[Diagnostic value of endoscopic ultrasonography, fibroscan, acoustic radiation pulse imaging, serological index, and their combination for early stage liver cirrhosis].

ABSTRACT

Objective: To investigate the diagnostic value of endoscopic ultrasonography (EUS), Fibroscan, acoustic radiation force impulse (ARFI), and aspartate aminotransferase-to-platelet ratio (APRI) and their combination for early stage liver cirrhosis. Methods: Three hundred and twenty-two hospitalized patients who had been diagnosed with chronic viral liver disease from March 2016 to April 2018 were included. According to the clinical diagnosis, patients were divided into chronic hepatitis and the early stage liver cirrhosis group (Child-Pugh A grade). All patients were examined by Fibroscan to detect liver stiffness measurement (LSM), ARFI to detect liver virtual touch tissue quantification (VTQ) value, esophagogastroduodenoscopy and EUS to detect esophagogastric varices, laboratory and imaging examination. The index of EUS, Fibroscan, ARFI, and APRI was analyzed and the regression model was established by binary logistic regression, and the diagnostic efficacy of the above index and regression model for early stage of cirrhosis was evaluated by the area under a receiver operating characteristic curve (AUROCs). Results: An early stage cirrhosis group had significantly higher detection rate with EUS (esophagogastric varices), Fibroscan (LSM), ARFI (VTQ) and APRI than chronic hepatitis group [76.7% vs. 10.7%, 10.4 (7.8, 17.3) vs. 6.1 (5.2, 8.4) kPa, 1.71(1.48, 2.07) m/s vs. 1.25(1.14, 1.43) m/s and 0.65 (0.38, 1.15) vs. 0.38(0.26, 0.62), respectively]. The corresponding chi-square test were 140.86, Z = -9.069, Z = -9.948 and Z = -5.764, respectively and the differences were statistically significant (P < 0.01). The areas under the receiver operating characteristic curve and regression model were 0.830 (0.783 ~ 0.877), 0.793 (0.744 ~ 0.841), 0.821 (0.775 ~ 0.868), 0.686 (0.628 ~ 0.744) and 0.947 (0.925 ~ 0.969) for the diagnosis of early stage cirrhosis, respectively. Among them, the regression model of three indices (EUS, LSM and VTQ) had the largest AUROCs (0.947) and the corresponding sensitivity and specificity were 0.878 and 0.867, respectively. Conclusion: The combination of EUS, LSM and ARFI had a superior diagnostic value for early stage liver cirrhosis, and may improve the diagnosis rate and reduce the misdiagnosis rate.