Arterioportal shunts in cirrhotic patients: evaluation of the difference between tumorous and nontumorous arterioportal shunts on MR imaging with superparamagnetic iron oxide

OBJECTIVE: The study objective was to distinguish between the features of tumorous and nontumorous arterioportal shunts on superparamagnetic iron oxide-enhanced MR imaging in patients with cirrhosis. SUBJECTS AND METHODS: Ten arterioportal shunts in eight patients, including four tumorous and six nontumorous arterioportal shunts, were evaluated on T2-weighted turbo spin-echo and T2(*)-weighted gradient-echo sequences before and after administration of superparamagnetic iron oxide. Qualitatively, the relative signal intensity of the arterioportal shunt compared with that of the surrounding liver parenchyma was categorized into three grades: high, slightly high, and not detected. Quantitatively, signal-to-noise ratio, contrast-to-noise ratio, lesion-to-liver contrast, and percentage enhancement were calculated and compared between tumorous and nontumorous arterioportal shunts by a nonparametric statistical test (Mann-Whitney test). RESULTS: Qualitatively, all four tumorous arterioportal shunts appeared as areas of slightly high or high intensity without and with superparamagnetic iron oxide on T2-weighted turbo spin-echo images and changed from isointensity to high intensity after the administration of superparamagnetic iron oxide on T2(*)-weighted gradient-echo images. All nontumorous arterioportal shunts except one could not be recognized without or with superparamagnetic iron oxide on either sequence. Quantitatively, with superparamagnetic iron oxide the contrast-to-noise ratio and the lesion-to-liver contrast of the tumorous arterioportal shunts were significantly higher than those of the nontumorous arterioportal shunts. CONCLUSION: Tumorous arterioportal shunts are seen as areas of reduced signal loss, whereas most nontumorous arterioportal shunts are seen as areas of normal signal loss, like the normal liver parenchyma. The difference is more marked on T2(*)-weighted gradient-echo images than on T2-weighted turbo spin-echo images.     

MSCT对肝硬化中非肿瘤性APS的诊断探讨

目的 探讨MDCT对肝硬化中非肿瘤性肝动脉门静脉分流(APS)的诊断价值. 资料与方法 对照数字减影血管造影术(DSA)结果 分析23例肝硬化患者的MDCT增强扫描及多平面重组(MPR)图像的表现. 结果 23例患者中发现28个APS灶,在动脉期均显示为高密度,门脉期为等密度,24个位于肝包膜下.在CT横断面上13个病灶表现为楔形或不规则形,另外15个虽表现为小结节样病灶但在MPR图像上13个也表现为楔形或不规则形.26个病灶DSA表现为门脉小分支早期显影,另2个DSA表现未见异常. 结论 肝硬化患者的MDCT增强扫描动脉期显示肝脏周边较小的楔性、结节样或不规则形高密度灶而在门脉期呈等密度提示非肿瘤性APS.     

Pyogenic hepatic abscesses: MRI findings on T1- and T2-weighted and serial gadolinium-enhanced gradient-echo images

The purpose of this study was to determine the magnetic resonance imaging (MRI) features of pyogenic hepatic abscesses on T1-weighted, T2-weighted, and serial gadolinium (Gd)-enhanced T1-weighted spoiled gradient-echo (SGE) images including images acquired in the immediate, intermediate, and late phases of enhancement. The MRI studies of 20 patients with pyogenic liver abscesses were retrospectively reviewed. All patients were examined on 1.5 (n = 19) and 1.0 (n = 1) T MR scanners. MR studies included T1-weighted, T2-weighted, and serial Gd-enhanced SGE images. The following determinations were made: signal intensity of the abscess cavity and perilesional liver tissue, and the presence of internal septations, layering material, or air in the abscess cavity. The pattern of enhancement of the abscess wall, internal septae and peri-abscess liver were evaluated on serial Gd-enhanced SGE images. A total of 53 abscesses were observed in the 20 patients. Fortyeight abscesses were hypointense on T1-weighted and hyperintense on T2-weighted images. Internal septations were present in four abscesses. Lower signal intensity material was observed in a dependent location on T2-weighted images in one abscess. Signal void foci of air located on the nondependent surface was observed in two abscesses. Two other abscesses contained signal void air that occupied the entire abscess cavity, observed on all imaging sequences. On serial gadolinium-enhanced images, all abscesses revealed early enhancement of the wall, which persisted with negligible change in degree of enhancement or thickness on delayed images. Abscess walls ranged in thickness from 2 to 5 mm. Internal septations ranged in thickness from 2 to 3 mm. Abscess walls and septations were relatively uniform in thickness with no evidence of focal nodularity. Periabscess liver tissue was mildly hypointense on T1-weighted and mildly hyperintense on T2-weighted images in 20 lesions, which were either circumferential (n = 12) or wedge-shaped (n = 8). All these regions enhanced more than the remainder of the liver on immediate post-gadolinium images and remained relatively hyperintense on late phase images. Periabscess liver parenchyma was isointense on both T1- and T2-weighted images in 18 lesions, and in these lesions wedge-shaped subsegmental (n = 6) or segmental (n = 12) enhancement was observed on immediate gadolinium-enhanced images, which faded to isointensity on intermediate phase images. No perilesional signal changes and enhancement difference was observed in 15 lesions. Characteristic features of abscesses include: intense mural enhancement on early gadolinium-enhanced images, which persists with negligible change in thickness and intensity on later post-gadolinium images, and the presence of periabscess increased enhancement on immediate post-gadolinium images. These MRI features may help to distinguish abscesses from other focal liver lesions during differential diagnosis.     

Focal manifestations of diffuse liver disease at MR imaging.

Detection and exclusion of focal liver lesions is especially difficult in patients with diffuse liver disease. Magnetic resonance (MR) imaging may be particularly valuable in these patients. By judicious comparison of appropriate pulse sequences, normal and hypertrophic liver may be distinguished from atrophic, neoplastic, or otherwise abnormal hepatic parenchyma. Chemical shift (lipid-sensitive) techniques allow definitive identification of fatty liver, including focal fatty infiltration or focal sparing. T2-weighted and T2*-weighted images allow identification of iron overload, depicting malignancies as focal masses without iron. Analysis of signal intensity and internal morphology allows confident distinction between regenerative nodules and hepatocellular carcinoma in most instances, and allows diagnosis of early carcinoma within regenerative nodules. MR imaging provides capabilities for noninvasive characterization of liver tissue beyond those available with other noninvasive modalities.     

Multicentre dose-ranging study on the efficacy of USPIO ferumoxtran-10 for liver MR imaging

AIM: A dose ranging multicentre phase-II clinical trial was conducted to evaluate the efficacy of ultrasmall superparamagnetic iron oxide (USPIO) ferumoxtran-10 for magnetic resonance (MR) imaging of focal hepatic lesions. MATERIAL AND METHODS: Ninety-nine patients with focal liver lesions received USPIO at a dose of 0.8 (n = 35), 1.1 (n = 32), or 1.7 (n = 32) mg Fe/kg. Liver MR imaging was performed before and after USPIO with T1-weighted and T2-weighted pulse sequences. Images were analysed by two independent readers for additional information (lesion detection, exclusion, characterization and patient management). Signal intensity (SI) based quantitative measurements were also taken. RESULTS: Post-contrast medium MR imaging showed additional information in 71/97 patients (73%) for reader one and 83/96 patients (86%) for reader two. The results with all three doses were statistically significant (P < 0.05). Signal intensity analysis revealed that all three doses increased liver SI on T1-weighted images and decreased liver SI on T2-weighted images. On T2-weighted images metastases increased in contrast relative to normal hepatic parenchyma whereas haemangiomas decreased in contrast. On T2-weighted images there was statistically improved efficacy at the intermediate dose, which did not improve at the highest dose. CONCLUSION: Ultrasmall superparamagnetic iron oxide was an effective contrast agent for liver MR imaging at all doses and a dose of 1.1 mg Fe/kg was recommended for future clinical trials.     

MR findings in eosinophilic infiltration of the liver

OBJECTIVE: This study describes the findings of magnetic resonance imaging (MRI) of focal eosinophilic infiltration of the liver. METHODS: Contrast-enhanced MR images of 8 patients with focal hepatic eosinophilic infiltration were reviewed retrospectively. We evaluated the signal intensity of focal lesions in T1-weighted and T2-weighted images and the pattern of enhancement in a dynamic contrast study. RESULTS: A total 22 focal hepatic lesions were observed; the lesions were isointense (55%) or hypointense (45%) on T1-weighted images and isointense (14%) or hyperintense (86%) on T2-weighted images. The arterial phase of the contrast study revealed 11 hyperintense lesions (50%). During the portal and delayed phases, 18 (82%) and 17 lesions (77%) were hyperintense, respectively. CONCLUSION: The focal eosinophilic infiltrations showed homogeneous enhancement in the portal and delayed phases in the dynamic contrast MR study. These findings should help to distinguish focal eosinophilic infiltration, especially from metastasis in patients with malignancy.     

Early radiation effects on the liver demonstrated on superparamegnetic iron oxide-enhanced T1-weighted MRI

Early radiation-induced liver injury during radiotherapy detected by a particulate reticuloendothelial MR contrast agent (superparamagnetic iron oxide; SPIO) is described in a patient with cholangiocarcinoma. The irradiated hepatic parenchyma appeared as a heterogeneous, less decreased signal intensity area than the nonirradiated area on MR images after SPIO administration. Resultant differences in signal intensity were better visualized on SPIO-enhanced T1-weighted images than SPIO-enhanced T2-weighted images, although SPIO-enhanced T2*-weighted fast field echo imaging was the most sensitive.     

Focal nodular hyperplasia of the liver: MR findings in 35 proved cases

MR images of 28 patients with 35 lesions of hepatic focal nodular hyperplasia were reviewed to determine the frequency of findings considered typical of this condition (isointensity on T1- and T2-weighted pulse sequences, a central hyperintense scar on T2-weighted images, and homogeneous signal intensity). Fifteen lesions were imaged at 0.6 T with T1- and T2-weighted spin-echo (SE) pulse sequences; 20 lesions were imaged at 1.5 T with T1-weighted SE and gradient-echo pulse sequences and T2-weighted SE pulse sequences. Diagnosis of focal nodular hyperplasia was made pathologically in 25 patients, with nuclear scintigraphy in four, and with follow-up imaging in six. Only seven lesions (20%) were isointense relative to normal liver on both T1- and T2-weighted images. On T1-weighted SE images, 21 lesions (60%) were isointense relative to normal liver, 12 (34%) were hypointense, and two (6%) were hyperintense. On T2-weighted SE images, 12 lesions (34%) were isointense and 23 (66%) were hyperintense relative to normal liver. A central scar was present in 17 lesions (49%) and was hypointense relative to the lesion on T1-weighted images and hyperintense on T2-weighted images. Twenty lesions (57%) were of homogeneous signal intensity throughout the lesion, except for the presence of a central scar. All three MR imaging characteristics were present in three cases (9%). We conclude that hepatic focal nodular hyperplasia has a wide range of signal intensity on MR imaging.     

Nondiffuse fatty change of the liver: discerning pseudotumor on MR images enhanced with ferumoxides-initial observations

PURPOSE: To clarify the findings of nondiffuse fatty change of the liver on ferumoxides-enhanced magnetic resonance (MR) images. MATERIALS AND METHODS: Of 202 patients who underwent ferumoxides-enhanced MR imaging, eight who had nondiffuse fatty change of the liver at computed tomography (CT) were examined as study subjects. MR imaging findings before and 1 hour after ferumoxides administration were compared with CT findings. RESULTS: Focal fatty areas of the liver showing low attenuation on CT images were depicted as areas of relatively high intensity on the ferumoxides-enhanced T1-weighted images in all patients. On enhanced T2-weighted images, focal fatty change showed relatively high intensity in three and isointensity in one of the four patients. Focal spared areas appearing as areas of relatively high attenuation on CT images were depicted as areas of relatively low intensity on the ferumoxides-enhanced T1- and T2-weighted images in all patients. CONCLUSION: Although prior reports of hepatic MR imaging with ferumoxides indicated that there is accumulation of ferumoxides within focal fatty areas that are no longer seen after the administration of contrast medium, this study revealed that focal fatty change and focal spared areas of fatty liver may be pseudotumors because of the relatively high intensity of fatty areas of the liver. Radiologists can distinguish these conditions from hepatic tumors by using the opposed-phase gradient-echo sequence or the fat-saturation technique.     

Segmental intensity differences in the liver on MR images: a sign of intrahepatic portal flow stoppage

Segmental intensity differences (SIDs) in hepatic parenchyma free of tumor were noted in six patients with hepatic masses (hepatocellular carcinoma in five and metastatic liver cancer in one). Areas of SID were homogeneous in intensity. The intensity of the affected region was high in all six patients on T2-weighted magnetic resonance (MR) images and low in two on T1-weighted images. Three of five patients examined with plain computed tomography (CT) had corresponding segmental areas of low attenuation. Angiograms obtained in five patients showed occlusion of the intrahepatic portal vein, segmental staining corresponding to the region of the SID, or both. Twelve of 82 patients examined with MR imaging and angiography had similar findings on angiograms, and ten of them had abnormal intensity of anatomic distribution around or beside the liver tumors on MR images. MR imaging may be more sensitive than plain CT in the detection of secondary changes caused by intrahepatic portal flow stoppage.     

Primary sclerosing cholangitis: MR imaging findings with pathologic correlation.

OBJECTIVE: The purpose of this study was to evaluate MR imaging findings of primary sclerosing cholangitis, to compare them with histopathologic findings, and to determine if these findings help differentiate primary sclerosing cholangitis from other disorders that result in end-stage liver disease. MATERIALS AND METHODS: MR imaging was performed in 40 patients (27 men, 13 women; age range, 13-72 years; mean, 47 years) with primary sclerosing cholangitis over a 9-year period. In 16 patients who underwent orthotopic hepatic transplantation and in seven patients who underwent needle biopsy, correlation was made between MR imaging and pathologic findings. RESULTS: Focal signal changes in the liver parenchyma were seen on T2-weighted images as peripheral wedge-shaped zones of increased signal intensity in 29 patients (72%), as a reticular pattern in 15 patients (38%), and as periportal edema in 16 patients (40%). Lobar atrophy involved the right lobe in three patients (8%) and the left lobe in 11 patients (28%); hypertrophy of the caudate lobe was seen in nine patients (23%). Features of portal hypertension were seen in 14 patients (35%). Histologic assessment showed zones of segmental atrophy and scarring on the periphery of the liver. CONCLUSION: Peripheral wedge-shaped areas of high T2 signal intensity and dilatation of bile ducts are characteristic MR features of primary sclerosing cholangitis. Pathologic correlation suggests that these features may be related to underlying perfusion changes and bile duct inflammation in patients with primary sclerosing cholangitis.     

Characterization of focal hepatic lesions with ferumoxides-enhanced T2-weighted MR imaging

OBJECTIVE: The purpose of this study was to evaluate whether ferumoxides-enhanced MR imaging of focal hepatic lesions provides distinctive signal intensity and lesion-to-liver contrast changes for benign and malignant lesions, helping to further characterize and differentiate these lesions. MATERIALS AND METHODS: Data analysis was performed on 70 patients, with previously identified focal hepatic lesions, who underwent MR imaging of the liver before and after IV administration of ferumoxides (10 micromol Fe/kg). Lesions analyzed with pathologically proven diagnoses included metastases (n = 40), hepatocellular carcinoma (n = 11), cholangiocarcinoma (n = 6), hemangioma (n = 4), focal nodular hyperplasia (n = 6), and hepatocellular adenoma (n = 3). Response variables measured and statistically compared included the percentage of signal-intensity change and lesion-to-liver contrast. RESULTS: Focal nodular hyperplasia showed significant signal intensity loss on ferumoxides-enhanced T2-weighted images (mean, -43%+/-6.7%, p < 0.01). All other lesion groups showed no statistically significant change in signal intensity on ferumoxides-enhanced T2-weighted images, although signal intensity loss was seen in some individual hepatocellular adenomas (mean, -6.6%+/-24.0%) and hepatocellular carcinomas (mean, -3.3%+/-10.3%). All lesions, with the exception of hepatocellular carcinoma, had a marked increase in lesion-to-liver contrast on ferumoxides-enhanced T2-weighted images, which was statistically significant for metastases and hemangioma (p < 0.02). CONCLUSION: Focal nodular hyperplasia shows significant decrease in signal intensity on ferumoxides-enhanced T2-weighted images, which may aid in the differentiation of focal nodular hyperplasia from other focal hepatic lesions. Other lesions, namely, hepatocellular adenoma and carcinoma, can have reticuloendothelial uptake, but usually to a lesser degree than that of focal nodular hyperplasia.     

Role of MR cholangiopancreatography in patients with failed or inadequate ERCP

OBJECTIVE: The purpose of this study was to evaluate the spectrum of MR imaging features of primary sclerosing cholangitis. MATERIALS AND METHODS: A retrospective review was performed of MR imaging findings including MR cholangiography and multiphasic contrast-enhanced dynamic sequences in 22 patients with primary sclerosing cholangitis. MR imaging analysis included abnormalities of intra- and extrahepatic bile ducts, abnormalities of liver parenchyma, changes in liver morphology, and lymphadenopathy. RESULTS: Abnormal findings of bile ducts were seen in all 22 patients; the most common finding was intrahepatic bile duct dilatation (77%), followed by intrahepatic bile duct stenosis (64%), extrahepatic bile duct wall enhancement (67%), extrahepatic bile duct wall thickening (50%), extrahepatic bile duct stenosis (50%), and intrahepatic bile duct beading (36%). Increased enhancement of the liver parenchyma on dynamic arterial-phase images, predominantly in the peripheral areas of the liver, was identified in 56% of patients. Other findings included periportal lymphadenopathy (77%), periportal high signal intensity on T2-weighted images (68%), hypertrophy of the caudate lobe (68%), and abnormal hyperintensity of the liver parenchyma on T1-weighted images (23%). CONCLUSION: On MR imaging, primary sclerosing cholangitis showed several characteristic features, including bile duct abnormalities and increased enhancement of the liver parenchyma. MR cholangiography and contrast-enhanced dynamic MR techniques are useful for revealing intra- and extrahepatic signs of primary sclerosing cholangitis.     

Proton spectroscopic imaging (Dixon method) of the liver: clinical utility

The contribution of proton spectroscopic (PS) imaging to magnetic resonance (MR) imaging of the liver was assessed at 0.5 T in 55 patients with known or suspected hepatic malignancy. PS images were compared subjectively with T1- and T2-weighted spin-echo (SE) images for hepatic lesion detection and conspicuity. For hepatic metastases (n = 27), PS images were equal to T1-weighted images in lesion detection in 17 patients but showed fewer lesions in five patients and false-negative results in two. When compared with T2-weighted images, PS images depicted more lesions in six patients, an equal number of lesions in 18, and fewer lesions in two. Hepatomas (n = 8) were detected with each sequence in all patients. Hepatomas were often more conspicuous on PS images than on T2-weighted images; they were of equal conspicuity on PS and T1-weighted images in most cases. Whereas fatty infiltration (n = 16) appeared on PS images as areas of low signal intensity similar to that of paraspinal muscle, it produced no detectable abnormality on either T1- or T2-weighted images. PS imaging is inferior to T1-weighted SE imaging in the detection of hepatic metastases. The major role of PS imaging at intermediate field strength is to differentiate focal fatty infiltration from hepatic metastases.     

CT and MR appearance of focal nodular hyperplasia of the liver in children with biliary atresia

Two children with biliary atresia are described in whom focal nodular hyperplasia of the liver occurred following portoenterostomy. The lesions were low-or iso-dense on unenhanced CT and became hypodense post-contrast enhancement. There was normal 99mTc phytate on hepatic colloid scintigraphy. On T2-weighted spin echo MR images, there was increased signal intensity within the masses, and the surrounding liver parenchyma was divided by linear septa in one of the two cases.     

Hepatic parenchymal changes after ethanol injection in rabbits: Correlation of conventional and dynamic MR imaging with pathologic findings

The purpose of this study is to clarify the changes of peripheral normal liver parenchyma before and after ethanol injection, with respect to MR appearances and pathologic findings and, in so doing, to clinically evaluate the therapeutic effectiveness of percutaneous ethanol injection (PEI) therapy. The normal liver in 12 rabbits was injected with ethanol. We performed conventional and dynamic MR imaging and prepared the histopathologic specimens 1 week (group 1), 2 weeks (group 2), and 1 month (group 3) after ethanol injection. On conventional MR images, coagulative necrosis in the normal liver was demonstrated as an area of low signal intensity on T1-weighted images and high signal intensity on T2-weighted images in all groups. On dynamic MR images, contrast enhancement in the coagulative necrosis was not seen in groups 1 and 2; however, gradual and concentric enhancement was seen in the direction of the central necrotic portion from early-to-delayed phase image in group 3. Although signal intensity of the coagulative necrotic area in the normal liver after ethanol injection may mimic that of untreated or viable hepatocellular carcinoma (HCC) when clinically encountered on conventional MR images, coagulative necrosis of the normal liver parenchyma will be discriminated from viable HCC by using dynamic MR imaging.     

T2-weighted MR imaging in the assessment of cirrhotic liver

PURPOSE: To assess if T2-weighted magnetic resonance (MR) imaging provides added diagnostic value in combination with dynamic gadolinium-enhanced MR imaging in the detection and characterization of nodular lesions in cirrhotic liver. MATERIALS AND METHODS: Two readers retrospectively and independently analyzed 54 MR imaging studies in 52 patients with cirrhosis. In session 1, readers reviewed T1-weighted and dynamic gadolinium-enhanced images. In session 2, readers reviewed T1-weighted, dynamic gadolinium-enhanced, and respiratory-triggered T2-weighted fast spin-echo images. Readers identified and characterized all focal lesions by using a scale of 1-4 (1, definitely benign; 4, definitely malignant). Multireader correlated receiver operating characteristic (ROC) analysis was employed to assess radiologist performance in session 2 compared with session 1. The difference in the areas under the ROC curves for the two sessions was tested. In a third session, readers assessed conspicuity of biopsy-proved lesions on T2-weighted MR images by using a scale of 1-3 (1, not seen; 3, well seen) and identified causes of reduced conspicuity. RESULTS: Two additional benign lesions were detected by each reader in session 2. Fifty-five lesions had pathologic verification, including 32 malignant, three high-grade dysplastic, and 20 benign nodules. There was no significant difference in the area under the ROC curves between the two sessions (P =.48). Thirty-two lesions were inconspicuous on T2-weighted MR images because of parenchymal heterogeneity, breathing artifacts (particularly in patients with ascites), and lesion isointensity with liver parenchyma. T2-weighted MR imaging was useful in the evaluation of cysts and lymph nodes. CONCLUSION: T2-weighted MR imaging does not provide added diagnostic value in the detection and characterization of focal lesions in cirrhotic liver.     

Colorectal mucinous carcinoma: findings on MRI

PURPOSE: The purpose of this work was to define the characteristic MR features of colorectal mucinous carcinomas and to correlate the mucin pool with the signal intensity of this tumor. METHOD: MRI of 12 cases of pathologically proven colorectal carcinoma containing mucin was evaluated. We analyzed the signal intensity of tumor on T1- and T2-weighted MR images and correlated the area of intratumoral high signal intensity on T2-weighted images with the mucinous pool on the pathologic specimens. Two radiologists independently estimated the area of high signal intensity in the tumor on T2-weighted images and one pathologist estimated the amount of mucinous pool in the pathologic specimen. RESULTS: In 9 (75%) of 12 cases, focal or diffuse high signal intensity areas were detected on T2-weighted fast spin echo images. In seven cases in which mucin pools were seen macroscopically, partial (n = 3) or diffuse high signal intensity areas were noted on the T2-weighted images. Among the five cases in which microscopic mucinous pools were detected on the pathologic slides, three cases showed no high signal foci on MR images, and in the remaining two cases, high signal intensity areas were noted as small foci. CONCLUSION: Intratumoral high signal intensity on T2-weighted fast spin echo MR images occurs in mucinous carcinomas and correlates with the mucin pools on pathologic specimens.     

MR imaging of subchondral osteonecrosis of the vertebral body after percutaneous laser diskectomy.

OBJECTIVE: We describe four cases of subchondral osteonecrosis of the vertebral body that occurred after percutaneous laser diskectomy. Follow-up MR imaging after laser intervention showed abnormal findings in the vertebral body immediately adjacent to the site of diskectomy that are consistent with subchondral osteonecrosis. CONCLUSION: MR imaging features of this complication include a wedge-shaped low signal intensity on T1-weighted images, high and low signal intensities on T2-weighted images, and a contrast-enhanced area corresponding to high signal intensity on T2-weighted images. Possible causative mechanisms include thermal injury and photoacoustic shock.     

Focal hepatic lesions: differentiation with MR imaging at 0.5 T

Magnetic resonance (MR) examinations of 43 patients with 95 focal hepatic lesions (diameter, greater than 1 cm) were analyzed for lesion shape, homogeneity, and relative signal intensity compared with normal liver parenchyma, spleen, and skeletal muscle. On T1-weighted, balanced, and T2-weighted images, most metastases (74%), cavernous hemangiomas (76%), and cysts (82%) were smooth and round or oval, while the hepatocellular carcinomas all had irregular borders (40%) or were lobulated (60%). All lesions with irregular borders were malignant. Seventy percent of metastatic lesions, 85% of cavernous hemangiomas, and 100% of simple hepatic cysts were of homogeneous signal intensity, while 60% of hepatocellular carcinomas were inhomogeneous. Logistic regression analysis of multiple lesion characteristics showed that inhomogeneous lesions had a high likelihood of malignancy, while markedly hyperintense lesions had a very low probability of being malignant, regardless of other traits. Homogeneous lesions that were isointense or hyperintense compared with spleen on balanced images but were not markedly hyperintense on T2-weighted images also had a high likelihood of malignancy.