The liver is an organ in which various benign or malignant primary

Transcription

1 Diagn Interv Radiol 2007; 13:81-86 Turkish Soiety of Radiology 2007 ABDOMINAL IMAGING ORIGINAL ARTICLE Contriution of diffusion-weighted MRI to the differential diagnosis of hepati masses Özgün İlhan Demir, Funda Ouz, Özgül Sağol, Oğuz Dile PURPOSE To evaluate the diagnosti ontriution of diffusionweighted magneti resonane imaging (MRI) using apparent diffusion oeffiient (ADC) values to the haraterization of hepati masses and differentiation of enign and malignant lesions. MATERIALS AND METHODS The study inluded 30 patients that underwent upper adominal MRI examinations eause of hepati masses that were found to e 1 m in size with onventional sequenes, and were additionally evaluated with diffusion-weighted MRI. Diffusion-weighted images and ADC maps in the axial plane were otained using a 1.5 Tesla MRI devie, single shot eho-planar spin eho sequenes on 3 axes (x, y, z), and diffusion sensitive gradients with 2 different values ( = 0 and = 1000 s/mm 2 ). Mean ADC measurements were alulated among the 30 ases involving 41 hepati masses. RESULTS Of the 41 hepati masses, 24 were enign and 17 were malignant. Benign lesions inluded 6 ysts, 14 hemangiomas, 2 asesses, and 2 hydatid ysts. Malignant masses inluded 8 metastases, 4 hepatoellular arinomas, 4 holangioellular arinomas, and 1 gall ladder adenoarinoma. The highest ADC values were for ysts and hemangiomas. The mean ADC value of enign lesions was 2.57 ± 0.26 x 10-3 mm 2 /s, whereas malignant lesions had a mean ADC value of 0.86 ± mm 2 /s. The mean ADC value of enign lesions was signifiantly higher than that of malignant lesions (P < 0.01). CONCLUSION Diffusion-weighted MRI with quantitative ADC measurements an e useful in the differentiation of enign and malignant liver lesions. Key words: liver magneti resonane imaging diffusion From the Departments of Radiology (Ö.İ.D., F.O. fouz@deu. edu.tr, O.D.), and Pathology (Ö.S.), Dokuz Eylül University Shool of Mediine, İzmir, Turkey. Reeived 17 Otoer 2006; revision requested 12 April 2007; revision reeived 12 April 2007; aepted 13 April The liver is an organ in whih various enign or malignant primary or seondary masses an e deteted. Today, foal masses are diagnosed using ultrasonography (US) and/or omputed tomography (CT). Additionally, magneti resonane imaging (MRI) is preferred when further haraterization of these masses is needed. MRI has many advantages (e.g., high ontrast resolution, the aility to otain images in any plane, lak of ionizing radiation, and the safety of using partiulate ontrast media rather than those ontaining iodine) that make it a favored modality. Lesion morphology, signal intensity, and ontrast enhanement pattern are taken into onsideration when haraterizing masses with MRI; however, even if the data are evaluated together, there an still e diffiulties in the differentiation of enign and malignant lesions. Although dynami ontrast enhaned examinations have eome a routine omponent of adominal imaging, the high ost/enefit ratio and risk of ontrast media side effets remain an issue. Moreover, sometimes it is not possile to distinguish etween highly vasular metastases and hemangiomas, even using dynami examinations (1). In hepati MRI, artifats due to ardia ativity, respiration, and intestinal peristalsis an negatively affet imaging quality, espeially in T2-weighted sequenes, whih require a relatively long time to aquire, partiularly in elderly patients. Diffusion-weighted MRI, first used for the early diagnosis of stroke in neuroradiology, is a tehnique that aquires an image during a single reath-hold and does not require ontrast medium (2 4). In the past, this tehnique was limited to ranial examinations eause of its sensitivity to ardia, respiratory, and peristalti movements; however its usage has spread among other ody parts after the development of fast MRI sequenes, like eo-planar imaging (EPI). Muller et al. first reported in 1994 on diffusion-weighted MRI of normal hepati, spleni, and musular tissues, as well as on foal and diffuse hepati diseases, and otained signifiant results (5). In the years that followed, several studies on liver, kidney, and other adominal organs examined with diffusionweighted MRI were pulished (6 13). In these studies it was shown that apparent diffusion oeffiient (ADC) values of normal tissues and lesions an e measured using diffusion-weighted images, and the differenes in ADC values an e used in the differential diagnosis. The major aim of the present study was to measure the ADC values of enign and malignant foal mass lesions of the liver using diffusionweighted MRI and to determine their ontriution to differential diagnosis. Materials and methods The study inluded onsious adult patient volunteers over 18 years of age with primary or metastati hepati tumors, or non-tumoral masses 81

2 that were determined y US or CT etween Novemer 2003 and June Patients with a poor general ondition, who were unale to maintain a reathhold, or had a ontraindiation for MRI (i.e., MRI inompatile prosthesis and ardia pae-maker holders) were exluded from the study. The study protool was approved y the ethial ommittee of our university and all the patients gave informed onsent. Patients were etween 18 and 88 years old (mean age, 54.4 years). In all, 30 patients (15 males and 15 females) with a total of 41 hepati masses partiipated in the study. Simple hepati ysts (n = 6) were diagnosed with typial US and MRI findings. Hemangiomas (n = 14) were diagnosed easily with their harateristi MRI findings and typial ontrast enhanement patterns. Histopathologial evaluations were performed to diagnose pyogeni and amoei asesses following surgery. One hydatid yst was diagnosed histopathologially and the other one ased on serologi and radiologial features. Of the 8 metastati masses, 5 were enountered in patients with known primary malignany (2 reast aners, 1 lung aner, 1 renal ell arinoma, 1 Hodgkin s lymphoma) and with diagnosed metastasis, as they were disovered during routine sreening and they tended to inrease in size with time. The 3 remaining metastati liver masses were evaluated with iopsy and diagnosed as metastati adenoarinoma of unknown origin. One of the ases was diagnosed with imaging tehniques (CT and MRI) and appeared to e a gall ladder tumor with loal hepati invasion. Of the 4 lesions of primary hepatoellular tumors of the liver, one was a hepatolastoma, diagnosed histopathologially. The 3 remaining lesions were hepatoellular arinoma (HCC) ases with portal vein thromosis, of whih 2 were Tale 1. Distriution of lesions aording to size diagnosed histopathologially, and one with MRI. Among the 4 ases of holangioellular arinoma, 2 were diagnosed histopathologially and the others with MRI. The 41 masses ranged in diameter from 1 to 17 m (mean diameter, 7.4 m) (Tale 1). Routine upper adominal MRI examinations were performed in the 30 patients using a 1.5 Tesla MRI devie (Gyrosan Intera, Philips, ACS-NT, Best, The Netherlands) and a phased array oil. Routine examinations were omposed of the following sequenes: fat suppressed TSE T2-weighted (TR/ TE, 1600/70 ms; flip angle, 90 ; slie thikness, 5 mm; FOV, 375 mm); TSE heavily T2-weighted (TR/TE, 1320/ 325 ms); gradient eho in-phase and opposed-phase T1-weighted (TR/TE, 192/5 ms [in-phase], 250/7 ms [opposed-phase]; flip angle, 80 ); ontrast enhaned dynami T1-weighted images (TR/TE, 176/7 ms; flip angle, 70 ) in the axial plane. Diffusion-weighted MRI examinations were performed efore ontrast enhaned slies were otained. Diffusion-weighted sequenes (TR/TE, 4200/95 ms; flip angle, 90 ; slie thikness, 5 mm; FOV, ; reath-holding time, 50 s) in the axial plane were performed, applying gradients (in order to sensitize SE sequene to diffusion) to single-shot eho-planar sequenes in all 3 axes (x, y, z), and 2 different values ( = 0 s/mm 2 and = 1000 s/mm 2 ). The first series of the image set was omposed of eho-planar spin eho T2-weighted images ( = 0 s/mm²), the next 3 series of images were applied to the first series in x, y, and z axes (value of diffusion sensitive gradients, = 1000 s/mm 2 ), and the last series of isotropi images were alulated from the projetion of the diffusion vetors in all 3 axes. Isotropi images onsisted of images that were alulated y otaining the ue root of multipliation of signal intensities that were measured y the devie in x, y, and z axes, and images that removed axes-dependent signal differenes. ADC maps regarding isotropi images were formed automatially y the devie and all mean ADC values of the lesions were measured on those maps. A irular region of interest (ROI) 1 m in diameter was used in order to measure the lesions. In large lesions the mean value of 3 different ROI measurements on the same slie was alulated. Again, for every lesion, a mean ADC value was determined y taking the mean of ADC measurements of suessive slies. For heterogeneous lesions, measurements were performed from ontrast enhaned solid parts on onventional sequenes and post-ontrast images. The ADC value of lesions 1 m in diameter was estalished using a single ROI. Statistial analyses were performed using the Mann-Whitney U test in a omputer software (SPSS In., Chiago, Illinois, USA). Results The mean ADC value of the 24 enign lesions was 2.57 ± mm 2 /s. ADC values of enign lesions were etween 1.09 ± and 3.36 ± mm 2 /s (Tale 2). The highest ADC value was for simple ysts (Fig. 1). Among the enign lesions, pyogeni asesses had the lowest ADC value (Fig. 2). Tale 2. Mean apparent diffusion oeffiient (ADC) values aording to lesion type Lesion type Mean ADC (mm²/s) Simple yst 3.05 ± ³ Hemangioma 2.46 ± ³ Hydatid yst 2.99 ± ³ Pyogeni asess 1.09 ± ³ Size of lesions Numer of lesions Amoei asess 1.83 ± ³ 1 m m m 13 >10 m 12 Metastasis 0.79 ± ³ Hepatoellular arinoma 0.90 ± ³ Cholangioellular arinoma 0.95 ± 0.13 x 10 ³ Gall ladder tumor 0.87 ± ³ 82 June 2007 Diagnosti and Interventional Radiology Demir et al.

3 a Figure 1. a. A 59-year-old female patient with a simple hepati yst. On axial fat suppressed T2-weighted MR image (a), a hyperintense lesion onsistent with a simple yst in the right loe of the liver is seen. On diffusion-weighted MR image (), the lesion is isointense ompared to the liver. Hyperintensity on the ADC map () regarding apparent diffusion inrease (mean ADC, 3.24 ± ). a Figure 2. a. A 50-year-old male patient with a pyogeni hepati asess. On axial fat suppressed T2-weighted MR image (a), a hyperintense lesion in right loe is seen. Inrease in intensity due to diffusion restrition on diffusion-weighted MRI (). Inrease in intensity onfirming diffusion restrition on ADC map () (mean ADC, 1.09 ± ). The ADC values of the 17 malignant lesions were etween 0.54 ± 0.07 and 1.24 ± mm 2 /s, with a mean value of 0.86 ± mm 2 /s (Tale 2, Fig. 3). Among the malignant lesions, the lowest ADC value was for reast aner metastasis, while holangioellular arinoma had the highest value (Fig. 4). The differene etween the mean ADC values of enign and malignant lesions was statistially signifiant (P < 0.01). Disussion Diffusion is the term used for the randomized mirosopi movement of water moleules. Diffusion is known to e a sensitive parameter in mirosopi tissue haraterization. Currently, it is possile to determine diffusion y measuring diffusion-weighted MRI and ADC in vivo (14). Diffusion-weighted imaging an e performed after strong ipolar pulses are added to spin eho or gradient eho sequenes, y sensitizing the water in tissue to diffusion. Thus, the moility and visosity of water moleules an e evaluated, and water alane etween intraellular Volume 13 Issue 2 and extraellular ompartments an e seen (15). Diffusion-weighted MRI examinations have many tehnial restritions, suh as respiratory, ardia, or peristalti physiologi ativity, all of whih affet image quality and make evaluation, whih is very sensitive to motion, more diffiult. Consequently, prior to the development of fast MRI tehniques, diffusion-weighted imaging was limited to ranial examinations. With the development of eho-planar imaging, a fast MRI tehnique, radiologists have overome the long imaging times and related artifats of onventional tehniques, and diffusion-weighted MRI is now availale for adominal evaluations as well (5, 16). The amount of diffusion is defined using the diffusion oeffiient. Diffusion oeffiient measurement in vivo is affeted y several fators in iologial tissues. Capillary perfusion, temperature, magneti sensitivity of the tissue, and motion affets the atual diffusion; therefore, the term apparent diffusion oeffiient (ADC) is used rather than diffusion oeffiient (17). The following formula is used to alulate ADC: SI/SI 0 = exp(- ADC) (18) where SI indiates the signal intensity of the diffusion gradient () applied to the image, SI 0 is the signal intensity prior to the gradient appliation and is the value of the applied diffusion gradient. The formula an e applied as follows, when there are 2 different values: ADC = [ln(s1/s2)] /(2-1) (19) In order to alulate the diffusion gradient (), the following formula is used whih inludes the gradient appliation time (λ), power of the gradient (G), time etween gradients ( ), and gyromagneti ratio (γ): = γ 2 G 2 λ 2 ( - λ/3) In the present study, ADC measurements of enign and malignant hepati masses were signifiantly different, whih supports similar previous findings (8, 19 21). Cysts and hemangiomas had the highest ADC values, Diffusion-weighted MRI of hepati masses 83

4 a Figure 3. a. A 47-year-old female patient with metastati lung aner. On axial fat suppressed T2-weighted MR image (a), a hyperintense mass lesion with nerosis in the enter is seen. Diffusion-weighted MR image () shows inreased intensity due to diffusion restrition. The lesion is oserved as hypointense on ADC map () (mean ADC, 0.76 ± ³). a Figure 4. a. A 62-year-old male patient with holangioellular arinoma. A hyperintense lesion with a slightly heterogeneous inner struture, partiularly in the right loe, and extending to segment 4a is seen on fat suppressed T2-weighted MR image (a). Inrease in intensity due to diffusion restrition on diffusion-weighted MRI (). Inrease in intensity onfirming diffusion restrition on ADC map () (mean ADC, 0.83 ± ). while malignant masses had the lowest. The mean ADC value for ysti lesions was 3.05 ± s/mm 2, whereas for hemangiomas it was 2.46 ± s/mm 2. Overlapping values were present among these 2 groups. Two hemangiomas in our study had ADC values > s/mm 2 (3.28 ± 0.19 and 3.07 ± s/mm 2 ). All the simple ysti lesions had higher ADC values than mean ADC value of hemangiomas (Fig. 1). The lowest ADC values among the malignant masses elonged to metastases (Fig. 3). This data is similar to Taouli et al. s findings (20). Mean ADC value for HCC was 0.90 ± s/mm 2 and for holangioellular arinoma it was 0.95 ± s/mm 2 (Fig. 4). Mean ADC value for all malignant masses was 0.86 ± s/mm². The mean ADC value for the pyogeni asess was 1.09 ± s/mm 2 (Fig. 2). This low value ould e related to the dense visous ontent of the asess. Aording to a study y Chan et al. on the use of MRI for the differentiation of asesses and neroti tumors (22), the mean ADC value was signifiantly lower for hepati asesses ompared to neroti tumors and simple ysts (0.67 ± s/mm 2 ). There were no neroti or ysti lesions among the malignant tumors in our study. Thus, the pyogeni asess had a signifiantly lower ADC value ompared to simple ysts. The mean ADC value was 1.83 ± s/mm 2 in our ase of an amoei asess. Different avity ontent and visosity ould have een the reason why it was higher than the pyogeni asess. The mean ADC values of the 2 hydatid ysts were 3.03 ± 0.22 and 2.95 ± s/mm 2. Unexpetedly, those values did not reflet an inrease in visosity related to yst ontents, and were not signifiantly different from simple ysts. To the est of our knowledge, there are no diffusion MRI studies dealing with hydatid ysts in the literature. With studies inluding larger series, we think that important data will e added to the literature on the use of diffusion-weighted MRI for the differential diagnosis of hydatid ysts and simple ysts. As reported y Le Bihan et al., when the value is lowered, the diffusion weight of the sequene eomes lower, signal loss aording to diffusion dereases, and ADC value inreases (23). In a study y Ihikawa et al., values were quite low (i.e., 1.6, 16, and 55) and ADC values for adominal organs were high (19). They reported that when the value is kept low, fators like perfusion and T2 time have greater relative affet on ADC measurements. For that reason, they onluded that for adominal diffusion studies, values >400 s/mm 2 might reflet ADC measurements more aurately (19). Our study was arried out with values of 0 and 1000 s/mm 2 ; however, again, Ihikawa et al. reported that higher values ause lower quality on diffusionweighted images and make evaluation harder (19). In our study, adequate image quality ould not e otained with diffusion-weighted images eause of high values; however, that was not onsidered prolemati sine ADC 84 June 2007 Diagnosti and Interventional Radiology Demir et al.

5 map measurements were taken into aount. Namimoto et al. (8) used 2 different values ( = 30 and = 1200 s/mm 2 ) in their study and on low -value diffusion-weighted MR images (in low diffusion weighting) all masses were oserved as hyperintense, whereas on high -value images (in high diffusion weighting) signals of ysts disappeared and signals of hemangiomas oviously dereased. In ontrast, sine there is a limitation of diffusion in solid tumors, they were also oserved as hyperintense on high -value diffusionweighted images. In a study y Yamada et al. (24), atual diffusion oeffiients (D) and ADC values of hepati lesions were measured, and D values were lower than ADC values. They onluded that in vivo apillary perfusion affeted the signals of diffusion-weighted images. Only in ysti lesions that did not have vasularity, ADC and D values were equal. Yamada et al. used following formula in order to alulate the D oeffiient: ) SI/SIo = (1-f) exp (-.D) + f exp (-.D where D and D represent atual and fake diffusion oeffiients, respetively, and f indiates perfusion fration (23). Aording to this formula and the study, f and D oeffiients may e useful for the haraterization of hepati lesions (23). In the present study, measurement of atual diffusion was not our aim, eause perfusion, temperature alterations, magneti sensitivity, and motion affet diffusion measurements in iologial tissues. Therefore, ADC measurements, with the ontriution of these fators, provided signifiant results in lesion haraterization. We used 2 different values in 3 axes (x, y, z) to ahieve diffusion-weighted MR images. ADC maps were formed and ADC measurements were made using isotropi images. Taouli et al. reported that there was no differene etween measured ADC values of normal and irrhoti liver parenhyma, and foal hepati lesions in 3 axes (20). Considering this data, it has een reported that liver parenhyma and foal liver lesions, ontrary to ereral white matter and kidneys, have an isotropi diffusion pattern, thus it is needless to use multi-dimensional diffusion gradients in liver diffusion studies (20). Volume 13 Issue 2 One major limitation of our study, was the low numer of lesions and the asene of enign hepatoellular lesions (e.g., hepati adenoma, foal nodular hyperplasia), when sugroups are taken into onsideration. Hene, omparison etween solid enign and malignant masses or etween different malignant masses ould not e made. Benign hepatoellular mass lesions were first evaluated y Taouli et al. and their ADC values were found to e lower than ysts and hemangiomas, and higher than malignant masses (20). Another limitation of our study was the extremely low spatial resolution due to high value seletion, espeially in lesions <1 m in diameter, and exlusion of those ases. In reent studies, image quality has een improved with faster parallel imaging methods (e.g., sensitivity enoding = SENSE) and so EPI-related artifats have een redued (25 27). Additionally, there are puliations that report improved image quality in diffusion MRI studies with 3 Tesla MRI devies (28). The latest improvements to fusion software make it possile to superimpose diffusion-weighted MRI images onto routine MRI images, automatially or manually, overoming diffiulties in the loalization of lesions. In onlusion, the diffusion-weighted MRI sequene is a useful diagnosti tool sine it an e otained during a single reath-hold, there is no need to use ontrast media, and it an ontriute to aurate diagnosis when disrimination of enign and malignant hepati masses annot e aomplished y onventional MRI sequenes. Referenes 1. Semelka RC, Shoenut JP, Kroeker MA, et al. Foal liver disease: omparison of dynami ontrast-enhaned CT and T2-weighted fat-suppressed, FLASH, and dynami gadolinium-enhaned MR imaging at 1.5 T. Radiology 1992; 184: Warah S, Chien D, Li W, Ronthal M, Edelman RR. Fast magneti resonane diffusion-weighted imaging of aute human stroke. Neurology 1992; 42: Bak T, Hoehn-Berlage M, Kohno K, Hossmann KA. Diffusion nulear magneti resonane imaging in experimental stroke: orrelation with ereral metaolites. Stroke 1994; 25: Lutsep HL, Alers GW, DeCrespigny A, Kamat GN, Marks MP, Moseley ME. Clinial utility of diffusion-weighted magneti resonane imaging in the assessment of ishemi stroke. Ann Neurol 1997; 41: Muller MF, Prasad P, Siewert B, Nissenaum MA, Raptopoulos V, Edelman RR. Adominal diffusion mapping with use of a whole-ody eho-planar system. Radiology 1994; 190: Kim T, Murakami T, Takahashi S, Hori M, Tsuda K, Nakamura H. Diffusion-weighted single-shot ehoplanar MR imaging for liver disease. AJR Am J Roentgenol 1999; 173: Ihikawa T, Haradome H, Hahiya J, Nitatori T, Araki T. Diffusion-weighted MR imaging with a single-shot ehoplanar sequene: detetion and haraterization of foal hepati lesions. AJR Am J Roentgenol 1998; 170: Namimoto T, Yamashita Y, Sumi S, Tang Y, Takahashi M. Foal liver masses: haraterization with diffusion-weighted eho-planar MR imaging. Radiology 1997; 204: Stahlerg F, Brokstedt S, Thomsen C, Wirestam R. Single-shot diffusion-weighted eho-planar imaging of normal and irrhoti livers using a phased-array multioil. Ata Radiol 1999; 40: Amano Y, Kumazaki T, Ishihara M. Singleshot diffusion-weighted eho-planar imaging of normal and irrhoti livers using a phased-array multioil. Ata Radiol 1998; 39: Moteki T, Ishizaka H, Horikoshi H, Matsumoto M. Differentiation etween hemangiomas and hepatoellular arinomas with the apparent diffusion oeffiient alulated from turoflash MR images. J Magn Reson Imaging 1995; 5: Ries M, Jones RA, Basseau F, Moonen CT, Grenier N. Diffusion tensor MRI of the human kidney. J Magn Reson Imaging 2001; 14: Namimoto T, Yamashita Y, Mitsuzaki K, Nakayama Y, Tang Y, Takahashi M. Measurement of the apparent diffusion oeffiient in diffuse renal disease y diffusion-weighted eho-planar MR imaging. J Magn Reson Imaging 1999; 9: Le Bihan D, Turner R, Douek P, Patronas N. Diffusion MR imaging: linial appliations. AJR Am J Roentgenol 1992; 159: Hagmann P, Jonasson L, Maeder P, Thiran JP, Wedeen VJ, Meuli R. Understanding diffusion MR imaging tehniques: from salar diffusion-weighted imaging to diffusion tensor imaging and eyond. Radiographis 2006; 26: Reimer P, Saini S, Hahn PF, Brady TJ, Cohen MS. Clinial appliation of adominal ehoplanar imaging (EPI): optimization using a retrofitted EPI system. J Comput Assist Tomogr 1994; 18: Le Bihan D. Moleular diffusion nulear magneti resonane imaging. Magn Reson Q 1991; 7: Le Bihan D, Breton E, Lallemand D, Grenier P, Caanis E, Laval-Jeantet M. MR imaging of intravoxel inoherent motions: appliation to diffusion and perfusion in neurologi disorders. Radiology 1986; 161: Ihikawa T, Haradome H, Hahiya J, Nitatori T, Araki T. Diffusion-weighted MR imaging with single-shot eho-planar imaging in the upper adomen: preliminary linial experiene in 61 patients. Adom Imaging 1999; 24: Diffusion-weighted MRI of hepati masses 85

6 20. Taouli B, Vilgrain V, Dumont E, Daire JL, Fan B, Menu Y. Evaluation of liver diffusion isotropy and haraterization of foal hepati lesions with two single-shot eho-planar MR imaging sequenes: prospetive study in 66 patients. Radiology 2003; 226: Moteki T, Horikoshi H, Oya N, Aoki J, Endo K. Evaluation of hepati lesions and hepati parenhyma using diffusionweighted reordered turoflash magneti resonane images. J Magn Reson Imaging 2002; 15: Chan JH, Tsui EY, Luk SH, et al. Diffusionweighted MR imaging of the liver: distinguishing hepati asess from ysti or neroti tumor. Adom Imaging 2001; 26: Le Bihan D, Breton E, Lallemand D, Auin ML, Vignaud J, Laval-Jeantet M. Separation of diffusion and perfusion in intravoxel inoherent motion MR imaging. Radiology 1988; 168: Yamada I, Aung W, Himeno Y, Nakagawa T, Shiuya H. Diffusion oeffiients in adominal organs and hepati lesions: evaluation with intravoxel inoherent motion eho-planar MR imaging. Radiology 1999; 210: Bammer R, Keeling SL, Augustin M, et al. Improved diffusion-weighted single-shot eho-planar imaging (EPI) in stroke using sensitivity enoding (SENSE). Magn Reson Med 2001; 46: Nasu K, Kuroki Y, Nawano S, et al. Hepati metastases: diffusion-weighted sensitivityenoding versus SPIO-enhaned MR imaging. Radiology 2006; 239: Yoshikawa T, Kawamitsu H, Mithell DG, et al. ADC measurement of adominal organs and lesions using parallel imaging tehnique. AJR Am J Roentgenol 2006; 187: Hunshe S, Moseley ME, Stoeter P, Hedehus M. Diffusion-tensor MR imaging at 1.5 and 3.0 T: initial oservations. Radiology 2001; 221: June 2007 Diagnosti and Interventional Radiology Demir et al.