Accuracy of liver stiffness-based model by different imaging modalities in compensated advanced chronic liver disease

A liver stiffness × spleen size/platelet count score (LSPS) model which can rule out high-risk varices and identify high likelihood of clinically significant portal hypertension in patients with compensated cirrhosis has been endorsed by American Association for the Study of Liver Diseases in the 2016 practice guidance on portal hypertension bleeding. This study aims to evaluate the accuracy of LSPS model assessed by ultrasound in well characterized patients with compensated advanced chronic liver disease.

Eligible patients with compensated advanced chronic liver disease were retrospectively enrolled between January 2017 and March 2018, who had undergone routine clinical and laboratory tests, liver stiffness measurement, ultrasound examination, and computed tomography scanning. Spleen sizes were evaluated by ultrasound and computed tomography reconstructed model, respectively. The correlation and agreement of spleen size and LSPS derived from ultrasound and computed tomography imaging modality were compared.

A total of 158 patients were included and analyzed. Spleen size showed a moderate correlation (R = 0.649, P < 0.001) according to ultrasound and computed tomography imaging. Also, the correlation between the two LSPS models based on ultrasound and computed tomography was excellent (R = 0.985, P < 0.001). The Bland-Altman plots demonstrated a superior agreement of LSPS model values evaluated by ultrasound and computed tomography, respectively.

This study demonstrated the accuracy of LSPS model based on ultrasound in a well characterized cohort of fully compensated patients with advanced chronic liver disease.