Differentiation between hemangiomas and cysts of the liver with nonenhanced MR imaging: Efficacy of T2 values at 1.5 T

The authors studied the utility of non-contrast-agent-enhanced magnetic resonance (MR) imaging for differentiating cysts and cavernous hemangiomas of the liver. Nineteen patients with hemangiomas (51 lesions) and 16 with cysts (30 lesions) were studied with a 1.5-T MR imager. T2 values were calculated with the two-point method to evaluate the efficacy of T2 values in the differentiation between hemangiomas and cysts of the liver. For lesions larger than 1 cm, the mean T2 value of cysts (306 msec ± 156) was significantly longer than that of hemangiomas (113 msec ± 15) (P < .0001); there was no overlap of the ranges for T2 values of hemangiomas and cysts. All cysts larger than 1 cm could be differentiated from hemangiomas by using a threshold T2 value of 140 msec. This study suggests that calculated T2 values permit differentiation between hemangiomas and cysts larger than 1 cm at 1.5 T.     

Spin-echo and dynamic gadoliniumenhanced flash MR imaging of hepatocellular carcinoma: Correlation with histopathologic findings

Evaluation of histologic subtype and degree of differentiation in hepatocellular carcinoma (HCC) is essential because it affects patient prognosis and treatment planning. To evaluate the histologic subtype of HCC with magnetic resonance (MR) imaging, conventional spin-echo and dynamic studies were correlated with histopathologic and angiographic findings in 72 HCCs. Dynamic MR imaging was performed with the fast low-angle shot (FLASH) technique after administration of gadopentetate dimeglumine. There was considerable overlap in signal intensity between various tumor grades on both T1- and T2-weighted images. On dynamic MR images, the peak contrast enhancement ratio correlated with tumor grade (well-differentiated, 29.5 ± 24.7; moderately differentiated, 63.5 ± 24.1; poorly differentiated, 86.9 ± 26.4) or degree of dilatation of the sinusoidlike vascular space between tumor cells. The maximum contrast-to-noise ratio in tumor (relative to surrounding liver) was achieved within 60 seconds in 45 HCCs (mostly of the trabecular or pseudoglandular type). Enhancement was slight or minimal in 17 tumors (mostly small, well-differentiated tumors). In 10 tumors, the degree of enhancement increased with time, with maximum enhancement in the delayed phase (most frequently in scirrhous HCC). These dynamic patterns correlated with angiographic findings. These data indicate that the degree and pattern of enhancement on dynamic MR images reflect tumor differentiation and architecture of HCC.     

Evaluation of hepatocyte-specific paramagnetic contrast media for MR imaging of hepatitis

The hepatocyte-specific paramagnetic magnetic resonance (MR) contrast agents manganese-DPDP [N,N′-dipyridoxylethylenediamine-N,N′-diacetate 5,5′ bis-(phosphate)] and gadobenate dimeglumine were used for diagnosing chemically induced hepatitis in rats. Ex vivo liver tissue relaxation times and in vivo MR image signal-to-noise ratios were compared before and after contrast agent administration. Ex vivo relaxometry and in vivo MR imaging showed that Mn-DPDP enhanced normal and diseased livers to the same degree at all time points from 5 to 120 minutes. Gadobenate dimeglumine showed reduced T1 and T2 enhancements in hepatitis relative to those of normal liver, in the early phase (5–30 minutes). However, these effects are offsetting, and as a result, MR imaging failed to allow distinction of diseased from normal livers. This surprising result observed in vivo was in fact predicted by applying the Bloch equation to our ex vivo data. Our results show that detection and quantitation of hepatitis with MR imaging enhanced with paramagnetic cellspecific contrast agents will be more difficult than anticipated.     

Mn-DPDP–enhanced MR imaging of malignant liver lesions: Efficacy and safety in 20 patients

Twenty patients with malignant liver lesions underwent magnetic resonance (MR) imaging with manganese (II) DPDP [N,N′-dipyridoxylethylenediamine-N,N′-diacetate 5,5′-bis(phosphate)] to evaluate the safety and efficacy of the contrast agent. In two groups of 10 patients each, 5 μmol/kg Mn-DPDP was administered intravenously (3 mL/min) at a concentration of either 50 or 10 μmol/mL. T1- and T2-weighted images were obtained with a 1.5-T imager. Six patients reported a total of eight instances of side effects (flush, feeling of warmth, metallic taste) of which seven occured at the 50 μmol/mL concentration. A significant decrease in alkaline phosphatase levels 2 hours after injection was recorded. On T1-weighted images, the 10 μmol/mL formulation yielded significantly greater increases in contrast-to-noise ratio (79.8%–137.5%) than the 50 μmol/mL formulation (46.2%–86.6%). In a blinded reader study of 10 patients with one to five lesions each, no lesion was missed on Mn-DPDP–enhanced T1-weighted images; however, four false-positive foci were identified. The authors conclude that slow administration of 5 μmol/kg Mn-DPDP at a concentration of 10 μmol/mL is safe and efficient enough to proceed to further clinical trials.     

Iron overload in the liver diagnostic and quantification

Hereditary Hemochromatosis is the most frequent modality of iron overload. Since 1996 genetic tests have facilitated significantly the non-invasive diagnosis of the disease. There are however many cases of negative genetic tests that require confirmation by hepatic iron quantification which is traditionally performed by hepatic biopsy. There are many studies that have demonstrated the possibility of performing hepatic iron quantification with Magnetic Resonance. However, a consensus has not been reached yet regarding the technique or the possibility to reproduce the same method of calculus in different machines. This article reviews the state of the art of the question and delineates possible future lines to standardise this non-invasive method of hepatic iron quantification.     

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.     

Nodular sarcoidosis of the liver and spleen: Appearance on MR images

Small nodular lesions in the liver and spleen have been reported as an infrequent manifestation of sarcoidosis. Five patients with this appearance on either dynamic contrast material—enhanced computed tomographic (CT) or ultrasound scans underwent magnetic resonance (MR) imaging with and without dynamic gadolinium enhancement. The lesions were relatively uniform in size, ranging from 0.5 to 1.5 cm. On CT scans, they were hypoattenuating relative to surrounding parenchyma. On MR images, the lesions were hypointense relative to background parenchyma with all sequences. No substantial enhancement was observed in the lesions, although lesion conspicuity decreased over time on serial postcontrast images. Lesion conspicuity was greatest on either T2-weighted fat-suppressed (T2FS) images or early-phase dynamic contrast-enhanced images. Abdominal adenopathy was seen in three of the five patients and was hyperintense relative to liver on T2FS images in two and intermediate in intensity in one patient.     

Comparison of breath-hold T1-weighted MR sequences for imaging of the liver

Three rapid T1-weighted gradient-echo techniques for imaging of the liver were compared: fast low-angle shot (FLASH) and section-selective (SSTF) and non-section-selective (NSTF) inversion-recovery TurboFLASH. Ten healthy volunteers were imaged at 1.5 T, with breath-hold images acquired in the transaxial and coronal planes and non-breath-hold images in the transaxial plane. Breath-hold images were evaluated quantitatively and qualitatively, and non–breath-hold images were evaluated qualitatively. FLASH images had significantly higher (P <.001) spleenliver signal difference–to-noise ratios (SD/Ns) than NSTF and SSTF images. Liver signal-to-noise ratios (S/Ns) were significantly higher (P <.001) on FLASH images than on NSTF and SSTF images. With breath hold. FLASH images were rated as having the highest quality in the axial plane, followed by NSTF and SSTF images. In the coronal plane, NSTF images were rated as having the highest quality. For images acquired during patient respiration, NSTF images had the highest quality and showed the least degradation. The results suggest that FLASH images have the highest SD/N and S/N for liver imaging and have the highest quality in the axial plane. In patients who cannot suspend respiration, NSTF images may be least affected by breathing artifact and provide reasonable image quality.     

Diagnosis of fatty liver with MR imaging.

The diagnosis of fatty liver with magnetic resonance (MR) imaging was evaluated in experimental rat models of simple fatty infiltration and fatty liver with hepatocellular injury. T1 and T2 were measured ex vivo and correlated with the histologic degree of fatty infiltration. Enhancement of fatty liver with four different cells-specific contrast agents was studied with ex vivo relaxometry and in vivo MR imaging. Quantitative analysis of conventional and chemical shift MR images was correlated with biochemically determined fat content of the liver. Diet-induced simple fatty infiltration of the liver caused a decrease in T1 of 15%, whereas the T1 of L-ethionine-induced fatty liver with hepatocellular injury increased by 12%. T2 showed a positive correlation with the degree of fatty infiltration in both models. Cell-specific hepatobiliary contrast agents showed the same liver uptake and relaxation enhancement in fatty livers as in normal livers. Conventional T1-weighted images and chemical shift images showed good correlation (r = .83 and .80, respectively) between signal intensity and the degree of fatty infiltration. However, only chemical shift imaging was reliable in the diagnosis of fatty liver.     

Recurrent biliary cystadenoma: MR imaging appearance

The authors report the case of a 37-year-old woman with a biliary cystadenoma that mimicked a liver cyst. The magnetic resonance (MR) imaging features of this rare lesion were correlated with the pathologic findings, showing the potential of MR imaging for depicting and aiding in the diagnosis of biliary cystadenoma and its recurrence.     

Cirrhosis: Multiobserver analysis of hepatic MR imaging findings in a heterogeneous population

To investigate the magnetic resonance (MR) imaging findings of hepatic cirrhosis and the potential of MR imaging in differentiating cirrhosis from other hepatic abnormalities, three observers with different levels of expertise in MR imaging (specialist, experienced radiologist, and novice with special training) reviewed hepatic MR imaging examinations of 52 patients with biopsy-proved presence (n = 29) or absence (n = 23) of cirrhosis. All examinations included motion-compensated T1-weighted, T2-weighted, and flow-sensitive gradientecho images. For all three observers, linear signal irregularity was more accurate than other findings. For the final diagnosis of cirrhosis, the specialist was most sensitive (76% at high threshold, 97% at low threshold), followed by the novice with special training (31% and 79%, respectively). Specificity was 100% for all observers at high threshold and 78%, 96%, and 87% for expert, experienced, and trained novice observers, respectively, at low threshold. Sensitivity did not vary regardless of severity of fibrosis, as determined independently by a hepatopathologist. MR imaging has the potential of offering a comprehensive noninvasive evaluation of patients with suspected cirrhosis, but considerable expertise is required at present.     

Malignant lesions of the liver identified on T1- but not T2-weighted MR images at 1.5 T

The authors reviewed their 21/2-year experience with a magnetic resonance (MR) imaging protocol for a 1.5-T MR imager that included T2-weighted fat-suppressed spin-echo, T1-weighted breath-hold gradient-echo, and serial dynamic gadolinium-enhanced T1-weighted gradient-echo imaging to identify histologic types of malignant liver lesions more apparent on T1- than on T2-weighted images. MR images of 212 consecutive patients with malignant liver lesions were reviewed. T2-weighted, T1-weighted, and dynamic contrast-enhanced T1-weighted images were examined separately in a blinded fashion. Seven patients demonstrated liver lesions (lymphoma [two patients] and carcinoid, hepatocellular carcinoma, colon adenocarcinoma, transitional cell carcinoma, and melanoma [one patient each]) on T1-weighted images that were inconspicuous on T2-weighted images. In all cases, the lesions were most conspicuous on T1-weighted images obtained immediately after administration of contrast agent. Histologic confirmation was present for all seven patients. The consistent feature among these lesions was that they were hypovascular, due either to a fibrous stroma or to dense monoclonal cellularity. These results suggest that in some patients with hypovascular primary neoplasms, the lesions may be identified only on T1-weighted images, and that immediate postcontrast T1-weighted images are of particular value in demonstrating lesions.     

Quantitative mr imaging data in the evaluation of hepatic metastases during systemic chemotherapy

To identify changes induced by chemotherapy in hepatic metastases, 34 patients with metastases underwent magnetic resonance (MR) imaging before the start of systemic chemotherapy and monthly thereafter. The number, size, and morphologic patterns of the lesions and changes in quantitative parameters (signal-to-noise ratio [S/N], contrast-to-noise ratio, and signal intensity ratio) were evaluated and correlated with response to treatment and prognosis. After treatment, seven patients showed a partial response, 18 had stable disease, and nine had progressive disease. No relevant changes in the patterns of the lesions were observed. Quantitative data showed that patients with a good prognosis had an increase in S/N on T1-weighted images and a relative decrease on T2-weighted images; patients with a poor prognosis showed a decrease in S/N on T1-weighted images and an increase on T2-weighted images. The differences between patient groups were significant for both T1- and T2- weighted images. This study demonstrates the value of MR imaging in follow-up of liver metastases and suggests the usefulness of quantitative MR imaging data.     

MR monitoring during cryotherapy in the liver: Predictability of histologic outcome

For well-controlled application of cryotherapy to focal liver lesions, real-time monitoring is necessary to limit the final necrotic effect in the treated area. In this study, near real-time magnetic resonance (MR) monitoring images of normal rabbit liver were obtained during the freezing procedure. The MR-monitored, freezing-induced lesions were followed with MR images for up to 3 weeks. Corresponding histologic samples were also obtained over the same time period. Our results indicate that MR images obtained during the freezing procedure can adequately depict the area of final necrosis. Furthermore, histologic changes at each stage of lesion development correlated well with MR signal intensities on follow-up images. With the development of an MR-compatible cryogen probe, MR imaging may prove to be a robust method for monitoring, controlling, and following up cryotherapy in the liver.     

MR imaging of blood oxygenation-dependent changes in focal renal ischemia and transplanted liver tumor in rat

The potential of using fast magnetic resonance (MR) imaging in conjunction with apnea-induced blood deoxygenation for the noninvasive monitoring of relative perfusion in the rat abdomen has been studied with two experimental models: glycer-ol-induced focal renal ischemia and transplanted liver tumor. Gradient-echo echo-planar imaging (GRE-EPI) (TE of 20 msec at 2 T) of liver and kidney was performed before, during, and after a 60-second apnea episode and then was followed in the same rat by contrast-enhanced (a) GRE-EPI and (b) T1-weighted spin-echo imaging (TR msec/TE msec = 200/6) with polylysine-(gadolmium-DTPA [diethylenetriaminepentaacetic acid]). The results indicate that a noninvasive vascular challenge due to apnea can be used for the detection of focal tissue perfusion abnormalities in rat kidney and liver tumor.     

Differentiation between hemangiomas and hepatocellular carcinomas with the apparent diffusion coefficient calculated from turboflash MR images

Serial turboFLASH (fast low-angle shot) images with and without diffusion-perfusion (DP) gradients were used for the evaluation of and differentiation between hemangiomas and hepatocellular carcinomas (HCCs) of the liver. Twenty-six patients with 27 hemangiomas, and 19 patients with 21 HCCs were studied. T1- and T2-weighted spin-echo images, serial turboFLASH images with and without DP gradients (b = 294.8 and 0 sec/mm², respectively) were obtained, and the apparent diffusion coefficient (ADC) was calculated for all lesions. Hemangiomas were detected as well-defined areas of decreased signal intensity on turboFLASH images obtained with DP gradients; HCCs showed slight or no obvious decrease in signal intensity on serial turboFLASH images obtained with DP gradients compared with turboFLASH images obtained without DP gradients, while a considerable percentage (76.5%) of nodular HCCs showed an obvious decrease in signal intensity at the margins. Hemangiomas had large ADC values compared with HCCs. Both ADC and T2 values were significantly different between hemangiomas and HCCs (P < 0.01). However, there was no obvious correlation between ADC and T2 values for either hemangiomas or HCCs (r = 0.18 and 0.48, respectively). On the basis of these results, the calculated ADC should be helpful for distinguishing hemangiomas from HCCs, and the ADC values may be useful even when T2 values are not helpful for making the distinction.     

Contrast agents for MR imaging of the liver

The evolution of contrast agents for MR imaging of the liver has proceeded along several different paths with the common goal of improving liver-lesion contrast. These contrast agents are used to accentuate the inherent differences in liver-lesion signal intensity through differential enhancement of proton relaxation within adjacent tissues. Contrast agents used for hepatic MR imaging can be broadly categorized into those that target the extracellular space, the hepatobiliary system, and the reticuloendothelial system. Although only a small number of liver contrast agents are currently available, others are rapidly proceeding through clinical trials and may soon be added to our clinical armamentarium. This article will briefly review the current clinical experience with these agents, discussing their mechanism of contrast enhancement, pharmacokinetics, and efficacy in the evaluation of focal liver lesions.     

Quantitation of total metastatic tumor volume in the rat liver: correlation of MR and histologic measurements.

Assessing tumor response to chemotherapy in the liver has always been difficult. Most investigators estimate tumor volume as either a product of the two perpendicular diameters of a tumor nodule, or, in animal studies, simply count surface tumor nodules. The authors evaluated magnetic resonance (MR) imaging as a technique for determining absolute tumor volume in the liver in an animal model. Specifically, histologic volumetric and MR imaging measurements of tumor and liver volumes were quantitatively compared over a wide range of tumor burdens in a rat model of hepatic metastasis of a colorectal carcinoma. Twenty-three rats were imaged, with two different section thicknesses used in each animal. Both section thicknesses showed highly significant correlations between MR and histologic measurements for both tumor and liver volumes (P less than .001). MR imaging may be useful for noninvasively quantifying tumor burden and temporal response of metastatic disease in the liver to novel antineoplastic regimens.     

MR imaging assessment of experimental hepatic dysfunction with mn-DPDP

Manganese (II) N, N'-dipyridoxylethylenediamine-N, N'-diacetate 5,5′-bis(phosphate) (DPDP) is a paramagnetic magnetic resonance (MR) contrast agent that enhances the liver and is predominantly excreted through the biliary tree. The authors evaluated its utility in diffuse liver disease by assessing liver and gallbladder enhancement in 24 rabbits. Total (n = 6) or segmental (n = 6) biliary occlusion or galactosamine-induced hepatitis (n = 6) was induced 3 days before imaging. Six rabbits served as normal controls. T1- and T2-weighted axial MR images were acquired at baseline, followed by T1-weighted images every 10 minutes for 1 hour after the intravenous administration of 20 μmol/kg Mn-DPDP. Except for the segmental occlusion group, the baseline study did not allow distinction between normal and abnormal livers. The temporal hepatic enhancement pattern was statistically different for each group. The normal, segmental occlusion, and hepatitis groups showed patterns similar to one another but markedly higher signal intensity than the total-occlusion group throughout the observation period. In contrast, the gallbladder showed a greater difference in both degree of enhancement and time to peak enhancement among the four groups. Mn-DPDP produces a temporal hepatic enhancement pattern that allows recognition of markedly impaired livers, and gallbladder enhancement patterns that allow distinction of more subtly impaired livers.     

Malignant liver lesions: Comparison of spiral CT arterial portography and MR imaging for diagnostic accuracy,cost, and effect on patient management

We compared two imaging techniques, spiral CT arterial portography (CTAP) and MR imaging, for diagnostic accuracy, procedural cost, and effect on management of 26 patients referred for hepatic surgery for suspected limited malignant liver disease. CTAP and MR imaging were done within a 1-week period (19 within 24 hours); the results of the studies were interpreted prospectively by separate reviewers. Surgical data were evaluated in conjunction with imaging data in 10 patients. Lesion detection and segmental involvement were determined and sensitivity and specificity were calculated. Procedural cost was determined from hospital billing codes. Effect on patient management was determined by the referring oncologic surgeon. CTAP and MR imaging showed 185 and 176 true-positive malignant lesions, 15 and zero false-positive malignant lesions, zero and 18 true-negative malignant lesions, and 13 and 22 falsenegative malignant lesions, respectively. CTAP and MR imaging showed 107 and 105 true-positive segments, 11 and zero false-positive segments, 80 and 91 true-negative segments, and four and six false-negative segments, respectively. There was a significant difference in specificity of segmental involvement between MR imaging (1.0 ± 0) compared with CTAP (0.88 ± 0.05), P=.03. Total procedural cost was $3,499 for CTAP and $1,224 for MR imaging. CTAP findings did not change patient management over MR imaging findings in any patient, whereas MR imaging findings resulted in a change in patient management over CTAP findings in seven patients (P=.015). The results of our study suggest that MR imaging has higher diagnostic accuracy and greater effect on patient management than CTAP does and is 64% less expensive.