This study aim to combine multimodal MRI with liver volume for quantitative assessment of liver function in patients with cirrhosis. Eighty-five subjects underwent GD-EOB-DTPA enhanced MR and IVIM-DWI.Through statistical analysis, RE, D* and Vliver/Vspleen from modeling group were selected as optimal indicators. Using multiple regression to establish a multi-parametric model for liver function assessing: F(x). The ROC analysis of the validation group showed that it had an AUC of 0.973 in distinguishing Child-Pugh A group and Child-Pugh B and C group. The results showed it could evaluate the overall liver function and expected to be a method for preoperative local liver function assessment.
INTRODUCTION
Current clinical assessment of liver function is based on systemic laboratory examinations and lacks a method for accurately assessing the function of the whole liver, liver lobe or liver segment. Gd-EOB-DTPA(Primovist, Bayer Schering Pharma, Germany) can be specifically taken up by normal liver cells. When liver function is impaired, the amount of Gd-EOB-DTPA absorbed by liver cells decreased. The resulting change in signal intensity can be used to assess liver function1. Intravoxel Incoherent Motion-Diffusion Weighted Imaging(IVIM-DWI) reflects both diffusion and perfusion of liver, which has certain potential for evaluating liver function2. Liver volume is another important clinical indicator of liver function3,4. This study aims to establish a non-invasive assessment model of liver function by combining Gd-EOB-DTPA enhanced MR, IVIM-DWI and liver volume to provide a reference for preoperative assessment of liver function.RESULTS
Among all the parameters from GD-EOB-DTPA enhanced MR, RE displayed significant statistical differences between different liver function groups(P<0.05). Among the three parameters derived from IVIM-DWI, D* displayed significant statistical differences between different liver function groups(P<0.05). Among the three parameters from liver volume, Vliver/Vspleen displayed significant statistical differences between different liver function groups(P<0.05). ROC analysis showed that these three parameters displayed the highest value in distinguishing Child-Pugh A group and Child-Pugh B and C in the respective evaluation methods.The area under the curve(AUC) were 0.917, 0.929, and 0.885, respectively(Figure 4). RE, D*, and Vliver/Vspleen were selected as effective parameters for model establishment. The model was acquired by multiple regression: F(x)=3.96-1.243×RE-0.034×D*-0.080×Vliver/Vspleen(R2=0.811, P<0.001). The ROC analysis of the validation group showed that the AUC of F(x) in distinguishing Child-Pugh A group and Child-Pugh B and C group was 0.973(Figure 5).DISCUSSION
Hepatocytes are damaged during liver cirrhosis, and the expression of organic anion transport polypeptide on hepatocyte membrane is decreased. It affects the absorption of Gd-EOB-DTPA, resulting in a decrease in the degree of liver specific signal enhancement. Therefore, RE could be used as an effective indicator for evaluating liver function5,6. Different degrees of collagen fibrosis during liver cirrhosis, hepatic sinus stenosis, decreased capillary perfusion, which leads to a decrease in liver perfusion. All of these could be reflected by a decrease of D*, so D* was selected as an effective indicator for evaluating liver function7. As liver fibrosis progresses, hepatocyte regeneration is weakened, resulting in a decrease in liver volume8. At the same time, the volume of the spleen increases, so Vliver/Vspleen was selected as an indicator for evaluating liver function4. An ideal liver function assessment model should be a combination of metabolism and morphology. Therefore, this study combined functional changes in hepatocytes, tissue perfusion and liver morphological changes to establish a liver function assessment model. The results of the validation group showed that the model displayed excellent value in assessing overall liver function. In addition, these three parameters can be obtained by placing the ROI in a certain liver lobe or liver segment. Therefore, it is speculated that it could be used to evaluate the function of the liver lobe or liver segment.CONCLUSION
This study established a multi-parameteric model for assessing liver function. It could evaluate the overall liver function and expected to be a method for preoperative local liver function assessment.