The Role Of Diffusion Weighted/Tensor And CT Perfusion Imaging To Evaluate Partial Hepatic Ischemia Reperfusion In Rabbit Models

BackgroundHepatic ischemia-reperfusion injury (I/R) is a problem in clinical medicine, which is a critical factor that affects organ and patient survival, especially after liver transplantation, hypovolemia, cardiac shock or massive hepatic tumor resection, and it is also responsible for acute liver failure and is usually associated with high morbidity and mortality. In short, hepatic I/R is a topic issue in the field of hepatic surgery research. Accumulating evidence suggests that energy metabolism reduction, calcium overload, microcirculation perfusion failure, cytokines disorders, neutrophils aggregation and infiltration, the effect of oxygen free radicals subsequent to I/R is a major factor in the pathogenesis of I/R-mediated liver injury. Although serum levels of aspartate and alanine transaminase (AST and ALT) activities, and the levels of tumor necrosis factor-a (TNF-a) have been used as indexes for estimating the degree of liver functional loss, they do not give morphologic information on the spatial distribution of liver injury. Moreover, laboratory parameters (liver enzymes activities and inflammatory factors levels), although valuable, could also be biased by postoperative bleeding and transfusion therapy with fresh-frozen plasma and blood, which may be lead to a pseudomorph and risk of a (resurgery) or (re-)transplantation.Diffusion weighted/tensor imaging (DWI/DTI) has become a function and micro-imaging MRI modality that allows visual assessment of the Brownian motion of water molecules and in vivo measurement of molecular diffusion and has become an important tool in life science research and clinical practice. At present, DWI/DTI has been extensively evaluated in the central nervous system. However, DWI/DTI is not a routine approach for the characterization of liver diseases.The pathologic mechanism of ischemic tissue injury includes ischemia and induced hypoxia initially. A cascade of deterioration is further aggravated by secondary reperfusion damage resulting from intra- and extracellular edema, through membrane disintegration, microcirculation perfusion disturbances, and eventually irreversible cytolytic necrosis. Therefore, such continuing changes can be detected by DWI/DTI as more restricted water diffusion initially on account of the diffuse cellular edema and subsequent enhanced water diffusion due to cytolysis. Yet, researches related to hepatic I/R have not reached to agreement and still in the initial phase.Multi-slice computed tomography perfusion imaging (CTP) as an effective, non-invasive technique that allows for quantitative and objective quantitative assessment of angiogenesis, blood perfusion, and vascular permeability with high temporal and spatial resolution. More evidence suggests that microcirculatory disturbances and capillary perfusion failure subsequent to I/R is a major factor in the pathogenesis of I/R-mediated liver injury, which increases in the heterogeneity of hepatic microvascular perfusion, thus, suggesting that CT together with results of laboratory tests, perfusion parameters may be well suited for the in vivo assessment of hepatic I/R.Therefore, the subjective of this study was to establish partial liver I/R in New Zealand white rabbit models and observe the changes of histopathology, liver function loss and level of TNF-a, and to the determine whether DWI/DTI/CTP could be used to characterize hepatic I/R quantitatively, providing a valuable and practicable method for diagnosis and evaluation of hepatic I/R. Part I Establishment of Partial Hepatic Ischemia Reperfusion in a Rabbit ModelObjectiveThis study was to establish partial liver I/R in New Zealand white rabbit models and observe the characteristics of gross morphology, histopathology, changes of liver function and level of TNF-a, and to provide pathological basis for functional imaging studies of hepatic I/R.Materials and Methods1. AnimalsThe experiments were performed with 3～4 month-old adult and male New Zeland rabbits, weighing 2.5 to 3.0 kg. All the experimental rabbits underwent computer tomography angiography (CTA) preoperatively to analyze anatomy variation of hepatic artery and portal vein.2. Surgical Procedures and Experimental GroupsSeven animal groups were used. The left portal vein and hepatic artery were occluded for 60 minutes with a vascular clamp to produce ischemia. The right hepatic lobe was perfused to prevent intestinal congestion. (Thereafter, the clamp was removed to allow blood flow through the liver (reperfusion) for 0.5,2,6,12,24 and 48 h, correspondingly. Animals undergoing laparotomy in parallel without vascular occlusion served as sham-operated controls (sham group).3. Biochemistry assays and histological studies3.1 Serum Assays and level of TNF-aBlood samples were collected from the supeirior mesenteric vein with heparinized capillary tubes after MRI and CT were completed at every time point. The serum was separated from the whole blood by centrifugation to determine the ALT, AST and ALP activities in the serum, and the results were expressed as international units per liter.Using ELISA kit, serum levels of TNF-a in I/R and sham groups were detected by 3 times, and the mean value were expressed as pg per milliliter. 3.2 Histological studiesTo perform biopsies, animals were sacrificed after drawing blood from the superior mesenteric vein, after that, to observe the characteristic of gross anatomy. Then, the liver was rapidly excised, fixed (formalin 10% neutral-buffered), embedded (paraffin), sectioned, and stained with H&E. Histologic analyses of tissue slices were performed randomly. The integration of liver injury using semi-quantitative method according to the degree of various diseases multiplied by different weights:liver cell necrosis×(0～3), cells swelling×(0～3), inflammatory cell infiltration×(0～3), congestion×(0～3). To calculate each animal's total scores, and then were analyzed by statistics.4. Statistical analysisSPSS13.0 statistical software was used for the analysis. (1):One-way analysis of variance (ANOVA) was used to determine differences between the indexes including serum levels of (AST, ALT and ALP), TNF-a and scores of histomorphologic in I/R and sham groups. (2):The correlations of histomorphologic scores with serum levels and TNF-a were analyzed using Pearson's correlation coefficient. For all tests, a P value less than 0.05 was considered to indicate a statistically significant difference.Results1. The result of establishment modelsThe study ultimately included seventy-eight experimental rabbits, which were divided into 6 I/R groups (0.5～24 h groups, n=12 each; 48 h group n=6) and one sham group (n=12 each).2. Histologic findings and biochemistry analysis2.1 Gross anatomyThe affected liver lobe appeared hyperemic or congestive compared to the sham group at the early stage(0.5 and 2 h group), subsequently, the liver's color gradually fades, became kermesinus, having "pattern-like" appearance(12-24 h), and infarction area present with dark green at edge of the liver(48 h). The right lobe without blocking the blood supply showed marked congestion and swelling.2.2 Under microscope The histopathologic changes after hepatic I/R included formation of intracellular edema and, hence, diffuse swelling of hepatocytes, narrowing of the lumen, and sinusoidal congestion. Destroyed erythrocytes were congested in and outside the sinusoidal space and the central vein (CV) in the initial phase (0.5 and 2 h group). With aggravated injury, inflammatory cells aggregated in the hepatic microvasculature and extravasated into the hepatic sinusoid lumen and hepatic parenchyma (6 and 12 h group). Some hepatocytes nuclear were staining and condensation of apoptotic cells in the affected tissues. Eventually, widespread cell necrosis/apoptosis, sinusoidal dissociation, and patchy focal necrosis were observed in the 24 and 48 h groups.2.3 The histomorphologic scoresThe general trend of histomorphologic scores in I/R groups kept on increasing except for a sharp increase of 0.5 h group, and peaked at 48 h after reperfusion. There were significant differences between sham and I/R groups(F=172.702, P<0.001).2.4 Biochemical findingsCompared to those in the sham group, the levels of ALT (F=325.531, P<0.001), AST(F=1009.602, P<0.001) and ALP (F=1430.899, P<0.001) in the I/R group were clearly higher at every time point. The levels of ALT in the I/R groups markedly increased significantly in the 2 h and peaked at 6 h after reperfusion and decline gradually from 12 h to 48 h of reperfusion. AST increased at 2 h of reperfusion, peaking at the end of the perfusion (48 h), and ALP increased rapidly in the 6 h group and peaked at 24～48 h after reperfusion.The levels of TNF-a in the I/R groups was increased significantly in the 2 h group and peaked at 12～24 h of reperfusion, and declined in the 48 h group. The level of TNF-a remained at stable and relatively low level in the sham group. Compared to the sham group, the levels of TNF-a in the I/R group were clearly higher at every time point (F=172.702, P<0.001).3. CorrelationsOur result demonstrated a good relationship between histomorphologic scores and serum levels (ALT, AST, and ALP) and TNF-a in the sham and I/R groups (r=0.788, r=0.841 and r=0.848, respectively, P<0.01).Conclusions1. Our study was to follow up and obverse the process of pathogenesis in a rabbit partial hepatic I/R model with relatively simple and high success rate during a long time window of 48 hours after reperfusion, therefore, may be suitable for functional imaging evaluation of hepatic I/R.2. Two phases of hepatic injury followed by I/R:the initial phase (0-2 h of reperfusion), associated with the generation of toxic hepatocytes swelling, congestion, and a later phase (6-48 h post-reperfusion) associated with a rather intense state of neutrophils aggregation and subsequent to hepatocyte necrosis. One of the basic pathological features was the heterogeneity response to the injury factors after reperfusion.3. These factors inducing in the course of the I/R, such as toxic hepatocytes swelling, congestion, neutrophil aggregation and infiltration, and release of TNF-a, were the major factors in the process of I/R pathogenesis. Morphological scores were positively correlated to the level of liver enzymes (AST, ALT, and ALP) and TNF-a.4. A cascade of deterioration is further aggravated by secondary reperfusion damage resulting from intra-and extracellular edema, neutrophils aggregation, microcirculation perfusion disturbances, sinusoidal dissociation, and subsequent patchy focal necrosis and eventually irreversible cytolytic necrosis, thus, suggesting that function imaging methods may be well appropriate for the in vivo assessment of hepatic I/R.Part II Diffusion weighted/tensor imaging of partial hepatic ischemia reperfusion injury in a rabbit model at 3.0TObjectiveThe intention of this study was to determine whether DWI/DTI could be used to characterize hepatic I/R quantitatively in an experimental rabbit model after 48 h of reperfusion and correlate the results with histopathology, liver enzyme activities, level of TNF-a and histopathology.Materials and Methods1. Experimental models and Experimental GroupsSeven groups of animals (n=6 each) were used (one sham and six I/R groups). The I/R models divided into 0.5 h,2 h,6 h,12 h,24 h and 48 h groups according the reperfusion. The hepatic I/R procedures were same with the Part One.2. Examination methods and ParametersT2WI, T1WI, DWI, DTI, and gadolinium-enhanced T1WI were performed in a 3.0T MR scanner (Signa Excite) with an eight-channel phased-array head coil. The rabbit was placed in a supine position in the coil with head first. The scanning parameters used for morphologic imaging sequences (T1WI) were TR8.5/TE4.0 ms and spoiled gradient-recalled echo (SPGR) in T1WI, and fast spin echo(FSE)in T2WI, TR 3200/TE 85 ms,20×20 cm FOV,3 NEX,3 mm thickness layer, and 288×224 matrix. The scanning parameters used for DWI included echo-planar imaging (EPI) series using multitude b values of 50,100,200,300,500, and 600 s/mm2, repetition time (TR) 2000 ms, echo time (TE) 49.3 ms,20 cm/20 cm field of view (FOV),4 NEX,4 mm thickness layer, and 128/128 matrix. For DTI, fat-saturated coronal single-shot echo-planar imaging (EPI) sequences were performed in 6 directions with b values of 100,300, and 600 s/mm2, repetition time (TR) 2600 ms, echo time (TE) 49.3 ms,20×20 cm field of view (FOV),4 NEX,4 mm thickness layer, and 128x128 matrix. The gadolinium-enhanced T1WI was acquired after 0.2 mmol/kg body weight gadopentetate dimeglumine (Gd-DTPA) was administered manually through three-way catheters, followed by 8 ml saline manned quickly.3. MR Image AnalysisThe liver parenchyma in DWI/DTI was manually delineated at a GE AW4.3 FuncTool Performance workstation with dedicated software (GE Medical Systems) and co-registered using Automated Image Registration to correct for any misregistration caused by the body motions during DWI/DTI and the gradient eddy current related image distortions. A region of interest (ROI) was defined in the axial DWI/DTI images within 15～18 mm2 in the liver parenchyma. On DWI/DTI-derived ADC/DCavg/FA maps involving multitude b values, the left lobes were delineated by copying and pasting the ROI from the corresponding DWI/DTI images at three consecutive slices and 3 ROI were determined in the affected liver lobe in each slice.4. Biochemical and histological studiesThe same was with the Part One.5. Statistical analysisStatistical analyses were performed with SPSS 13.0 software. Values were expressed as means±SD. For all tests, a P value less than 0.05 was considered to indicate a statistically significant difference. One-way analysis of variance (ANOVA) was used to determine differences. Multiple linear regressions was used to compare multiple correlation coefficient (R) of histomorphologic scores, serum levels and TNF-a with DWI/DTI parameters.Results1. Morphologic MRI findingCompared to the sham group, the slightly high signal of the whole left lobe was observed in the 0.5 h groups (T2WI) as well as the enhancement of the affected lobe decreased significantly. However, the signal of the left lobe was reappearing relatively normal on the T2WI and enhanced T1WI images in the 2 h group. As I/R progressed, the marginal areas of affected liver appeared as uneven spots of higher signal (T2WI). The patch-like or wedge-shaped high signal on T2WI had become more evident, usually at the periphery of the liver, which corresponded to a marked unenhanced necrotic area in the 24 h and 48 h groups.2. ADC changesThe overall changes in ADC decreased significantly at the early I/R phase (0.5 h), and drastically increased in the 2 h group, and then ascended slightly from 6 h to 48 h after reperfusion, except for a transient decrease (24 h).The ADC values in 0.5 h (all P<0.001)and 24 h groups (b=50 s/mm2, P=0.037; b=100s/mm2, P=0.020; b=200 s/mm2, P=0.016; b=300 s/mrn2, P=0.001))were significantly lower than that of the sham group when the b-values were no more than 300 s/mm2. Note that the ADC of the 2 h group rose abruptly in contrast to the sham group, especially when the b-value was lower than 300 s/mm2 at every time point, and the difference did not reach statistical significance compared to sham group. The ADC increased significantly in the 6 h group, but still lower than that of sham group (b=50 s/mm2, P=0.029; b=200 s/mm2, P=0.013; b=300 s/mm2, P=0.021). There were no significant differences between sham and 12h group. The ADC significantly increased in the 48 h group in contrast to the sham group when the b-values were 300,500 and 600 s/mm2 (all P<0.001).3. DCavg valuesThe overall change in DCavg was similar to ADC's:DCavg decreased significantly at the early I/R phase (0.5 h) and was significantly lower than those in the sham group when b values of 100 (P<0.001)and 300 s/mm2(P=0.008). Then drastically increased in the 2 h group and significantly higher than sham group when b=100 s/mm2(P=0.021). The DCavg values of the 6 h group were significantly higher than that of the sham group when the b-values were 100(P=0.021) and 300 s/mm2 (P=0.002). In the 24 h group, the DCavg values were also lower in comparison with the sham group when b=100(P=0.008) and 300 s/mm2 (P=0.042). The differences in DCavg value between 48 h group and sham group were more obviously with the b values increasing(b=300 s/mm2, P=0.012; b=600 s/mm2, P=0.002).4. FA valuesMeanwhile, FA had an almost opposite trend and increased drastically in the 0.5 h group(b=100 s/mm2, P=0.032; b=300 s/mm2, P<0.001), and then declined slightly in the 2 h group(b=100 s/mm2, P<0.001; b=300 s/mm2, P=0.004). After that, FA decreased slightly in the 6 h and 12 h group, but the difference did not reach statistical significance compared to sham group. With injury progressed, the FA values were kept decreasing 24-48 h of reperfusion, and the FA values in the 24 h (b=300 s/mm2, P=0.018; b=600 s/mm2, P=0.006) and 48 h (b=100 s/mm2, P=0.044; b=300 s/mm2,P< 0.001; b=600 s/mm2, P< 0.001) were significantly decreased compared to the sham group.5. Biochemical findings There were significant differences between levels of ALT (F=121.404, P<0.001), AST (F=489.615, P<0.001), ALP (F=835.325, P<0.001) in I/R and sham groups.Compared to those in the sham group, the levels of TNF-a in I/R groups were clearly higher at every time point (F=171.803, P<0.001).6. Histomorphologic scoresThe overall changes of integral scores increased significantly from 2 h to 48 h groups compared with sham group except for a transient sharply rise in the 0.5 h group (F=119.350, P<0.001).7. Multiple linear stepwise regressionsOur data demonstrated a good relationship between ADC values (b=100 s/mm2, R=0.756; b=500/mm2, R=0.782), biochemical levels and pathologic index. The FA values (b=300 s/mm2, R=0.781) were also closely related to histomorphologic scores and hepatic biochemistry.Conclusions1. This study showed that complex and dramatic changes in ADC/DCavg/FA could be dynamically. We have proposed explanations for the complex mechanisms and multifactorial pathophysiologic processes related to the pathophysiology of hepatic I/R and determined the corresponding various and dramatic changes. These parameters markedly increased/decreased after 2 h of reperfusion, which make the overall trend of I/R have a transient recovery phenomenon similar to the trend of histomorphologic scores.2.3.0T DTI/DWI can dynamically monitor the pathologic processes of liver ischemia reperfusion injury and reveal the microvascular disorder with a perfusion-sensitive DTI/DWI at the lower b values (<300 s/mm2), particularly at early stages, and can accurately reflect the injury of the liver tissue, necrosis and other pathological changes in the late phase of I/R due to the significant difference between sham and I/R groups with larger b values (b=500 and 600 s/mm2).3. The trend of DCavg values was similar to ADC's. In the early phase of I/R, FA had an opposite trend in contrast to DCavg, and declined gradually in the late stage, which was accordance with the increasing dissolution of necrotic liver tissue and the disappearance of normal micro-spatial.4. More importantly, there were good correlation between DTI/DWI parameters, and biochemical levels and pathologic index, which suggests that the DTI/DWI be a feasible and useful tool to prevent early postoperative liver failure, and clinical application.5. DTI/DWI can monitor the pathologic processes of liver ischemia reperfusion injury, and reveal the micro vascular disorder and morphology changes by dramatic changes in parameters, together with perfusion parameters, might explore the mechanism of I/R, especially for potential benefits of decision making and clinical management after liver surgery.PartⅢCT Perfusion imaging of Partial Hepatic Ischemia Reperfusion in a Rabbit ModelObjectiveThis experimental study was designed to evaluate changes in perfusion parameters in rabbits with partial hepatic ischemia reperfusion injury (I/R) and correlate the components of hepatic perfusion with histopathology and liver enzyme, and which hepatic perfusion parameters may be used to monitor the microvascular perfusion of hepatic I/R.Materials and Methods1. Experimental models and Experimental GroupsSeven groups of animals (n=6 each) were used (one sham and six I/R groups). The I/R models divided into 0.5 h,2 h,6 h,12 h and 24 h groups according the reperfusion. The hepatic I/R procedure was same with the Part One.2. Examination methods and ParametersBefore CT scanning, rabbits were fixed on a board in a supine position and an abdominal bandage was applied to reduce movement artifact. CT perfusion imaging was performed on a 256-slice detector CT (Philips Brilliance iCT). After location scans were acquired CT perfusion examinations were performed by a JOG mode scan with a fixed 0.625 mm thick slice including the entire liver and spleen, and post injection delay 2 s before the intravenous bolus injection of contrast material (2 ml/kg, Ultravist 370 mgI/ml) and 10 ml saline at the rate of 1.5 ml/s through the auricular vein using a double power injector, and continued for 50～60 s. The CT perfusion imaging scanning parameters were as follows:detector collimation,128x0.625 mm; thickness/ increment,1 mm/-0.5 mm; rotation time,0.4 s; pitch 0.915; 80 mAs,120 kVp; and matrix,512×512; Number of scans 11; scan interval 3 s.3. Image analysisData were collected and transferred into the EBW4.0I Philips workstation. Then perfusion images of HAP, HPP, TLP and HPI were produced. For every rabbit,3 consecutive axial slices were analyzed and 3 ROI (ranged in size from 0.5 to 2.5 cm2) were determined in the affected liver lobe in each slice. Within 2 h after reperfusion, ROI randomly was distributed on left lobe. Three ROI were placed into the infarcted liver tissue and 3 ROI in the non-infarcted area in the 6-24 h groups. The mean values of each 9 ROI for the viable liver tissue and nonenhancement of the necrotic tissue for every I/R model were calculated, respectively, and were expressed as relative mean values. Also, the infarcted and non-infarcted tissues in the 6,12, and 24 h groups were measured and expressed as mean values for each I/R model.4. Biochemical and histologic studiesThe same with the Part One.5. Statistical analysisOne-way analysis of variance (ANOVA) was used to determine differences between the CT perfusion parameters. The paired t-test was used for comparison of the differences in CT perfusion parameters (1):between infarcted and non-infarcted area in the 6,12 and 24 h groups; (2):between sham group and infarcted area in the 6, 12 and 24 h groups; (3):between non-infarcted area in the 6,12,24 h groups and sham group. Multiple linear regressions was used to compare multiple correlation coefficient (R) of histomorphologic scores, serum levels and TNF-a with CT perfusion parameters. For all tests, a P value less than 0.05 was considered to indicate a statistically significant difference. Results1. Characteristics of CT perfusion map and perfusion parameters changesHeterogeneity of CT perfusion patterns appeared in the 6 h group (the relatively normal enhancement of viable liver tissue or a low-enhancement of the necrotic tissue could be discriminated). As aggravated injury, the area of low perfusion was gradually expanding and showed a decreased trend of enhancement from 6 h to 24 h of reperfusion.2 CT perfusion parameters changes2.1 The overall trend of CT perfusion parametersThe overall changes of HAP, HPP and TLP decreased significantly at the early I/R phase (0.5 h group) except for HPI. HAP drastically increased in the 2 h of reperfusion, and then declined slightly from 6-24 h after reperfusion. HPP were always significant lower than that of sham group, especially in the 0.5 h and 2 h groups.TLP had the similar changes with HAP; However, HPI of I/R groups were always higher than that of sham group.2.2 Changes of CT perfusion parameters in the low-enhancement of necrotic tissues and enhancement of viable liver tissuesCT perfusion parameters in the low-enhancement tissues increased significantly in the 6 h group and decreased gradually in 12, and 24 h groups, however, the HAP in enhancement of viable liver tissues was decreasing gradually. The HPP decreased significantly not only in the low-enhancement and but also in the enhancement of viable liver tissues. The TLP had a similar change with HPP. In a word, the sharply rise in the HPI implied action of "arterial buffer response"3. CorrelationThe results of multiple linear regressions showed CT perfusion parameters in the necrosis tissues were closely related to histomorphologic scores,serum levels, and TNF-a. Among them, perfusion parameters HPP had the better relationship between them(R=0.855).Conclusions1. The enhancement patterns and the CT perfusion parameters showed a temporary recovery in the 2 h group, which reflect the "false normalization"during the process of hepatic I/R.2. Heterogeneity of CT perfusion patterns appeared in the 6 h group. As aggravated injury, the area of low perfusion was gradually expanding and showed a decreased trend of enhancement from 12 h to 24 h of reperfusion. Eventually, patchy focal necrosis was observed in the 24 h groups.3. Heterogeneity of CT perfusion parameters occurred after reperfusion. Although the HPP declined in the low-enhancement tissue in the 2 h groups, the HAP increased significantly so as to the TLP remained relatively normal and HAP of infarcted tissues were higher in 6～24 h groups than that of non-infarcted tissues, thus, that may prove the "arterial buffer response".4. In spite of HAP increased significantly in the 6-24 h groups, HPP and TLP were decreased significantly in the low-enhancement tissue which make the "arterial buffer response" out of action result in the reduction of HPP mainly because of thoroughfare channel. The perfusion in the low-enhancement tissue was actually reducing that further aggravated the liver tissues injury(24 h group).5. Serum levels, TNF-a and morphological scores were closely correlated to the severity of liver tissue damage as determined by the CT perfusion parameters of the low-enhancement liver tissue. Our result suggested CT perfusion parameters may be a useful parameter in assessing pathologic processes of hepatic I/R.