Objective Liver Fibrosis Estimation from Shear Wave Elastography

摘要: Diffuse liver disease is common, primarily driven by high prevalence of non-alcoholic fatty (NAFLD). It currently assessed biopsy to determine fibrosis, often staged as F0 (normal) - F4 (cirrhosis). A noninvasive assessment method will allow a broader population be monitored longitudinally, facilitating risk stratification and treatment efficacy assessment. Ultrasound shear wave elastography (SWE) promising technique for measuring tissue stiffness that has been shown correlate with fibrosis stage. However, this approach limited variability in measurements. In work, we developed evaluated an automated framework, called SWE-Assist, checks SWE image quality, selects region interest (ROI), classifies the ROI whether stage at or exceeds F2, which important clinical decisionmaking. Our database consists 3,392 images from 328 cases. Several classifiers, including random forest, support vector machine, convolutional neural network (CNN)) were evaluated. The best utilized CNN yielded area under receiver operating curve (AUROC) 0.89, compared conventional only based AUROC 0.74. Moreover, new on single per decision, vs. 10 decision baseline. larger dataset needed further validate approach, potential improve accuracy efficiency non-invasive staging.