Experimental Study Of Blood Biochemiology, MR Diffusion Weighted Imaging And Proton MR Spectroscopy In Liver Fibrosis

Experimental Study of Blood Biochemiology, MR Diffusion Weighted Imaging and Proton MR Spectroscopy in Liver FibrosisPrefaceLiver fibrosis is a pathologic state with abnormal synthesis and deposition of extracellular matrix. It is a procedure of cirrhosis resulted from all kinds of chronic liver diseases. Make early diagnosis and assess the severity of it accurately is very important for therapy and prognosis.Recently, several MR functional imaging techniques emerged. Among them, MR diffusion-weighted imaging sensitizes the MR signal related to the incoherent microscopic motion of water molecules and ¹H magnetic resonance spectroscopy prvides direct biochemical information on hepatic metabolic processes in liver. Both of them can assess ultrastructural changes and alterations of metabolism and function before morphological alteration in liver fibrosis.In our experimental study, blood biochemiology, DWI and ¹H-MRS examinations were underwent in rabbit model of liver fibrosis. The changes of them in different stage of liver fibrosis were analyzed. DWI and ¹H-MRS results were compared with histopathological and ultramicrostructural features to make further interpretations. The diagnosis efficacies of serum markers, DWI and ¹H-MRS parameters for liver fibrosis as well as the correlations among them were analyzed, so as to choose noninvasive method for assessing liver fibrosis reasonablely.Materials and Methods45 rabbits were divided into six groups: 1 as control group with 5 rabbits, 2 to 6 as experimental group with 8 rabbits each. 0.1ml/kg saline were injected in peritoneal cavity once a week in rabbits of control group, while in experimental groups were injected 0.1 ml/kg carbon tetrachloride (CCl₄) once a week. At 8,12,16,20,24 weekend after injection, one rabbit of control group and rabbits of 2 to 6 experimental groups were underwent MR examination respectively.Blood samples were obtained in all animals before MR examination to observe serum markers of liver fibrosis including hyaluronic acid (HA), procollagen typeâ…¢(PCâ…¢), typeâ…£collagen (C-â…£) and laminin (LN). The differences of results at various stages of liver fibrosis were compared. Analysis of relationships among serum markers and stage of liver fibrosis were performed.All MR examinations were performed on a 1.5T unit (GE Signa Twinspeed) using the knee coil. Three different b values were tested: 300, 600 and 1000s/ram². Spin echo echoplanar DWI was performed using the following acquisition parameters: TR/TE=3700/(60.1/68.2/75.5)ms (correspond to three b values), slice thickness/slice spacing=3mm/lmm, matrix=128×128, FOV=16×16, nex=12. Image postprocessing was performed on GE ADW4.0 workstation using software functool. Color maps of ADC and EADC were provided by the software automatically. Three ROI avoiding large vessels and artifacts were placed over the liver on three representative sections. Signal intensity (SI), correspond ADC and EADC value were recorded and compared among different stages of liver fibrosis and among different grades of inflammation activity. The relationships among DWI parameters, liver fibrosis stage and inflammation activity grade were analyzed.For all spectra, a PROVE SV PRESS sequence was used with the following acquisition parameters: TR/TE=1500/35ms, FOV=16×16, nex=8. A VOI of more than 10mm3 was placed in left lobe and right external lobe separately, with care taken to avoid large intrahepatic vessels, perihepatic fat and gas. 2 rabbits of control group and ten of experimental groups underwent spectra acquisition repeatedly when other parameters still stable only TE changed to 144ms. Postprocessing was carded out by using the MR spectroscopic analysis package provided by the manufacturer (SAGE 7.0; GE Medical System). The composition of ¹H-MRS and the differences between MRS with two TE were analyzed. The peak amplitude and peak area of each metabolite were recorded. The relative metabolite-to-lipid ratios were obtained by dividing the peak amplitude and peak area of metabolite by those of lipid. The differences in appearances and parameters of MRS at various stages of liver fibrosis were compared, analysis of relationship among them was also performed. Liver was resected immediately after MR scanning and fixed in 10% formalin. The samples were stained by HE, Masson and reticular fiber staining. The stage of liver fibrosis, grade of inflammation activity and degree of fatty infiltration were decided by light microscopy. 1×1×3mm liver sample was obtained and fixed in 2.5% glutaral for making electron microscopic section in 2 rabbits of control group and of each experimental group. Hepatic ultrastructure was observed by PHILIPS CM10 transmission electron microscope.When S≥1 for one with liver fibrosis was arbitrarily defined, receiver operating characteristic (ROC) curve was used to generate AUC and determine the sensitivity and specificity of serum fibrosis indices, DWI and ¹H-MRS parameters when cut-off point was selected according to max Youden index. The relationships of indices among the three methods were analyzed.ResultsEighteen rabbits of experimental group were eliminated from experiment due to death in the process of breeding. With the prolongation of injection, the severity of hepatic degeneration, necrosis and inflammatory cell infiltration progressed, the amount of collagen and reticular fiber increased. The injury of hepatic cell progressed and the deposition of fibrous tissue increased under electron microscope as the injection time prolonged. Morphology of fat-storing cell changed and hepatic sinusoidal capillarization developed in the process of liver fibrosis.1. Results of serum fibrosis indicesAll serum fibrosis indices increased as the stage of liver fibrosis increased, but the statistically differences was only found in HA (P＜0.05). There was a significantly positive correlation among all serum fibrosis indices and between serum indice and stage of liver fibrosis (P＜0.01 ).2. Results of DWIWhen using b values of 300 and 600s/mm² separately, SI, EADC increased and ADC decreased gradually as the stage of fibrosis progressed (P＜0.05); while when b value was 1000s/mm², only SI increased as the stage of fibrosis progressed (P＜0.05). When b value was 300, 600 and 1000s/mm² separately, SI increased gradually as the grade of inflammation activity increased, while ADC value decreased and EADC value increased fluctuately as the severity of inflammation progressed (P＜0.05). There was a good relationship between SI, ADC, EADC and stage of fibrosis, grade of inflammation activity, especially when b value was 300s/mm².3. Appearances and results of ¹H-MRSThree metabolite peaks were seen as lipid (lip), choline complex (cho) and glycogen complex (glc) on ¹H-MRS of liver. When TE was 35ms, the signal-to-noise ratio was high and baseline was stable in the spectra; while when TE was 144ms, the signal-to-noise ration was low and baseline was unstable. The peak amplitude and peak area of all metabolites declined (P＜0.01) while (cho/lip)Amp and (cho/lip)Are elevated significantly (P＜0.01, P＜0.05) when TE was 144ms. The peak ofcho and glc increased in fibrosis liver except stage S2. The changes in peak amplitude and peak area of cho, glc and lip were different as the stage of fibrosis progressed. Of all parameters, only (cho/lip)Amp, (glc/lip)Amp and (cho/lip)Are were statistically different among various stages of fibrosis (P＜0.05) and had a positive correlation with stage of fibrosis (P＜0.05).4. ROC analysis of serum fibrosis indices, DWI and ¹H-MRS parameters in evaluating of liver fibrosis as well as the relationships among all indexesThe AUC of serum fibrosis indices in ROC curves was in the order of HA＞C-Ⅳ＞LN＞PCâ…¢. Take both sensitivity and specificity into account, HA was the best index for diagnosis. The AUC of DWI parameters in ROC curves was in the order of EADC1＞ADC1＞SI1＞SI2＞EADC2＞ADC2. The diagnositic accuracy of all DWI parameters was slightly decreased as b value increased. The AUC of ¹H-MRS parameters in ROC curves was in the order of (cho/lip)Amp＞(cho/lip)Are＞(glc/lip)Amp＞choAre＞choAmp. Take both sensitivity and specificity into account, (cho/lip)Are and (cho/lip)Amp were the best indexes for diagnosis.There was a correlation between SI, ADC, EADC and HA, PCâ…¢, C-â…£, LN when b value was 300s/mm²; while only a positive correlation was found between SI and HA when b value was 600s/mm² (P＜0.01). The positive correlation was statistically significant between C-â…£and choAmp, choAre (P＜0.01), and was statistically different between LN and choAre, (cho/lip)Are (P＜0.05). There was a negative and positive correlation between ADC, EADC and choAmp, choAre when b value was 300s/mm² (P＜0.05). When b value was 600s/mm², there was a positive correlation only found between EADC and choAre (P＜0.05).Conclusion1. All serum fibrosis indices increased as the stage of liver fibrosis increased and HA was an index to quantity liver fibrosis. There was a significantly positive correlation among all indices and between serum indice and stage of liver fibrosis.2. DWI was not only an effective method to quantify severity of liver fibrosis as well as to detect early cirrhosis and severe fibrosis but also a useful tool to quantify severity of inflammation activity and detect severe inflammation. The differences of DWI parameters among fibosis groups and inflammation groups were based on different pathological foundations.3. The variations of cho, glc and lip were different as the stage of fibrosis progressed which can be explained by corresponding changes in pathology. (cho/lip) Amp, (glc/lip)Amp and (cho/lip)Are were of the ability to quantify fibrosis and differentiate severe fibrosis as well as had positive correlation with stage of fibrosis.4. Serum fibrosis indices, DWI and 1H-MRS parameters were all of certain value in diagnosing liver fibrosis, but none of them had the best sensitivity and specificity. There were some correlations between the indexes among three examinations. These correlations were a reflect of common pathological and physiological basis within liver fibrosis.