Obsjectives:There are many ways of diagnosis for the HBV-rela tive cirrhosis,especially,the noninvasive diagnosis,which had been widely used. Before these tests are applied to the clinic,which should be made a diagnosis evaluation,for acquiring the specificity and sen-sitivity.Because of the ubperfect standard biopsy, which the diag-nosis evaluation depended on can’t be reliable.Without the gold st-andard, the aim of this research was to assess the accuracy of seve-ral noninvasive diagnosis for the HBV-relative cirrhosis by the laten tclass method with random effects.Methods:From2013June to2014March, we collect the clinical data of patients with the chronic hepatitis B(CHB) from the liver disease clinic. A total of575patients with CHB was assessed the fibrosis stage by the four techniques(FibroScan, APRI, FIB-4and A/G).The diagnostic evaluation was made by the latent class method with random effect which analysis the clinical data above to assess the accuracy of the four ways of noninvasive diagnosis.Result:The latent class model with random effects permitted to conciliate the observed data and estimates of test performances. For advanced fibrosis,the specificity/sensitivity was for APRI0.99/0.28FIB-40.99/0.98,FibroScan0.83/0.55;And for cirrhosis, APRI0.99/0.19,FIB-40.99/0.90,FibroSan0.93/0.36,A/G0.99/0.03.Conclusion:Without the use of a gold standard, the latent class model with random effects estimated the accuracy of these tests of noninvasive diagnosis. The FIB-4and FibroScan performance well in the diagnosis of fibrosis and cirrhosis. |