Liver And Spleen Dynamic Contrast-enhanced Magnetic Resonance Imaging For Staging Liver Fibrosis: An Experimental Study

Objective: To explore whether and how liver and spleen dynamiccontrast-enhanced magnetic resonance imaging (DCE-MRI) could monitorprogression of liver fibrosis in a piglet model, and which DCE-MRIparameter is most accurate for staging this disease.Methods: sixteen piglets, weighing between20.0kg and24.0kg, wereprospectively used to model liver fibrosis induced with carbon tetrachloride,and underwent abdominal DCE-MRI followed by biopsy on0,5th,9th,16thand21st weekend after modeling of fibrosis. The following Liver and spleenDCE-MRI parameters were measured: Time to peak (TTP), positiveenhancement integral (PEI), maximum slope of increase (MSI) and maximumslope of decrease (MSD) were measured. Liver specimens were scored usingthe human METAVIR classification system after standardd Masson trichromestaining. Statistically analyzed for the monitoring and staging liver fibrosis.Results:(1) As liver fibrosis progress, liver TTP tended to increasewhereas MSI, MSD and PEI trended to decrease (r=0.689、-0.820、-0.343and-0.492， all P <0.05). Although PEI at liver fibrosis stage F4was higherthan at stage F3, there was a slight increase in MSD from liver fibrosis stageF1to F3, however, this increase had no statistical significance (P>0.603).Liver TTP and MSI could discriminate fibrosis stage F0from F1-4, F0-1from F2-4, F0-2from F3-4, and F0-3from F4; Liver PEI coulddistinguish above-mentioned stages except F0-3from4; and MSD could distinguish stage F0-3from4, and F0from1-4(all P<0.05). For predictingstage F≥1(cutoff,11.591/s),≥2(cutoff,9.651/s) and≥3(cutoff,8.931/s),area under receiver operating characteristic curve (AUC) of MSI was largestamong all parameters(AUC=0.920、0.913、0.917, respectively); and for stage4, the best parameters could be used to predict liver fibrosis S4was liver TTPamong all the liver DCE-MRI parameters (AUC=0.963).(2) As liver fibrosis progress, Spearman’s rank correlation tests showedthat spleen TTP tended to increase with increasing stages of liver fibrosis (r=0.647, P <0.001), and that PEI tended to decrease from stage F0to4(r=-0.709, P <0.001). MSD increased slightly from stage F0to2(P>0.05),and decreased from stage F2to4(P <0.05). MSI increased from stage F0to1, and decreased from stage1to4(all P>0.05). Spleen TTP and PEI couldclassify fibrosis between stage F0and1-4, between F0-1and2-4, betweenF0-2and3-4, or between F0-3and4(all P <0.01). Spleen MSD coulddiscriminate between F0-2and3-4(P=0.006), or between F0-3and4(P=0.012). Spleen MSI could not differentiate between any two stages. Receiveroperating characteristic analysis illustrated that area under receiver operatingcharacteristic curve (AUC) of spleen TTP was larger than of PEI forclassifying stage F≥1(cutoff,70.82s) and≥2(cutoff,73.75s)(AUC=0.851and0.783, respectively). Spleen PEI could best classify stage F≥3(cutoff,5.46) and F4(cutoff,4.75) with the best AUC of0.903and0.968,respectively.Conclusion: Liver and spleen DCE-MRI has potential to dynamicallystage progression of liver fibrosis. Liver MSI was most useful for predictingfibrosis stage≥1,≥2and≥3and spleen PEI was best to predicate liver fibrosisstage F4.