Using Superparamagnetic Iron Oxide Enhanced MRI to Differentiate Metastatic Hepatic Tumors and Nonsolid Benign Lesions

Transcription

1 Downloaded from by on 11/27/17 from IP address Copyright RRS. For personal use only; all rights reserved Seishi Kumano 1 Takamichi Murakami Tonsok Kim Masatoshi Hori tsuya Okada Takashi Sugiura Yumi Noguchi Syuji Kawata Kaname Tomoda Hironobu Nakamura Received ugust 19, 2002; accepted after revision May 14, ll authors: Department of Radiology, Osaka University Graduate School of Medicine, 2-2 Yamadaoka, Suita, Osaka Japan. ddress correspondence to T. Murakami. JR 2003;181: X/03/ merican Roentgen Ray Society Using Superparamagnetic Iron Oxide Enhanced MRI to Differentiate Metastatic Hepatic Tumors and Nonsolid enign Lesions OJECTIVE. We evaluated the ability of superparamagnetic iron oxide (SPIO)-enhanced MRI to differentiate solid metastatic tumors and nonsolid benign lesions by clarifying the characteristic signal-intensity pattern of each lesion on SPIO-enhanced T2-weighted and heavily T1-weighted gradient-echo images. MTERILS ND METHODS. SPIO-enhanced MRI was performed using a 1.5-T system in 33 consecutive patients without cirrhosis who had 81 focal hepatic lesions (42 cysts, 13 hemangiomas, 26 metastatic tumors). The relative signal intensity of lesions on SPIO-enhanced heavily T1- and T2-weighted gradient-echo images was classified into one of the following three categories: high intensity, isointensity, or low intensity relative to the surrounding liver parenchyma. The diagnostic accuracy for differentiating solid metastatic tumors from nonsolid benign lesions (cysts or hemangiomas) was determined. RESULTS. combination of the relative signal intensity of the lesion on T2- and heavily T1-weighted gradient-echo images could be classified into the following five categories: high intensity and high intensity (category 1), high intensity and isointensity (category 2), high intensity and low intensity (category 3), isointensity and isointensity (category 4), and isointensity and low intensity (category 5). ccording to these categories, category 1 contained two hemangiomas, category 2 had 11 hemangiomas, category 3 had 25 metastatic tumors and two cysts, category 4 had three cysts, and category 5 had 37 cysts and one metastatic tumor. When a tumor with a relative signal intensity of categories 1 or 2 was considered to be a hemangioma (category 3 metastatic tumors and categories 4 and 5 cysts), diagnostic accuracy for characterizing such hepatic lesions was 96% (78/81). CONCLUSION. When evaluating metastatic liver tumors on SPIO-enhanced MRI, we recommend that heavily T1- and T2-weighted gradient-echo images be obtained with our parameters to exclude hemangiomas or cysts. T he detection and characterization of focal hepatic lesions are an important aspect of liver imaging that has implications for patient treatment and prognosis. MRI has excellent soft-tissue contrast and has improved our ability to detect and characterize focal hepatic lesions by allowing us to evaluate the hemodynamics of each lesion in combination with the selective use of extracellular contrast agents such as gadopentetate dimeglumine [1]. Superparamagnetic iron oxide (SPIO) was developed to improve the detectability of hepatic tumors without Kupffer s cells. For this type of imaging, T2-weighted sequences are usually used because SPIO shortens the T2 value of liver parenchyma [2 11]. SPIO also provides new possibilities for characterization of the lesion based on its cellular composition and function rather than on its vascularity and diffusion within the extracellular space [12]. However, when evaluating metastatic liver tumors, it is sometimes difficult to differentiate solid metastatic tumors from nonsolid benign lesions such as hemangiomas or cysts because each lesion shows high intensity on T2-weighted images and SPIO cannot show the hemodynamics of each lesion. Recently, heavily T1-weighted gradientecho images with SPIO obtained during a breath-hold have been reported to be useful in detecting metastatic liver tumors because SPIO shortens the T1 value of liver parenchyma, but not of tumors without Kupffer s cells [13, 14]. combination of T1- and T2- JR:181, November

2 Kumano et al. Downloaded from by on 11/27/17 from IP address Copyright RRS. For personal use only; all rights reserved weighted gradient-echo images can be expected to show different signal intensity patterns relative to the liver parenchyma in each kind of solid and nonsolid lesion because the T1 and T2 values for solid and nonsolid lesions are different and SPIO enhances hemangiomas on T1-weighted images as a blood pool agent [15 17]. The purpose of this study was to evaluate the ability of SPIO-enhanced MRI to differentiate solid metastatic tumors from nonsolid benign lesions by clarifying characteristic signal-intensity patterns of each tumor on SPIO-enhanced T2-weighted and heavily T1-weighted gradient-echo images. Materials and Methods etween October 2000 and ugust 2001, 65 patients without cirrhosis with hepatic tumors suspected on the basis of findings on enhanced CT or sonography underwent SPIO MRI for further examination to confirm the presence of tumor. Of these 65 patients, 33 (25 men and eight women) had 81 focal hepatic lesions consisting of 42 cysts, 13 hemangiomas, and 26 metastatic tumors. The scans of 33 patients were reviewed for this study. Their mean age was 65 years (range, years), and all patients had given their informed consent to be included in the study, which was conducted in accordance with the principles of the Declaration of Helsinki [18]. To exclude primary malignant tumor, such as hepatocellular carcinoma, we did not include patients with cirrhosis in this retrospective study. The maximal dimension of the lesions ranged from 5 to 35 mm (mean, 10 mm) for cysts, from 8 to 60 mm (mean, 24 mm) for hemangiomas, and from 8 to 35 mm for metastases (mean, 19 mm). Two hemangiomas in two patients were histologically confirmed by surgical resection. ll 42 of the cysts and the 11 hemangiomas were confirmed on the basis of a combination of clinical and radiologic criteria consisting of tumor markers and dynamic MRI and CT examinations. Five metastatic tumors in four patients were histologically confirmed by surgical resection. The other metastatic lesions were confirmed by using a combination of clinical and radiologic criteria, including response to transcatheter arterial chemotherapy or progression or regression in size. In addition, patients underwent follow-up CT and MRI more than 6 months later. The metastases originated from the following primary tumors: colon carcinoma (n = 16), breast carcinoma (n = 5), mesenteric leiomyosarcoma (n = 3), gastric carcinoma (n = 1), and cervical carcinoma (n = 1). ny lesion detected on MRI alone was excluded from this study because we could not confirm the lesion by our imaging criteria mentioned previously. Ferumoxides (Feridex, Eiken and Tanabe Pharmaceutical, Osaka, Japan), a SPIO preparation coated with low-molecular-weight dextran, was used. The physical, pharmacokinetic, and biologic properties of the SPIO have been reported previously [19]. This agent was supplied in 5-mL vials containing 11.2 mg of iron and 61.3 mg of mannitol per milliliter. dose of 0.56 mg (10 µmol) of iron per kilogram of body weight was diluted with 100 ml of 5% dextrose solution and infused IV at a rate of 3 ml/min through an in-line 5-µm filter. MRI was performed on either a Signa Horizon LX 1.5-T system (General Electric Medical Systems, Milwaukee, WI) (23 patients) or a Magnetom Vision 1.5-T system (Siemens, Erlangen, Germany) (10 patients). For all cases, phased array coils (Torso-array coil, General Electric Medical Systems; or CP body array coil, Siemens) were used for reception of the signal. ll MRIs were obtained in the axial plane. pproximately 60 min after the initiation of IV drip infusion of the SPIO for 30 min, heavily T1- and T2-weighted fast low-angle shot or fast spoiled gradient-echo with steady-state images were obtained. TR of 150, TE range of , and a flip angle of 90 were used for heavily T1-weighted gradient-echo images, and a TR of 150, a TE range of 10 12, and flip angle of 60 were used for T2- weighted gradient-echo images. ll images were obtained during a 17-sec breath-hold in the axial plane with a section thickness of 8 mm, an interslice gap of 2 mm, a matrix size of , a field of view of mm, and 1 excitation. TR of 150 was used to maximize anatomic coverage within the time frame of a reasonable breath-hold. reath-held images were obtained at end-expiration in all cases. No presaturation pulses were used in order to eliminate the increase in the requisite breath-hold times. Respiratory-triggered T2-weighted fast spin-echo sequences with a TR range of 3,750 8,000 (mean, 4,974), effective TE range of 65 80, echo-train length of 8, 3 signals acquired, and a matrix size of were also obtained. In this study, the signal intensity of the lesion relative to that of the surrounding liver parenchyma only on ferumoxides-enhanced MRI with the two sequences (heavily T1- and T2-weighted gradient-echo images) was evaluated qualitatively by two radiologists in consensus who were experienced in liver imaging. During the analysis, no other sequences were available. The two observers knew of the presence and location of liver lesions but were unaware of other imaging findings and the patients diagnoses. The relative signal intensity of the lesions was classified into the following three categories: high intensity relative to the surrounding liver parenchyma, isointensity, and low intensity. Only in the lesions over which there was disagreement between the two radiologists, even after discussion, were MRI values of region of interest (circle, 5 mm in diameter) of both the lesion and the surrounding liver parenchyma measured for reference. fter blinded reviewing, the signal intensity of each lesion was also observed on T2- weighted fast spin-echo images, retrospectively. The diagnostic accuracy for differentiating solid metastatic tumors from nonsolid benign lesions (cysts or hemangiomas) using the characteristic signal-intensity pattern of each lesion on the two sequences was determined. Results ll lesions showed high intensity on T2- weighted fast spin-echo images, although cysts and hemangiomas tended to show higher intensity than metastatic liver tumors. Of the 42 cysts, 40 lesions showed isointensity on the T2-weighted gradient-echo images (Fig. 1). Of these 40 lesions, 37 showed low intensity (Fig. 1) whereas the remaining three lesions showed isointensity on the heavily T1-weighted gradient-echo Fig. 1. Cyst (arrow) in 60-year-old man with preoperative examination of colon cancer using superparamagnetic iron oxide enhanced MRI (category 5)., Ferumoxides-enhanced T2-weighted gradient-echo image shows cyst as isointense to liver parenchyma because liver parenchyma decreased in signal after contrast administration and cyst has long T1 values, suppressing T2 contrast, and does not show high intensity in parameter., Ferumoxides-enhanced T1-weighted gradient-echo image shows cyst as low intensity relative to liver parenchyma because liver parenchyma increased in signal after contrast administration and cyst did not. This signal-intensity pattern of category 5 is characteristic of cyst JR:181, November 2003

3 Superparamagnetic Iron Oxide Enhanced MRI of the Liver Downloaded from by on 11/27/17 from IP address Copyright RRS. For personal use only; all rights reserved images. The three lesions were less than 5 mm in diameter. The remaining two of the 42 cysts revealed high intensity on the T2-weighted gradient-echo images and low intensity on the heavily T1-weighted gradient-echo images (Fig. 2). The two lesions were 35 mm and 18 mm in diameter and were larger than the other cysts. ll 13 hemangiomas showed high intensity on the T2-weighted gradient-echo images (Fig. 3). Eleven of the 13 lesions showed isointensity (Fig. 3), and the other two showed high intensity on the heavily T1- weighted gradient-echo images. Of the 26 metastatic lesions, 25 showed high intensity (Fig. 4) on the T2-weighted gradient-echo images. ll 25 lesions showed low intensity on the heavily T1-weighted gradient-echo images (Fig. 4). nother lesion, which was less than 10 mm in diameter in the subphrenic location of segment VIII and originating from mesenteric leiomyosarcoma, showed isointensity on the T2- weighted gradient-echo images. This lesion showed low intensity on the heavily T1- weighted gradient-echo images (Fig. 5). ll 26 lesions revealed low intensity on the heavily T1-weighted gradient-echo images. Fig. 2. Cyst (arrow) in 54-year-old woman with preoperative examination for breast cancer using superparamagnetic iron oxide enhanced MRI (category 3)., Ferumoxides-enhanced T2-weighted gradient-echo image shows large cyst, 35 mm in diameter, as high intensity because T2 value of large cyst with large amount of serous fluid and without partial volume effect is long., Ferumoxides-enhanced T1-weighted gradient-echo image depicts cyst as low intensity. This signal-intensity pattern is classified as category 3, which is considered to be metastasis; however, this cyst is atypical. Fig. 3. Hemangioma (arrow) in 62-year-old man with preoperative examination of colon cancer using superparamagnetic iron oxide enhanced MRI (category 2)., Ferumoxides-enhanced T2-weighted gradient-echo image depicts hemangioma as high intensity relative to liver parenchyma that deceased in signal after contrast administration., Ferumoxides-enhanced T1-weighted gradient-echo image shows that hemangioma is enhanced as isointense relative to liver parenchyma and cannot be seen. This signal-intensity pattern of category 2 is characteristic for hemangioma. combination of the relative signal intensity of the lesion on each sequence could be classified into the following five categories: high-intensity heavily T1-weighted gradientecho images and high-intensity T2-weighted gradient-echo images (category 1), isointense heavily T1-weighted gradient-echo images and high-intensity T2-weighted gradient-echo images (category 2), low-intensity heavily T1- weighted gradient-echo images and high-intensity T2-weighted gradient-echo images (category 3), isointense heavily T1-weighted gradient-echo images and isointense T2- weighted gradient-echo images (category 4), low-intensity heavily T1-weighted gradientecho images and isointense T2-weighted gradient-echo images (category 5). ccording to this classification, category 1 contained two hemangiomas, category 2 had 11 hemangiomas, category 3 had 25 metastatic tumors and two cysts, category 4 had three cysts, and category 5 had 37 cysts and one metastatic tumor. When a tumor with a relative signal intensity of category 1 or 2 was considered to be a hemangioma (category 3 metastatic tumors and category 4 or 5 cysts), diagnostic accuracy for characterizing such hepatic lesions was 96% (78/81). Discussion Detection of metastatic liver tumors before and after surgery for regional tumors is an important aspect of liver imaging that has implications for patient treatment and prognosis. However, metastatic liver tumors sometimes coexist with benign focal hepatic lesions like hemangiomas or cysts. Therefore, differentiation between metastatic liver tumors and benign lesions is important in reducing the number of findings of false-positive lesions. MRI with gadopentetate dimeglumine allows us to evaluate the hemodynamics of each kind of tumor and improve our capability for tumor characterization [1]. On the other hand, SPIO was reported to improve metastatic tumor detection, but not tumor characterization, although it may help to evaluate histologic grading of hepatocellular tumors with different numbers of Kupffer s cells [12]. To our knowledge, the efficacy of characterization of hepatic tumors by SPIO-enhanced MRI combined with the most recently optimized sequences, such as various gradient-recalled echo sequences, has not been evaluated in previous comparative studies. In this study, we used gradient-echo sequences with long TEs (T2-weighted gradient-echo JR:181, November

4 Downloaded from by on 11/27/17 from IP address Copyright RRS. For personal use only; all rights reserved Kumano et al. contrast. The two cysts that showed high intensity on T2-weighted gradient-echo images were relatively larger than the other cysts. We believe that the T2 value of large cysts with a large amount of serous fluid and without partial volume effect was longer than small cysts. This signal pattern was considered to be indicative of metastasis according to our category. In this study, blinded reviewers did not assess T2weighted fast spin-echo images at the time of blinded reviewing, but we can speculate that T2-weighted single-shot fast spin-echo imaging may be helpful because large serous cysts usually show markedly higher intensity, although this evaluation was not an objective of our analysis. On the heavily T1-weighted gradient-echo images, 37 cysts showed low intensity relative to the liver parenchyma enhanced by SPIO, but three cysts showed isointensity. We speculated that the reason was that these small cysts showed increased signal because of a partial volume effect. Hemangiomas showed iso- or high intensity on the heavily T1-weighted gradient-echo images with short TEs. Hemangiomas were re- images). In this sequence, T2 relaxations are strongly affected by local field inhomogeneities (intravoxel dephasing). SPIO can have a large longitudinal relaxivity value, which can reduce the T1 value more than gadopentetate dimeglumine. Therefore, we also used heavily T1-weighted gradient-echo images obtained during breath-hold because this sequence was reported to be useful in detecting metastatic liver tumors because SPIO shortens the T1 value of liver parenchyma but not that of lesions without Kupffer s cells. Each kind of solid and nonsolid lesion can be expected to show different signal-intensity patterns relative to the liver parenchyma on T1- and T2weighted gradient-echo images because T1 and T2 values of solid or nonsolid lesions are different and SPIO enhances hemangiomas on T1-weighted gradient-echo images as a blood pool agent [15 17]. Forty of 42 cysts showed isointensity on T2weighted gradient-echo images. ecause the T2-weighted gradient-echo images in this study used a relatively large flip angle of 60, T1 contrast of cysts with long T1 values suppressed T2 ported to be enhanced by SPIO on T1-weighted gradient-echo images [15 17] because SPIO is not only taken up by the macrophages and endothelial cells of hemangiomas but also accumulates in a blood pool of hemangioma because of the prolonged circulation of the contrast agent. ll 13 hemangiomas showed high intensity on T2-weighted gradient-echo images with long TEs. The reason for this was thought to be because both T1 values shortening of the lesion due to SPIO and originally long T2 values of the lesion reflected the high signal intensity of hemangiomas on the T2-weighted gradient-echo images with relatively large flip angles and long TE sequences. Twenty-five of 26 metastatic tumors showed high intensity, but the other one showed isointensity on T2-weighted gradient-echo images. This lesion was less than 10 mm in diameter. We thought that this small lesion could not be detected because of a partial volume effect. On unenhanced T2weighted fast or single-shot fast spin-echo images, cysts and hemangiomas may show a higher signal intensity than metastatic tu- Fig. 4. Metastatic tumor (arrow) in 59-year-old woman with follow-up examination of colon cancer using superparamagnetic iron oxide enhanced MRI (category 3)., Ferumoxides-enhanced T2-weighted gradient-echo image depicts metastasis as high intensity relative to liver parenchyma that decreased in signal after contrast administration., Ferumoxides-enhanced T1-weighted gradient-echo image shows metastasis as low intensity relative to liver parenchyma. This signal-intensity pattern of category 3 is characteristic for solid lesion, indicating metastatic liver tumor. Fig. 5. Metastasis in 54-year-old woman with follow-up examination of mesenteric leiomyosarcoma using superparamagnetic iron oxide enhanced MRI (category 5)., Ferumoxides-enhanced T2-weighted gradient-echo image cannot show small metastasis because it may be isointense because of partial volume effect., Ferumoxides-enhanced T1-weighted gradient-echo image shows metastasis (arrow) as low intensity relative to liver parenchyma. This signal intensity pattern is classified as category 5, which is considered to be cyst; however, this metastasis is atypical JR:181, November 2003

5 Superparamagnetic Iron Oxide Enhanced MRI of the Liver Downloaded from by on 11/27/17 from IP address Copyright RRS. For personal use only; all rights reserved mors. Therefore, as mentioned previously, SPIO-enhanced T2-weighted single-shot fast spin-echo imaging may also provide us with useful information for the differential diagnosis of hepatic lesions, although this possibility is just speculation. The T2-weighted gradient-echo images used in this study might be reflected by both T1 and T2 values because a large flip angle of 60 and a long TE of 10 were used. We used a large flip angle to increase the signalto-noise ratio of the images and a long TE to emphasize signal loss due to the susceptibility effect of SPIO. We called the sequence T2-weighted gradient-echo images in this study because we think this sequence largely reflects the T2 value (susceptibility) on SPIO-enhanced MRI. This study evaluated only contrast-enhanced images because a long time is required for obtaining both unenhanced and SPIO-enhanced MRI. However, our data show that contrast-enhanced images are sufficient for differentiating hepatic lesions. Our study did not include patients with cirrhosis and hepatocellular carcinomas because tumor contrast relative to the liver parenchyma may change due to Kupffer s cell function, which is one kind of hepatic function. Hepatocellular carcinomas show variable signal intensity relative to the liver parenchyma on T1-weighted gradient-echo images with short TEs. Moreover, some hepatocellular carcinomas, especially welldifferentiated tumor grades, have a certain number of Kupffer s cells and may uptake SPIO [12]. This uptake function by Kupffer s cells may be decreased in cirrhotic livers. Therefore, it is difficult to diagnose tumors of hepatocellular origin by comparing the signal intensity of the tumor with that of liver parenchyma on SPIO-enhanced MRI. However, patients with metastatic tumors, cysts, or hemangiomas in this study had normal liver function. SPIO was developed largely for improving detection of metastatic hepatic lesions, and we therefore believe that the fact that this study included no patients with hepatocellular carcinoma is not a shortcoming. In this study, metastatic lesions were mostly derived from colon cancer. Many kinds of metastatic tumors have different T1 or T2 values. Some of them might not be included in the diagnostic criteria used by this study. Further studies are needed. In this study, previously obtained images were reviewed by the two observers. However, only the study designer and coordinator knew how the patients were selected, and although the two observers knew of the presence and location of the liver lesions, they were unaware of other imaging findings and their corresponding diagnoses. We therefore believe there was no bias at the time of the reviewing. nother criticism of our study could be the lack of histologic proof for each lesion. However, all lesions had several confirmatory studies, such as sonography, enhanced CT, enhanced MRI, and follow-up imaging studies. Moreover, we could follow up the course of some lesions over time and in response to therapy, especially transcatheter arterial chemotherapy. The results of this study show potential in determining specific diagnoses of metastatic tumors by using SPIO-enhanced heavily T1- and T2-weighted gradient-echo images. We recommend that when evaluating metastatic liver tumors on SPIO-enhanced MRI, heavily T1- and T2-weighted gradient-echo images be obtained with our parameters to exclude hemangiomas or cysts. References 1. Vogl TJ, Hammerstingel R, Schwarz W, et al. Superparamagnetic iron oxide-enhanced versus gadolinium-enhanced MR imaging for differential diagnosis of focal liver lesions. Radiology 1996;198: Ross PR, Freeny PC, Harms SE, et al. Hepatic MR imaging with ferumoxides: a multicenter clinical trial of the safety and efficacy in the detection of focal hepatic lesions. Radiology 1995;196: Ward J, Naik KS, Guthrie J, et al. Hepatic lesion detection: comparison of MR imaging after the administration of superparamagnetic iron oxide with dual-phase CT by using alternative-free response receiver operating characteristic analysis. Radiology 1999;210: Reimer P, Jahnke N, Fiebich M, et al. Hepatic lesion detection and characterization: value of nonenhanced MR imaging, superparamagnetic iron oxide-enhanced MR imaging, and spiral CT-ROC analysis. Radiology 2000;217: ellin MR, Zaim S, uberton E, et al. Liver metastases: safety and efficacy of detection with superparamagnetic iron oxide MR imaging. Radiology 1994;193: lakeborough, Ward J, Wilson D, et al. Hepatic lesion detection at MR imaging: a comparative study with four sequences. Radiology 1997;203: Hagspiel KD, Neidl KFW, Eichenberger C, et al. Detection of liver metastases: comparisons of superparamagnetic iron oxide-enhanced and unenhanced MR imaging at 1.5 T with dynamic CT, intraoperative US, and percutaneous US. Radiology 1995;196: Seneterre E, Taourel P, ouvier Y, et al. Detection of hepatic metastases: ferumoxides-enhanced MR imaging versus unenhanced MR imaging and CT during arterial portography. Radiology 1996;200: luemke D, Paulson EK, Choti M, DeSena S, Clavien P. Detection of hepatic lesions in candidates for surgery: comparison of ferumoxides-enhanced MR imaging and dural-phase helical CT. JR 2000;175: Choi DI, Kim SH, Lim JH, et al. Preoperative detection of hepatocellular carcinoma: ferumoxides-enhanced MR imaging versus combined helical CT during arterial portography and CT hepatic arteriography. JR 2001;176: Ward J, Chen F, Guthrie J, et al. Hepatic lesion detection after superparamagnetic iron oxide enhancement: comparison of five T2-weighted sequences at 1.0 T by using alternative-free response receiver operating characteristic analysis. Radiology 2000;214: Imai Y, Murakami T, Yoshida S, et al. Superparamagnetic iron oxide-enhanced MR images of hepatocellular carcinoma: correlation with histological grading. Hepatology 2000;32: Oudkerk M, van den Heuvel G, Wielopolski P, et al. Hepatic lesions: detection with ferumoxides-enhanced T1-weighted MR imaging. Radiology 1997;203: Kim JH, Kim MJ, Suh SH, Chung JJ, Yoo HS, Lee JT. Characterization of focal hepatic lesions with ferumoxides-enhanced MR imaging: utility of T1-weighted spoiled gradient recalled echo imaging using different echo times. J Magn Reson Imaging 2002;15: Denys, rrive L, Servois V, et al. Hepatic tumors: detection and characterization at 1T MR imaging enhanced with MI-25. Radiology 1994;193: Poeckler-Schoeniger C, Koepke J, Gueckel F, Sturm J, Georgi M. MRI with superparamagnetic iron oxides: efficacy in the detection and characterization of focal hepatic lesions. Magn Reson Imaging 1999;17: Paley MR, Mergo PJ, Torres GM, Ros PR. Characterization of focal hepatic lesions with ferumoxides-enhanced T2-weighted MR imaging. JR 2000:175: World Medical ssociation. Declaration of Helsinki: ethical principles for medical research involving human subjects, revised. Edinburgh: World Medical ssociation, Weissleder R, Stark DD, Engelstad L, et al. Superparamagnetic iron oxide: pharmacokinetics and toxicity. JR 1989;152: JR:181, November