Abstract: Objective To study on the differential diagnosis value of T2 weighted image (T2WI) sequence image texture analysis (TA) of magnetic resonance imaging (MRI) in the differential diagnosis of hepatic infiltrative hepatocellular carcinoma (iHCC) and hepatic focal fusion fibrosis (FFF). Methods Fifty cirrhotic patients with invasive liver cancer were enrolled in the iHCC group. Thirty-two patients with focal fusion fibrosis with complete preoperative magnetic resonance imaging (MRI) data and pathological confirmation of their surgical specimens were enrolled in the FFF group. The clinical data of both groups of patients were collected and retrospectively analyzed. All patients underwent MRI examination and TA sequence analysis on T2WI at one week before the operation. The maximum axial T2WI images of the lesion were selected, and the region of interest (ROI) was manually sketched by Image J Image processing software. The spatial gray co-occurrence matrix (GLCM) texture parameters (entropy, energy, correlation, contrast, deficit distance) and Histogram texture parameters (mean, kurtosis, skewness) were extracted. GLCM and Histogram texture parameters were compared between the two groups. Receiver operating characteristic curve (ROC) was used to evaluate their efficiencies in differential diagnosis of invasive liver cancer and focal hepatic fusion fibrosis. Results The entropy was 7.04±0.29 in iHCC group, which was significantly higher than that of 6.70±0.60 in FFF group (P<0.05). The energy and correlation of the iHCC group were (15.30±7.19) ×10^-4 and (15.41±7.28) ×10^-4, which were significantly lower than those of (25.02±11.37) ×10^-4 and (33.65±14.40) ×10^-4 in FFF group, respectively (P<0.05). There were no significant difference in contrast, deficit distance, mean, kurtosis and skewness parameters between the two groups (P>0.05). The entropy, energy and correlation parameters have had certain efficiencies in the differential diagnosis of invasive liver cancer and liver focal fusion fibrosis, and the areas under curves (AUC) were 0.672, 0.701 and 0.812, respectively. Conclusion The entropy, energy and correlation of GLCM texture parameters under T2WI TA are useful in distinguishing infiltrating liver cancer and focal fusion liver fibrosis, which are worthy of further clinical verification.

Key words: T2-weighted image, Texture analysis, liver cirrhosis with infiltrative hepatocellular carcinoma, Focal fusion fibrosis