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.     

Focal hepatic lesions: comparative MR imaging at 0.5 and 1.5 T

Twenty-nine patients with known or suspected malignancy were studied with identical T1-weighted (spin echo [SE] and inversion recovery [IR]) and T2-weighted SE magnetic resonance (MR) imaging at 0.5 and 1.5 T to evaluate the relative sensitivities of these sequences for detecting focal hepatic lesions. At 0.5 T, 98 lesions were detected with the IR sequence, 86 with the T1-weighted SE sequence, and 96 with the T2-weighted sequence. At 1.5 T, 93 lesions were detected with the IR sequence, 70 with the T1-weighted SE sequence, and 99 with the T2-weighted sequence. Although the lack of pathologic correlation precluded establishment of true sensitivity and specificity rates, data showed that magnetic field strength resulted in no significant difference for detecting focal hepatic lesions. No single sequence was shown to be significantly superior, although the T1-weighted SE sequence at 1.5 T was significantly inferior to the other sequences for detecting focal hepatic lesions. T1-weighted SE imaging at 0.5 T was significantly inferior to T1-weighted IR and T2-weighted imaging at both magnetic field strengths for detecting focal lesions in the left lobe of the liver. The authors conclude that T1-weighted IR and T2-weighted sequences alone will result in optimal MR imaging for the detection of focal hepatic lesions at 0.5 and 1.5 T.     

Phased array breath-hold versus non-breath-hold MR imaging of focal liver lesions: A prospective comparative study

This study was undertaken to determine whether phased array breath-hold T1- and T2-weighted sequences can replace non-breath-hold spin echo (SE) sequences in the imaging of focal liver lesions by comparing overall image quality, liver-lesion contrast, and artifact. Both breath-hold and non-breath-hold T1-weighted and T2-weighted imagings of focal liver lesions were prospectively compared in 120 patients with suspected focal liver lesions imaged at 1.5 T with use of a body phased array multicoil. Breath-hold images were acquired with T1-weighted fast low-angle shot (FLASH) and T2-weighted turbo spin echo (TSE) sequences, and non-breath-hold images were made with conventional T1- and T2-weighted SE sequences. Qualitative image analysis was done by three blinded readers, and quantitative analysis was done. The highest signal-to-noise ratios were obtained with breath-hold T1-weighted FLASH sequence. The signal-to-noise ratios of breath-hold T2-weighted TSE sequence were slightly inferior to those of non-breath-hold SE sequence. Both T1-weighted and T2-weighted breath-hold sequences had less image artifact. Overall image quality of breath-hold sequences was better than that of non-breath-hold sequences for both T1- and T2-weighted sequences (P < .01). The tissue contrast of T1-weighted FLASH sequence was superior to that of SE sequence (P < .01). On T2-weighted imaging, tissue contrast of solid lesions was better on conventional SE sequence than that on breath-hold TSE sequence (P < .01). Respiratory ghost artifact was less prominent on T1-weighted FLASH sequence, although this artifact was occasionally seen on breath-hold T2-weighted TSE sequence. In a state-of-art MR unit with use of a phased array multicoil, conventional T1-weighted can be replaced by breath-hold sequences. On T2-weighted imaging, because solid tumor-liver contrast on breath-hold TSE imaging is inferior to that on non-breath-hold SE image, breath-hold imaging may not replace conventional non-breath-hold T2-weighted SE sequence.     

MR imaging of experimentally induced intracranial hemorrhage in rabbits during the first 6 hours.

PURPOSE: To evaluate the MR appearance of intracranial, especially intraparenchymal, hemorrhage during the first 6 hours after bleeding with various pulse sequences in an animal model. MATERIAL AND METHODS: Intracerebral hematomas and subarachnoid hemorrhage were created by injecting autologous blood in 9 rabbits. MR studies were performed using a 1.5 T scanner with pixel size and slice thickness comparable to those used in clinical practice before blood injection, immediately after injection, and at regular intervals during 6 hours. The images were compared with the hematoma sizes on formalin-fixed brain slices. RESULTS: In every animal, susceptibility-weighted gradient-echo (GRE) pulse sequences depicted the intraparenchymal hematomas and blood escape in the ventricles or subarachnoid space best as areas of sharply defined, strong hypointensity. The findings remained essentially unchanged during follow-up. The sizes corresponded well to the post-mortem findings. Gradient- and spin-echo (GRASE) imaging revealed some hypointensities, but these were smaller and less well defined. Spin-echo (SE) sequences (proton density-, T1- and T2-weighted) as well as a fluid-attenuated inversion recovery turbo spin-echo sequence (fast FLAIR) depicted the hemorrhage sites as mostly isointense to brain. CONCLUSION: Susceptibility-weighted GRE imaging at 1.5 T is highly sensitive to both hyperacute hemorrhage in the brain parenchyma and to subarachnoid and intraventricular hemorrhage.     

Multisection FLASH: method for breath-hold MR imaging of the entire liver.

One hundred ten patients with various focal liver lesions were imaged with a multisection fast low-angle shot (FLASH) gradient-echo sequence with an echo time of 4.6 msec. This sequence enabled the acquisition of 19 T1-weighted magnetic resonance (MR) images of the liver within a single 26-second breath hold. Patients were also examined with standard T1- and T2-weighted spin-echo (SE) sequences. The multisection FLASH sequence provided significantly higher (P less than .01) liver-spleen contrast, liver-spleen signal-difference-to-noise ratio (SD/N), liver-tumor contrast, and liver-tumor SD/N than the T1-weighted SE sequence but lower values than the T2-weighted SE sequence. Motion artifacts were reduced with the multisection FLASH sequence compared with both SE sequences (P less than .01). The overall image quality of the multisection FLASH images was similar to that of the T1-weighted SE images and superior to that of T2-weighted SE images. The most important characteristics of the multisection FLASH technique in MR imaging of the liver are the high T1 contrast, the prevention of motion artifacts, and a dramatic reduction in imaging time.     

MR imaging of liver metastases at 1.5 T: similar contrast discrimination with T1- and T2-weighted pulse sequences

The authors evaluated soft-tissue contrast on spin-echo (SE) proton density-weighted, SE T2-weighted, SE short-echo-time (TE) T1-weighted, and gradient-echo (GRE) images of 34 patients with known hepatic tumors who underwent high-field-strength (1.5-T) magnetic resonance imaging. For solid liver tumors, the difference in the mean lesion-liver contrast-to-noise ratios (C/Ns) with T1- (GRE and SE) and T2-weighted pulse sequences was not statistically significant (P greater than .05). For nonsolid liver tumors, the T2-weighted images provided significantly greater (P less than .05) mean lesion-liver C/N than T1-weighted GRE images. Mean liver signal-to-noise ratio was significantly greater on T1-weighted GRE (P less than .0001) and T1-weighted SE (P less than .05) images than on T2- and proton density-weighted images. Qualitative analysis of T1-weighted (SE and GRE) images and proton density- plus T2-weighted images showed that lesion conspicuity was similar in 25 of 32 patients (78%). The results suggest that liver tumor imaging at high field strength can be performed with short-TE T1-weighted (SE or GRE) or conventional T2-weighted pulse sequences.     

Adrenal hemorrhage after orthotopic liver transplantation: MR appearance

The purpose of this paper is to describe the MR imaging findings of right adrenal hemorrhage after orthotopic liver transplantation. Twenty-seven orthotopic liver transplantation patients underwent MR studies of the liver and/or biliary system. Patients were referred to MR examination because of suspected biliary complications ( n=22) or for evaluation of mass lesions ( n=5). The standard MR protocol included T1-weighted spin-echo (SE) or gradient-recalled echo (GRE) images and T2-weighted turbo SE (TSE) images with fat suppression. In addition, cholangiography pulse sequences and/or contrast-enhanced T1-weighted images were obtained according to specific indications. In 2 patients a right adrenal mass was detected at MR imaging. Three to 4 weeks after transplantation, the lesions were markedly hyperintense on T2-weighted images and showed a hypointense capsule. Follow-up MR examinations revealed a slight decrease in size and a change in morphology. Computed tomography examinations of these 2 patients, obtained 10 weeks after transplantation, showed resolution of the hemorrhage and transformation into a cystic lesion in one case and a complete resolution of the hemorrhage and a normal right adrenal gland in the other case. Adrenal hemorrhage after liver transplantation shows typical MR features and should not be mistaken for an adrenal tumor or a postoperative abscess.     

Experimental acute intracerebral hemorrhage. Value of MR sequences for a safe diagnosis at 1.5 and 0.5 T

Purpose: To determine the detectability of intracerebral hematomas with MR imaging at 1.5 T and 0.5 T with fluid attenuated inversion recovery turbo spin-echo (FLAIR) and gradient-echo sequences.Material and Methods: Twenty-seven intracerebral hematomas were created in 25 piglets by injection of venous blood into the brain through a burr hole. All were imaged with T2*-weighted gradient echo sequences (fast field echo, FFE), T2-weighted fluid attenuated inversion recovery turbo spin-echo sequences (FLAIR), T2-weighted turbo spin-echo (TSE) and T1-weighted spin-echo sequences. Follow-up was performed on the 2nd, 4th and 10th postoperative days. Ten animals were additionally investigated with similar sequences at 0.5 T. Histologic correlation was obtained in all cases.Results: T2* FFE sequences detected all acute intracerebral hematomas and demonstrated the size correctly at 1.5 T and 0.5 T. The conspicuity was better at 1.5 T. FLAIR sequences were unreliable in the hyperacute phase at 1.5 T. However, subarachnoid and intraventricular extension was best appreciated with FLAIR images. T2 TSE images were incapable of detecting paraventricular and subarachnoid hemorrhages, but clearly demonstrated intracerebral blood in other locations. T1-weighted images were insensitive to hemorrhage in the acute state but very useful in subacute and chronic hematomas.Conclusion: The safe and reliable diagnosis of intracerebral hemorrhage is probably possible with MR imaging at 1.5 T and 0.5 T even of hematomas less than 90 min old, but requires the application of at least FLAIR, T2* FFE and T1 sequences and is therefore time consuming.     

MR imaging of focal lung lesions: elimination of flow and motion artifact by breath-hold ECG-gated and black-blood techniques on T2-weighted turbo SE and STIR sequences.

Respiratory and cardiac motion correction may result in better turbo spin-echo (SE) imaging of the lung. To compare breath-hold cardiac-gated black-blood T2-weighted turbo SE and turbo short-inversion-time inversion-recovery (STIR) magnetic resonance (MR) imaging pulse sequences with conventional breath-hold turbo SE and half-Fourier acquisition single-shot turbo spin-echo (HASTE) sequences for lesion conspicuity of focal lung lesions, 42 patients with focal lung lesions were prospectively studied with MR imaging at 1.5 T. Helical computed tomography was used as a reference. In comparison with the conventional breath-hold turbo SE sequence, all black-blood sequences had fewer image artifacts arising from the heart and blood flow. The overall image quality for the black-blood turbo SE and turbo STIR sequences was superior to that for the breath-hold turbo SE and HASTE sequence (P < 0.01). Not only focal lung lesions but also surrounding inflammatory changes were clearly visualized with these two sequences. With the HASTE sequence, although several slices could be obtained in one breath-hold, both the tumor and vessels appeared blurred. We conclude that T2-weighted turbo SE and turbo STIR imaging of the lung with effective suppression of flow and motion artifacts provide high-quality images in patients with focal lung lesions.     

MR imaging of acute hemorrhagic brain infarction

Six patients with acute hemorrhagic brain infarct were imaged using spin-echo (SE) pulse sequences on a 1.5 Tesla MR scanner. Including two patients with repeated MR imaging, a total of eight examinations, all performed within 15 days after stroke, were analyzed retrospectively. Four patients revealed massive hemorrhages in the basal ganglia or cerebellum and three cases demonstrated multiple linear hemorrhages in the cerebral cortex. On T1-weighted images, hemorrhages were either mildly or definitely hyperintense relative to gray matter, while varied from mildly hypointense to hyperintense on T2-weighted images. T1-weighted images were superior to T2-weighted images in detection of hemorrhage. CT failed to detect hemorrhages in two of five cases: indicative of MR superiority to CT in the diagnosis of acute hemorrhagic infarcts.     

Hepatic metastases: randomized, controlled comparison of detection with MR imaging and CT

To determine the accuracy of magnetic resonance (MR) imaging relative to computed tomography (CT) in the diagnosis of liver metastases, a randomized, controlled study was conducted of 135 subjects, including 57 with cancer metastatic to the liver, 27 with benign cysts or hemangiomas, and 51 without focal liver disease. The sensitivity of MR imaging for detecting individual metastatic deposits was 64%, significantly greater than 51% for CT (P less than .001); the difference in sensitivity for identifying patients with one or more hepatic metastases was less (82% for MR imaging vs. 80% for CT). In patients without hepatic metastases, the specificity of MR imaging was 99% versus 94% for CT. Significant differences were found between individual MR pulse sequences in detection of individual lesions. The sensitivity of both T1-weighted spin-echo (SE) (64%) and inversion-recovery (IR) (65%) pulse sequences was significantly (P less than .001) greater than either the TE (echo time) 60 msec (43%) or TE 120 msec (43%) T2-weighted pulse sequences. Overall, the accuracy of a single T1-weighted (10-minute) pulse sequence was superior to that of contrast-enhanced CT.     

MR imaging of hepatic focal nodular hyperplasia: characterization and distinction from primary malignant hepatic tumors

Spin-echo MR imaging at 0.35 T was used to image hepatic focal nodular hyperplasia (FNH) and to attempt to distinguish it from primary malignant hepatic tumors. There were six FNH and 10 malignant tumors including seven hepatocellular carcinomas, two cholangiocarcinomas, and one hepatoblastoma. Our results show that FNH has a fairly consistent appearance, dissimilar from that of malignant primary hepatic tumors. Four of six FNH lesions were isointense (except for a central scar in three) and indistinguishable from normal hepatic parenchyma on all pulse sequences, whereas two of six were homogeneous but slightly hyperintense on T2-weighted sequences. Furthermore, a central hyperintense scar was seen in three of six lesions on T2-weighted sequences. In contrast, each of the malignant primary hepatic tumors was hyperintense on T2-weighted sequences and seven of 10 were hypointense on T1-weighted sequences; in nine of 10, heterogeneous areas of intensity were noted. In two fibrolamellar hepatocellular carcinomas a central scar was seen that was hypointense on all pulse sequences. By using quantitative data, the best characterization was obtained by using lesion/normal-liver intensity ratios from a T2-weighted sequence; all FNH had a ratio less than 1.33, while in nine of 10 primary malignant tumors it was greater than 1.41. We conclude that focal nodular hyperplasia may have a consistent appearance on spin-echo MR imaging and probably can be distinguished from primary malignant lesions in most instances.     

Experimental intracerebral and subarachnoid/intraventricular haemorrhages

PURPOSE: To compare the detectability of small experimental intracranial haemorrhages on MR imaging at 0.5 T and 1.5 T, from hyperacute to subacute stages. MATERIAL AND METHODS: 1 ml of autologous blood was injected into the brain of 15 rabbits to create intraparenchymal haematomas. Since the blood partially escaped into the cerebrospinal fluid (CSF) spaces, detectability of subarachnoid and intraventricular blood was also evaluated. MR imaging at 0.5 T and at 1.5 T was repeated up to 14 days, including T1-, proton density- and T2-weighted (w) spin-echo (SE), FLAIR and T2*-w gradient echo (GE) pulse sequences. The last MR investigation was compared to the formalin-fixed brain sections in 7 animals. RESULTS: The intraparenchymal haematomas were best revealed with T2*-w GE sequences, with 100% of sensitivity at 1.5 T and 90-95% at 0.5 T. Blood in the CSF spaces was significantly (p < 0.05) better detected at 1.5 T with T2*-w GE sequences and detected best during the first 2 days. The next most sensitive sequence for intracranial blood was FLAIR. SE sequences were rather insensitive. CONCLUSION: 1.5 T equipment is superior to 0.5 T in the detection of intracranial haemorrhages from acute to subacute stages. T2*-w GE sequences account for this result but other sequences are also needed for a complete examination.     

腹部磁共振成像的临床应用──探讨和比较几种扫描脉冲序列

笔者通过分析１００例腹部ＭＲＩ图像，研究和比较３对扫描脉冲序列的应用；研究结果表明：（１）Ｔ＿１加权像脉冲序列，快速ＳＰＧＲ（屏气）扫描速度快、呼吸伪影少、流动伪影多；而ＳＥ序列扫描时间长、流动伪影少、呼吸伪影多些；两者脉冲序列图像中解剖分辨和病变显示率相似（Ｐ＞０．０５，无显著差异）。（２）Ｔ＿２加权像脉冲序列，ＦＳＥ序列扫描时间短、呼吸伪影和流动伪影多；而ＳＥ序列扫描时间长、呼吸伪影和流动伪影少些；ＦＳＥ和ＳＥ图像中病变显示率相似（Ｐ＞０．０５，无显著差异）；ＳＥ图像中解剖分辨好于ＦｓＥ（Ｐ＜０．０１．有极显著差异）。（３）Ｔ＿２加权像加脂肪抑制序列图像中呼吸伪影、化学位移伪彤少，解剖分辨和病变显示清晰度均好于Ｔ＿２加权像不加脂肪抑制序列（Ｐ＜０．０ｌ，有极显著差异）。     

MR detectability and appearance of small experimental intracranial hematomas at 1.5 T and 0.5 T. A 6-7-month follow-up study

PURPOSE: To investigate the detectability and appearance of small experimental intracranial hemorrhages on MR at 0.5 T and 1.5 T in a long-term follow-up. MATERIAL AND METHODS: Autologous blood (1 ml) was injected into the brain of 7 rabbits to create intraparenchymal hematomas. The injected blood leaked partially into the cerebrospinal fluid (CSF) spaces. MR imaging at 0.5 T and 1.5 T were performed immediately before and after hematoma creation, at 2 weeks and monthly up to 6 or 7 months using T1-, proton density- and T2-weighted (w) spin-echo (SE), FLAIR and T2*-w gradient echo (GE) pulse sequences. RESULTS: Blood was detected both in the brain and in the CSF spaces of all animals during the first hours after hematoma creation at 1.5 T. In the last examination after 6-7 months, the T2*-w GE sequences still showed residues of the intraparenchymal hematomas in all the rabbits at 1.5 T, but the signal pattern was not specific for the age of the hematomas. SE and FLAIR sequences were insensitive. The histopathology revealed iron deposits in all brains. CONCLUSION: Residues of small intraparenchymal hematomas can be seen for months with T2*-w GE sequences on brain MR imaging at 1.5 T. The age of the microhematomas cannot be estimated with MR imaging.     

Detection of articular cartilage lesions: experimental evaluation of low- and high-field-strength MR imaging at 0.18 and 1.0 T

The objective of this study was to compare the diagnostic performance of a dedicated orthopedic magnetic resonance (MR) imaging system (0.18 T) and a conventional MR imaging system (1.0 T) in the detection of articular cartilage lesions. Fifty knee joint specimens of pigs with artificially created articular cartilage lesions of different diameters, grades (2-3), and localizations, as well as 50 joints with intact articular cartilage, were imaged at 0. 18 and 1.0 T. Diagnostic performance was determined by means of receiver operating characteristics (ROC) analysis with three independent observers. For none of the pulse sequences used at 0.18 T or 1.0 T areas under ROC curves (A(z)) showed significant differences between the three observers. A(z) values from averaged data were as follows: a) 0.18 T: T1-weighted spin echo (SE): 0.70, proton-density-weighted SE: 0.59, T2-weighted SE: 0.61, two-dimensional (2D) gradient-echo (GRE): 0.73, 3D GRE: 0.75; and b) 1.0 T: T1-weighted SE: 0.73, fat-suppressed T2-weighted turbo-SE: 0. 79, 2D fast low-angle shot (FLASH): 0.79, fat-suppressed 3D FLASH: 0. 96, and water-excited 3D double-echo steady state (DESS): 0.96. With the use of 3D pulse sequences, the high-field system demonstrated a significantly better diagnostic performance than the low-field system in the detection of grades 2 and 3 articular cartilage lesions (P < 0.001).     

Improved detection of intraventricular cysticercal cysts with the use of three-dimensional constructive interference in steady state MR sequences

BACKGROUND AND PURPOSE: Before the advent of MR imaging, intraventricular cysts were difficult to diagnose noninvasively. Among the invasive procedures used were contrast ventriculography and CT ventriculography. MR imaging, with its multiplanar imaging capabilities, excellent depiction of tissue contrast, and versatile parameters, is an important tool in the assessment of intraventricular cystic lesions. We investigated the role of three-dimensional constructive interference in steady state (3D-CISS) MR sequences in the evaluation of intraventricular cysticercal cysts. METHODS: The study group comprised 11 patients with intraventricular cysticercal cysts. MR studies included spin-echo (SE) T1-weighted, turbo-SE T2-weighted, and 3D-CISS sequences. All images were obtained on a superconducting 1.5-T MR unit. The routine and 3D-CISS sequences were reviewed and interpreted separately by two neuroradiologists. RESULTS: All patients underwent surgery for excision of intraventricular cysticercal cysts. Eight patients had cysts in the fourth ventricle, two in the lateral ventricle, and one in the third ventricle. SE T1-weighted images showed the cystic wall in nine cases, the scolex in four, and the cystic fluid in two. Turbo-SE T2-weighted images showed the cystic wall and scolex in three and four cases, respectively. The routine sequences did not show the scolex, cystic wall, or cystic fluid together in any of the 11 patients. 3D-CISS images showed the scolex in all 11 patients and the cystic wall and cystic fluid in eight patients each. In seven of the 11 patients, 3D-CISS images showed the scolex, cystic wall, and fluid together. CONCLUSION: The 3D-CISS sequence is more sensitive and specific than routine SE sequences in the diagnosis of intraventricular cysticercal cysts.     

Focal liver lesions: MR imaging with Mn-DPDP--initial clinical results in 40 patients.

Manganese (II) N,N'-dipyridoxylethylenediamine-N,N'-diacetate-5,5'-bis(phosphate) (DPDP) was evaluated as a contrast agent for magnetic resonance (MR) imaging (1.5 T) of focal liver lesions in 40 patients. Doses of 5 and 10 mumol/kg were administered intravenously. Mn-DPDP-enhanced T1-weighted images were compared quantitatively and subjectively with standard T1- and T2-weighted nonenhanced images. Use of Mn-DPDP resulted in a statistically significant increase in signal intensity of liver parenchyma in T1-weighted images at both doses. No enhancement was seen in metastases, cholangiocarcinomas, or lymphomas, while all hepatocellular carcinomas were enhanced. Enhancement was seen in focal nodular hyperplasia and in regenerative nodules. The lesion-to-liver contrast in Mn-DPDP-enhanced gradient-recalled-echo images was superior to that of all precontrast images (P less than .01). The number of nonenhancing malignant liver lesions detected in spin-echo (SE) images was increased (272 in T2-weighted SE images vs 390 in T1-weighted Mn-DPDP-enhanced SE images). Image interpretation (eg, visualization and demarcation of the lesions) was markedly better in Mn-DPDP-enhanced images than in all precontrast images (P less than .001).     

Malignant lesions of the liver with high signal intensity on T1-Weighted MR images

Our purpose was to identify the histologic types of malignant liver lesions with high signal intensity (SI) on T1-weighted images and to describe the MR imaging features. Thirteen patients with malignant liver lesions high in SI on T1-weighted images were studied with a 1.5-T MR imager using pre- and serial postcontrast spoiled gradient-echo (SGE) sequences (all patients), T2-weighted fat-suppressed spin-echo sequences (all patients), precontrast T1-weighted fat-suppressed spin-echo sequences (five studies in five patients), and precontrast out-of-phase SGE sequences (seven studies in six patients). Images were reviewed retrospectively to determine number of lesions; lesion size; SI of lesions on T1-weighted, T2-weighted, and fat-attenuated T1-weighted images; distribution of high SI in lesions on T1-weighted images; and tumor enhancement pattern. Seven patients had multiple tumors high in SI on T1-weighted images and six patients had solitary tumors. Seventy-two lesions were less than 1.5 cm in diameter and 35 lesions were more than 1.5 cm in diameter. Nine patients had solid malignant lesions and four patients had cystic malignant lesions. All tumors more than 1.5 cm in diameter were heterogeneously high in SI on T1-weighted images, and all tumors less than 1.5 cm were completely homogeneous or homogeneous with a small central hypointense focus. All tumors were more conspicuous on T1-weighted fat-attenuated images, both on excitation spoiled fat-suppressed spin-echo or on out-of-phase SGE images with the exception of one fat-containing hepatocellular carcinoma (HCC). In one patient with melanoma metastases and one patient with multiple myeloma nodules, appreciably more lesions were detected on out-of-phase SGE images. Causes of hyperintensity were considered to be either fat, melanin, central hemorrhage, or high protein content, all of which may be seen in a variety of tumors. Fat-attenuation techniques are helpful in the detection of these lesions.     

Liver metastases: detection by phase-contrast MR imaging

Forty patients with biopsy-proved metastatic liver cancers were studied by magnetic resonance (MR) imaging using one or more conventional (in-phase) pulse sequences and a corresponding phase-contrast (opposed-phase) pulse sequence. Pulse-sequence performance was quantitated by measuring signal-difference-to-noise (SD/N) ratios between cancerous tissue and liver. The SD/N performance of T2-weighted spin-echo (SE) pulse sequences improved when used with the phase-contrast technique. SE 2,000/30 opposed-phase images showed improved (P less than .001) SD/N in 72% of patients over in-phase images. The SD/N of T1-weighted SE or inversion recovery pulse sequences deteriorated when used with the phase-contrast technique. Changes in measured SD/N correlated well with image appearance and actual lesion detectability in individual cases. Phase-contrast imaging should be employed routinely when T2-weighted SE pulse sequences are relied on to detect liver cancer.