Clinical Study Of MR Susceptibility-weighted Imaging On Multistep Hepatocarcinogenesis In Cirrhotic Livers

PART ⅠDetection and Quantification of Iron deposition in Cirrhotic Livers Using MR Susceptibility-weighted Imaging:Comparison with T2*-weighted Imaging and Correlation with HistopathologyPurpose To preospectively compare susceptibility-weighted Imaging (SWI) and T2*-weighted imaging in the detection and quantification of iron deposition in cirrhotic livers using histopathologic result as the reference standard. Materials and methods Fifty-five patients (46men, nine women; age range26-79years) with pathologically confirmed cirrhosis underwent hepatic MR imaging with SWI, single-echo and multi-echo T2*sequences. For qualitative analysis, two sets of images from SWI and single-echo T2*were analyzed in random order for assessment of presence of hepatic iron deposition. Nodules that were hypointense to background liver parenchyma were considered siderotic nodules (SN). For cases with SN detected by SWI and T2*, number of SN as well as minute SN with diameter less than3mm were counted, respectively. The conspicuity of SN was assessed using a scale from1to3(1, weak;2, moderate;3, prominent). For quantitative analysis, SWI phase value and T2*value were measured. Prussian blue staining was performed for semiquantification of hepatic iron content. Correlation between SWI phase value and hepatic iron concentration were determined. SWI phase values were compared between patients stratified by hepatic iron grade and were correlated with histopathologic iron grade. Results Positive iron staining was found in54of55patients (98.2%) with grade1(n=15), grade2(n=16), grade3(n=18) and grade4(n=5). A significant improvement in the sensitivity for detecting iron deposition in cirrhotic livers was achieved by SWI compared with T2*(66.7%vs90.7%, p=0.002)., especially for grade1(26.7%vs73.3%, p=0.027) and grade2(56.3%vs93.8%, p=0.037). The number of SN and minute SN detected by SWI were significantly greater than that of T2*(p=0.000and0.000, respectively). Overall conspicuity of SN on SWI images was better than that on T2*images. Patients with iron grade3or greater had lower SWI phase values than those with iron grade2or less (p=0.000). There was a strong negative correlation between SWI phase values and histopathologic iron grade (r=r=-0.803) and a significant correlation between SWI and T2*(r=0.771). Conclusion Detection and quantification of iron deposition in cirrhotic livers could be better performed by SWI compared with T2*. PART ⅡSusceptibility-weighted Imaging of Multistep Hepatocarcinogenesis in Cirrhotic Livers:Correlation with PathologyPurpose To observe imaging characteristics of multistep hepatocarcinogenesis in cirrhotic livers on susceptibility-weighted imaging and correlate with histopathologic results. Materials and Methods Seventy-one patients with83nodules in cirrhotic livers underwent hepatic MR imaging with SWI, unenhanced (T1-and T2-weighted imaging) and gadolinium-enhanced dynamic MRI. Two radiologists reviewed MR images by consensus. Imaging characteristics of dysplastic nodules (DN), DN with malignant foci and HCC were evaluated. Prussian blue staining was performed for semiquantification of hepatic iron content and cirrhosis-associated nodules. Results Positive iron staining of background liver parenchyma was found in68of71patients (95.8%) and3HCC patients were iron-negative staining of background liver parenchyma. Nine DNs appeared as hypointensity or isointensity with pathologically confirmed similar (n=7) or slighly decreased (n=2) iron deposition compared with background liver parenchyma. SWI detected14of15DNs with malignant foci. Seven cases appeared as homogeneous hyperintensity and1case appeared as heterogeneous hyperintensity due to intratumoral hemorrhages. The remaining6cases demonstrated as nodule-in-nodule appearance with iron deposition in all background nodules, iron deposition with grade1in one internal HCC foci, and iron-free in5internal HCC foci. The remaining50patients with hepatic iron deposition had55HCC lesions. Three HCC lesions had iron deposition with grade1to2and the remaining52HCC lesions were pathologically iron-resistant. HCC appeared as hyperintensity compared with siderotic surrounding liver parenchyma. However, HCCs with diameter larger than3cm usully demonstrated heterogeneous hyperintensity due to intratumoral hemorrhage. Conclusion SWI could accurately visualize dynamic iron clearance on multistep hepatocarcinogenesis in cirrhotic livers with siderosis and iron-free hyperintense nodules on SWI images should be considered the possibility of HCC. PART IIIAdded Value of Susceptibility-weighted Imaging to Unenhanced and Gadolinium-enhanced Dynamic MR imaging for Characterization of Small Hepatocellular Carcinoma in Cirrhotic LiversPurpose To assess the added value of MR susceptibility-weighted imaging (SWI) to unenhanced and gadolinium-enhanced dynamic MRI for characterization of small hepatocellular carcinomas (sHCCs) in cirrhotic livers. Materials and Methods Sixty-seven cirrhotic patients with105nodules with diameter less than3cm in cirrhotic livers underwnet hepatic MR imaging with SWI, unenhanced (T1-and T2-weighted imaging) and gadolinium-enhanced dynamic MRI. Two abdominal radiologists independently evaluated two image sets (conventional MRI set, unenhanced and gadolinium-enhanced dynamic images; SWI set, SWI, unenhanced and gadolinium-enhanced dynamic MRI) and assigned confidence levels for diagnosis of HCC using a five-point scale for each lesion. Interreader varaibility was assessed and sensitivity, positive predicitive value (PPV), and area under the alternative-free response receiver operating characteristic curve (Az) was calculated for each image set. Results The Az value of SWI set was larger than the conventional MRI set by both readers (reader1,0.804vs0.898, p<0.05; reader2,0.782vs0.859, p<0.05). Sensitivity with the combined susceptibility-weighted imaging and gadolinium set (93.5%) was significantly higher than those with the gadolinium set (80.6%)(P<0.05). The addition of susceptibility-weighted images lead to a change in diagnosis for12lesions by both observers, which at set1were assigned a confidence level of1or2but at additional reading of susceptibility-weighted images were assigned a confidence level of3or4. For the positive predictive values, each image showed a similar value for each observer. For confidence levels for diagnosis of HCC with diameter smaller than2cm (reader1,3.86±0.47vs3.86±0.47, p=0.005; reader2,3.41±0.92vs3.85±0.57, p=0.014)。 Conclusion SWI could provide additional valuable inforamtion to unenhanced and gadolinium-enhanced dynamic imaging for characterization of sHCC in cirrhotic livers. PART ⅣEvaluating Tumor Angiogenesis of Advanced Hepatocellular Carcinoma Using Susceptibility-weighted Imaging:A Preliminary StudyPurpose To investige the value of susceptibility-weighted imaging (SWI) for evaluating tumor angiogenesis of advanced hepatocellular carcinoma (HCC). Materials and methods Thirtyty-four patients (32men,2women; age range28-76years) with pathologically confirmed HCC larger than3cm in diameter underwnet hepatic MR imaging with SWI, unenhanced and gadolinium-enhanced dynamic MRI. SWI phase value was measured and correlated with microvessel density (MVD) counted with immunhischemistry. SWI phase values and frequency of intratumoral hemorrhage of HCC with histopathological grade were compared. Results A moderate positive correlation was found beween the SWI phase values and MVD (r=0.486, p=0.004). There was no significant difference between SWI phase values, frequency of intratumoral hemorrhage of HCC of Endonson-Steiner1and2compared with HCC of grade3(P=0.795). Conclusion SWI demonstrated limited value for noninvasive assessment of angiogenesis of advanced HCC due to the possiblity of intratumoral high oxygen level.