Patients and Methods
Between August 2008 and July 2011, we studied 246 HCV-infected patients with histologically proven liver tissues who underwent EOB-MRI. Clinical data and EOB-MRI were obtained within 1 month before the liver biopsy or operation. HCV infection was defined by a positive anti-HCV test and detection of HCV-RNA in either a quantitative or qualitative assay. Exclusion criteria were infection with hepatitis B or human immunodeficiency viruses, alcohol abuse, numerous liver tumors, and previous partial splenic arterial embolization (PSE) or splenectomy. Nine patients were excluded due to inadequate biopsy specimens for the fibrosis stage, two patients with HBV infection, five patients with high alcohol consumption, three who had undergone PSE, one who had undergone splenectomy, and two patients because of numerous liver tumors. Finally, 224 patients were enrolled in the study. Data were randomly split almost two to one for model estimation and model validation: 149 patients (male 99, female 50, mean age 71 ± 8) for estimation and 75 (male 45, female 30, mean age 65 ± 13) for validation.
Liver tissue was obtained by percutaneous needle biopsy using a 14–16G biopsy needle or from liver specimens resected during surgery. Biopsy specimens were stained with hematoxylin and eosin, and Masson's trichrome staining. Liver fibrosis was evaluated according to the METAVIR scoring system: F0, no fibrosis; F1, portal fibrosis without septa; F2, few septa; F3, numerous septa without cirrhosis; F4, cirrhosis. One pathologist, who was unaware of the patient characteristics, assessed and scored the specimens. Overall, fibrosis staging for estimation was 15(3)/24(1)/19(0)/46(2)/45(2) for F0/F1/F2/F3/F4, respectively, and the samples for validation were 11(2)/15(0)/12(1)/20(1)/17(1) for F0/F1/F2/F3/F4, respectively; () is surgical resection.
We used a 1.5-Tesla MR system (PHILIPS Co., Amsterdam, the Netherlands); 0.025 mmol/kg bodyweight gadoxetate disodium was injected intravenously, and quantitative measurements were conducted using unenhanced and gadoxetate disodium-enhanced imaging at 20, 35, 70, and 180 s. Hepatobiliary phases at 15, 20, and 25 min were obtained. Imaging parameters were as follows: repetition time/echo time = 4.17/2.05 ms. Then, 1- to 2-cm2 regions of interest of the mean SI value of the liver were measured for each MR image to calculate the means of three different regions (right anterior, right posterior, and left lateral segment) of the liver devoid of large vessels or prominent artifacts (Fig. 1a). We calculated MR enhancement with four different methods: (i) relative enhancement (RE) was calculated using the following equation: (post-enhanced SI—pre-enhanced SI)/pre-enhanced SI; (ii) using the liver-to-muscle (paraspinal muscles) SI (LMSI), liver SI/muscle SI, we calculated the post-enhanced LMSI/pre-enhanced LMSI (described as LM); (iii) using the liver-to-intervertebral disk SI (LISI), liver SI/intervertebral disk SI, we calculated the post-enhanced LISI/pre-enhanced LISI (described as LI); (iv) using the liver-to-spleen SI (LSSI), liver SI/spleen SI, we calculated the post-enhanced LSSI/pre-enhanced LSSI (described as LS). We compared several predictive methods of fibrosis published previously (Forn's score, FibroIndex, APRI, Lok index). Blood tests were performed in the hospital laboratories of the participating clinical centers. Age, sex, AST, ALT, albumin, total bilirubin, and platelet counts were examined, and the PT was measured as a percentage of the daily internal control. The diagnostic value of the MRI was assessed by calculating the areas under the receiver operating characteristic (ROC) curves. An area under the curve of 1.0 is characteristic of an ideal test, whereas 0.5 indicates a test of no diagnostic value. The best models derived from the categorical variables were compared by the χ2 or Fisher's exact test, whereas continuous variables were compared with Student's t test. Correlation was evaluated by the Spearman correlation coefficient. A two-sided P value of less than 0.05 was considered significant. Diagnostic accuracy was calculated by sensitivity, specificity, and positive and negative predictive values, considering significant fibrosis as the disease. The best cut-off points were selected from the ROC curve to identify the presence and absence of significant fibrosis. A representative cut-off point was determined by calculating the maximum sensitivity and specificity, and comparing their accuracy.
(a) One- to two-squared centimeter regions of interest of the mean signal intensity value of the liver were measured for each magnetic resonance image (MRI) to calculate the means of three different regions of the liver devoid of large vessels or prominent artifacts. (b) Time-course of relative enhancement (RE) after gadolinium-ethoxybenzyl (EOB)-diethylenetriamine pentaacetic acid injection. After EOB injection, relative enhancement increased gradually up to 25 min. RE of F0 and F1 was significantly higher than that of F2 at 15, 20, and 25 min. F0 (n = 15); F1 (n = 24); F2 (n = 19); F3 (n = 46); F4 (n = 44).
J Gastroenterol Hepatol. 2013;28(6):1032-1039. © 2013 Blackwell Publishing