Hepatocellular carcinoma with central scar on gadoxetic acid-enhanced and diffusion-weighted magnetic resonance imaging.

Background Central scars are rarely reported in conventional hepatocellular carcinoma (HCC). The presence of central scars on imaging might lead to erroneous diagnosis of hepatic tumors. Purpose To determine imaging features of HCC with central scars on magnetic resonance imaging (MRI) including gadoxetic acid-enhanced and diffusion-weighted imaging (DWI). Material and Methods Fifty-one patients with 51 surgically confirmed HCCs with central scars (fibrotic scar: n = 50; myxoid scar: n = 1; range = 1.2-15 cm; mean = 3.7 cm) underwent liver MRI that consisted of T1- and T2-weighted (T2W) imaging, gadoxetic acid-enhanced arterial, portal, 3-min late phase, and 20-min hepatobiliary phase (HBP), and DWI. Two reviewers evaluated morphology, signal intensity, and enhancement features of tumors and central scars for each image and reached consensus. Results Lobulated contour was seen for 30 tumors (58.8%); the rest were round or oval masses. Central scars (range = 0.2-6.0 cm; mean = 0.9 cm) were most commonly seen as defects within hyperenhancement on arterial phase images (n = 47, 92.2%), bright (n = 28, 54.9%) or dark areas (n = 15, 29.4%) on T2W imaging, areas of central darkness on high b-value DWI (b = 800) (n = 31, 60.8%), and/or central enhancement on HBP (n = 36, 70.6%), mimicking a target appearance. Tumor capsule was seen in 35 (39 pathology, 74.5%) and intratumoral septum in 35 (41 pathology, 78.4%) tumors on gadoxetic acid-enhanced MRI. Conclusion Non-fibrolamellar HCC may show central scar. HCC with central scar mimics cholangiocarcinoma by showing a target appearance on HBP and DWI. Tumor capsule and intratumoral septum might be useful for characterizing HCC with central scar.

Comparison between MRI with MR cholangiopancreatography and endoscopic ultrasonography for differentiating malignant from benign mucinous neoplasms of the pancreas.

OBJECTIVE: To compare diagnostic performance of magnetic resonance imaging (MRI) and endoscopic ultrasonography (EUS) for differentiating malignant from benign intraductal papillary mucinous neoplasms (IPMN) and mucinous cystic neoplasms (MCN) of the pancreas. METHODS: This retrospective study included 55 patients with 47 surgically confirmed IPMNs (12 malignant, 35 benign) and eight MCNs (two malignant, six benign) who underwent contrast-enhanced pancreas MRI and EUS. Contrast enhancement was not routinely used at EUS examination. Two observers independently evaluated the MRIs, and another reviewed EUS images. They recorded their confidence for malignancy with each imaging modality. We calculated diagnostic performance using the area under the receiver operating characteristic curves (A z ), and to determine the accuracy, sensitivity, specificity, and positive (PPV) and negative predictive (NPV) values. RESULTS: The A z values of MRI were higher than those of EUS (0.712 and 0.688 for MRI vs. 0.543 for EUS; p = 0.007). The diagnostic accuracies (74.5%), specificity (78.0% and 80.5%) and PPV (50.0%) of MRI in two observers were higher than those (56.4%, 58.5% and 29.2%, respectively) of EUS (p = 0.013-0.049). CONCLUSION: MRI showed better diagnostic performance than EUS for differentiating malignant from benign pancreatic IPMN and MCN. KEY POINTS: • The A z values of MRI were higher than those of EUS. • The diagnostic accuracies of MRI were higher than those of EUS. • The specificities of MRI were higher than those of EUS.

Capsule, septum, and T2 hyperintense foci for differentiation between large hepatocellular carcinoma (≥5 cm) and intrahepatic cholangiocarcinoma on gadoxetic acid MRI.

OBJECTIVE: To determine the added value of capsule, septum, and T2 hyperintense foci for differentiating large hepatocellular carcinoma (HCC; ≥ 5 cm) from intrahepatic cholangiocarcinoma (ICC) using gadoxetic acid MRI. METHODS: The study included 116 patients (94 men, 22 women; mean age, 56.8 years) with surgically confirmed HCCs (n = 87, 5.0-18.0 cm) or ICCs (n = 29, 5.0-14.0 cm) who underwent gadoxetic acid MRI. Three observers independently reviewed MRIs in two sessions, examining enhancement patterns only and then adding capsule, septum, and T2 hyperintense foci. Reviewers used a five-point scale to score accuracy, sensitivity, and specificity. RESULTS: A significant increase was observed in accuracy when ancillary features (96.1-98.3%) were added compared to enhancement pattern only (83.6-88.4%; p ≤ 0.02). Sensitivity was significantly increased with combined reading (97.1-98.3%) compared to enhancement features only (81.6-88.5%; p ≤ 0.006) for two observers, with no difference in specificity (84.5-89.7% vs. 86.2-98.3%; p > 0.05). We found substantial to excellent interobserver agreement for ancillary features (0.598-0.976). CONCLUSION: Adding capsule, septum, and T2 hyperintense foci to enhancement patterns for gadoxetic acid MRI increased diagnostic performance for characterizing large HCC by differentiating it from ICC. KEY POINTS: • Capsule, septum, and T2 hyperintense foci were useful for characterizing large HCC. • Adding ancillary features to enhancement pattern increased accuracy for diagnosing large HCC. • Interobserver agreement was substantial to excellent for ancillary features.

Predicting Hepatic Steatosis in Living Liver Donors via Noninvasive Methods.

Hepatic steatosis assessment is of paramount importance for living liver donor selection because significant hepatic steatosis can affect the postoperative outcome of recipients and the safety of the donor. The validity of various noninvasive imaging methods to assess hepatic steatosis remains controversial. The purpose of our study is to investigate the association between noninvasive imaging methods and pathology to detect steatosis in living liver donors and to propose a prediction model for hepatic steatosis. Liver stiffness measurements (LSMs) and controlled attenuation parameter values in vibration controlled transient elastography, ultrasonography, computed tomography (CT), and magnetic resonance imaging were used as pretransplant screening methods to evaluate living liver donors between 2012 and 2014. Only 1 pathologist assessed tissue sample for hepatic steatosis. The median age of the 79 living donors (53 men and 26 women) was 32 years (16-68 years). The CT liver-spleen attenuation (L-S) difference and the controlled attenuation parameter values were well correlated with the level of hepatic steatosis on liver pathology. Multivariate analysis showed that liver stiffness measurement (LSM) (ß = 0.903; 95% CI, 0.105-1.702; P = 0.027) and the CT L to S attenuation difference (ß = -3.322; 95% CI, -0.502 to -0.142; P = 0.001) were closely associated with hepatic steatosis. We generated the following equation to predict total hepatic steatosis: Hepatic steatosis = 0.903 × LSM - 0.322 × CT L to S attenuation difference (AUC = 86.6% and P = 0.001). The values predicted by the equation correlated well with the presence of hepatic steatosis (r = 0.509 and P < 0.001). The combination of nonenhanced CT L to S attenuation difference and transient elastography using vibration controlled transient elastography provides sufficient information to predict hepatic steatosis in living liver donor candidates.