Assessment of transient elastography in Japanese patients with non-alcoholic fatty liver disease

Aim Transient elastography (TE) is a non-invasive method for predicting liver fibrosis. However, there are limited data regarding the performance of TE in Japanese patients with non-alcoholic fatty liver disease (NAFLD). We aimed to evaluate the association between liver stiffness measurement (LSM) by TE and liver fibrosis stage, and define a cut-off value for predicting liver fibrosis. Methods A total of 171 Japanese patients with biopsy-proven NAFLD underwent LSM using TE with FibroScan. The area under the receiver operating characteristic curve of LSM and other non-invasive markers of liver fibrosis were compared to determine the most accurate method of predicting liver fibrosis. Results Liver stiffness measurement significantly correlated with fibrosis stage (P<0.001). The areas under the receiver operating characteristic curve of LSM for fibrosis stage >= 1 and >= 3 was 0.85 and 0.91, respectively and were higher than those of the aspartate aminotransferase/alanine aminotransferase ratio, aspartate aminotransferase to platelet ratio index, fibrosis-4 index, and NAFLD fibrosis score. The best cut-off values of LSM fibrosis stage >= 1 and >= 3 were 7.2kPa (sensitivity 78.5%, specificity 78.3%) and 10.0kPa (sensitivity 89.5%, specificity 87.6%), respectively. The combination of LSM (>= 10kPa) and type IV collagen 7s (>= 6.0ng/mL) had a specificity of 97.6% for advanced fibrosis. The LSM in patients with high alanine aminotransferase levels or high body mass index was associated with false positive results regarding advanced fibrosis. Conclusions In NAFLD patients, TE has excellent utility for the assessment of liver fibrosis, particularly for advanced stage cases. The cut-off value of LSM by TE for predicting liver fibrosis stage >= 3 is 10.0kPa in Japanese NAFLD patients.