predicted with good sensitivity and specificity by using the logistic formula,which will substitute the liver biopsy for some patients to survey the progression of chronic hepatitis and liver fibrosis.With good specification,sensitivity,simple performance and low cost, the ELISA kit detecting serum CTGF could be widely applicated in institutions and hospitals.

Related Dissertations

CLC: > Medicine, health > Internal Medicine > Digestive and abdominal diseases > Liver and gall bladder disease > Cirrhosis
© 2012 www.DissertationTopic.Net Mobile