Ultrasound based elastographic methods for the prediction of esophageal varices in liver cirrhosis

A Popescu; I Sporea; R Lupusoru; AM Stepan; R Sirli; M Danila; T Moga; C Pienar; R Mare; F Bende; B Miutescu

doi:10.1055/s-0036-1587747

Ultraschall in der Medizin - European Journal of Ultrasound, Table of Contents

Ultrasound based elastographic methods for the prediction of esophageal varices in liver cirrhosis

A Popescu ¹, I Sporea ¹, R Lupusoru ¹, AM Stepan ¹, R Sirli ¹, M Danila ¹, T Moga ¹, C Pienar ¹, R Mare ¹, F Bende ¹, B Miutescu ¹

¹"Victor Babeş" Univesity of Medicine and Pharmacy Timişoara, Gastroenterology and Hepatology, Timisoara, Romania

Abstract

Purpose: Ultrasound based elastographic methods are non-invasive techniques for the evaluation of liver stiffness (LS) that might be also useful in the assessment of portal hypertension. The aim of this study was to evaluate the performance of 4 ultrasound based elastographic methods for predicting the presence of esophageal varices (EV) in patients known with liver cirrhosis.

Material and method: The study included 109 consecutive subjects diagnosed with liver cirrhosis (with clinical, biological, ultrasound, endoscopical or histological signs of liver cirrhosis), in whom LS was evaluated in the same session by means of 4 elastographic methods: transient elastography (TE) (M and XL probes), ARFI (VTQ), 2D-SWE and 2D-SWE.GE. Reliable LS measurements were defined as: for TE, VTQ and 2D-SWE.GE-the median value of 10 measurements and for 2D-SWE the median value of 3 measurements acquired in a homogenous area. In 60 patients out of 109 all 4 elastographic methods had valid measurements and were included in the final analysis.

Results: 20/60 patients from the study group had EV while 40/60 had not. The mean LS values for patients without EV were lower as compared to those of patients with EV: TE (21.60 ± 9.24 vs. 29.43 ± 11.23, p = 0.005), 2D-SWE (20.24 ± 12.60 vs. 24.38 ± 15.24, p = 0.26), ARFI (2.54 ± 0.71 vs. 2.69 ± 0.64, p = 0.42), 2D-SWE.GE (12.06 ± 6.52 vs. 12.95 ± 3.49, p = 0.49)

TE had the best performance for predicting EV. The following cut-off were establish for predicting the EV: TE ³21.1 kPa, Se-94.4%, Sp-57.1%, NPV-48.6%, PPV-96%, AUROC = 0.76, p < 0.0001; ARFI (VTQ): > 2.01 m/s, Se-88.9%, Sp-45.6%, NPV-48.6%, PPV-96%, AUROC 0.56, p = 0.002; 2D-SWE: > 13.7 kPa, Se-100%, Sp-48.6%, NPV-100%, PPV-37.5%, AUROC = 0.61, p = 0.05; 2D-SWE.GE: > 13.48 kPa, Se-77.8%, Sp-64.3%, NPV-87.1%, PPV-48.3%, AUROC = 0.69, p = 0.1.

Conclusion: LS values assessed by any ultrasound based elastographic methods are higher in patients with EV as compared to those without EV, but TE seems to be the most predictive for the presence of EV.