Liver fibrosis assessment by FibroScan compared with pathological findings of liver resection specimen in hepatitis C infection

Background & Aims FibroScan® is a tool for the noninvasive diagnosis of hepatic fibrosis. Previous studies have compared liver stiffness to percutaneous liver biopsy findings, but there were no studies which compared liver stiffness to liver resection specimen findings. The aim of this study was to compare FibroScan® measurements to resected liver specimen findings. Methods From April 2011 to November 2015, a total of 114 patients with liver tumor and hepatitis C were enrolled. We divided them into two groups (training set and validation set). The first 66 patients who had liver resection from April 2011 to June 2014 were considered the training set, and the other 48 patients from August 2014 to November 2015 formed the validation set. Liver stiffness was measured by FibroScan® before surgery, and specimens obtained by liver resection were evaluated according to the METAVIR system. Results Using Spearman rank correlation analysis, a positive correlation between liver stiffness measurement and liver fibrosis stage was observed (r = 0.786, p < 0.0001). In the training set, Area under receiver operating curves for diagnosis of F ≥ 2 was 0.971 (95% CI: 0.928-1.000, cut-off value: 5.9), for diagnosis of F ≥ 3 was 0.911 (0.825-0.997, 9.8), and for diagnosis of F = 4 was 0.917 (0.849-0.985, 15.5). In the validation set, at a cut-off value of 5.9 kPa, sensitivity, specificity, positive predictive values and negative predictive values for F ≥ 2 were 95.7%, 0.0%, 97.8% and 0.0% respectively, of 9.8kPa for F ≥ 3 were 86.2%, 52.6%, 73.5% and 71.4%, and of 15.5kPa for F = 4 were 100%, 71.8%, 45.0% and 100%. The distribution of the stiffness values according to fibrosis stage and the curves were examined to obtain the most discriminant cut-off values, which were 5.9 kPa for F ≥ 2, 9.8 kPa for F ≥ 3, and 15.5 kPa for F = 4. When liver stiffness determined by FibroScan® was set at the cut-off values, the sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy were high in the training set. These cut-off values were adequate for application in the validation set. Conclusions The stage of stiffness graded by FibroScan® has a good correlation with the liver fibrosis of resected liver specimens. It has the ability to diagnose fibrosis stage noninvasively.