Diffusion-Weighted Imaging of the Liver

Diffusion-WeightedImaging of the Liver
Techniques and Applications
Sara Lewis, MDa, Hadrien Dyvorne, PhDb, Yong Cui, MDb,Bachir Taouli, MDa,b,*

KEYWORDS

� Diffusion � Liver MRI � Apparent diffusion coefficient � Liver lesion detection� Liver lesion characterization � IVIM � Cirrhosis � Echo planar imaging

KEY POINTS

� Diffusion-weighted imaging (DWI) is based on differences in the mobility of water protons in tissues.Single-shot echo planar DWI sequences are most commonly used in liver imaging.

� DWI is generally more sensitive than fast-spin-echo fat-suppressed T2-weighted imaging for liverlesion detection. The combination of DWI and contrast-enhanced T1-weighted imaging is mostsensitive for the detection of malignant liver lesions.

� Apparent diffusion coefficient (ADC) quantification can be used to characterize liver lesions ascystic/necrotic or solid. However, ADC alone is insufficient for lesion characterization.

� ADC has potential value in the evaluation of tumor treatment response, with changes in ADC pre-ceding changes in lesion size.

� ADC quantification and intravoxel incoherent motion DWI have diagnostic value in the noninvasivedetection of liver fibrosis and cirrhosis.

INTRODUCTION

Diffusion-weighted imaging (DWI) is a magneticresonance (MR) imaging technique that reportson the physical process of microscopic thermalmotion of water molecules in biologic tissues.1,2

The differences in the mobility of water protonscreate image contrast, which is influenced by theinteraction of water molecules with cellular mem-branes, macromolecules, degree of cellular den-sity, and the size of the extracellular extravascularspace.3 DWI is increasingly being used in liverMR imaging given the recent technologic ad-vances and improvements in image quality,including the introduction of echo planar imaging

Funded by National Institutes of Health, United States, ga Department of Radiology, Icahn School of Medicine atYork, NY 10029, USA; b Translational and Molecular ImSinai, One Gustave Levy Place, Box 1234, New York, NY* Corresponding author. Department of Radiology, IcahLevy Place, Box 1234, New York, NY 10029.E-mail address: [email protected]

Magn Reson Imaging Clin N Am 22 (2014) 373–395http://dx.doi.org/10.1016/j.mric.2014.04.0091064-9689/14/$ – see front matter � 2014 Elsevier Inc. All

(EPI), parallel imaging, multichannel coils, andhigh amplitude gradients.

Accurate lesion detection and characterizationare essential for treatment planning for patientswith primary or secondary liver tumors, especiallyin selecting patients who may undergo liver resec-tion or locoregional or systemic therapies.4,5 DWIcanbeused for focal liver lesiondetectionandchar-acterization, for the assessment of tumor response,and for the evaluation of diffuse liver disease.3 Thissequence is easily incorporated into routine clinicalprotocols, especially given that DWI is a noncon-trast technique andmaybeperformedeither beforeor after contrast administration. DWI can be ac-quired rapidly within a breath hold, and it provides

rant number 1R01DK087877.Mount Sinai, One Gustave Levy Place, Box 1234, Newaging Institute, Icahn School of Medicine at Mount10029, USAn School of Medicine at Mount Sinai, One Gustave

rights reserved. mri.th

eclinics.com

mailto:[email protected]

http://crossmark.crossref.org/dialog/?doi=10.1016/j.mric.2014.04.009&domain=pdf

http://dx.doi.org/10.1016/j.mric.2014.04.009

http://mri.theclinics.com

Lewis et al374

both qualitative and quantitative information as anadjunct to conventional sequences. Quantificationof apparent diffusion coefficient (ADC) has enableddifferentiation of solid cellular lesions from cystic ornecrotic lesions.6 ADC quantification has alsodemonstrated promising results for predictingtumor response to therapy.7,8 The purpose of thisarticle is to review the basic principles of DWI,discuss protocol optimization, summarize the per-formance of DWI compared with conventional se-quences, discuss the role of DWI in the evaluationof treatment response, and examine the potentialrole of DWI in the assessment of liver fibrosis.

DWI TECHNIQUEConcepts

Principles of molecular diffusionDiffusion describes random molecular motionoccurring in tissues as a result of thermally acti-vated translations of atoms and molecules. The re-sulting motion can be described as a stochasticprocess, with a gaussian probability distributiongiven as follows:

Pðr; tÞwe�r2=4Dt 1

where D is the diffusion coefficient and r is the dis-tance traveled by the diffusing molecule duringtime (t)D depends on the species undergoing diffusion

and on the medium in which diffusion occurs. Forwater self-diffusion at 37�C,D5 3.0mm2/s.9 Largermolecules tend toundergoslowerdiffusion resultingin lowerD, as is also the case of diffusion in fluids ofhigher viscosity. The gaussian distribution in Equa-tion 1 describes free diffusion of molecules in an in-finitemedium.An important variationof this conceptis restricted diffusion whereby molecular motionis restrained within hard boundaries. Restricteddiffusion, which is common in tissues, generallyleads to lower D and nongaussian distributions.

DWI physicsNuclear MR offers a novel way to measure diffu-sion, via the application of magnetic field gradi-ents. If such gradients are deployed in pulses ofopposed polarity (using a gradient or spin-echosequence), moving spins undergo dephasing whilestatic spins show a null phase at the echo time. Byconsidering the effect of a stochastic diffusion pro-cess on the magnetic signal, one can derive thesignal attenuation resulting from the applicationof pulsed gradients in a spin-echo experiment10:

S5S0e�bD

with S0 the signal in the absence of gradient and ba function of the applied gradient

b5 ðgGdÞ2�D� d

3

�

with g the spin gyromagnetic ratio, G and d thegradient strength and length, and D the time sepa-rating the gradient pair.The factor b, called b value, determines the

strength of the diffusion weighting. The diffusiongradients can be inserted in an imaging experi-ment as a preparation module in order to provideadditional contrast to the MR imaging signal andto estimate molecular diffusion in different organsand tissues.

Quantification of diffusion properties in tissuesIn its simplest form, the diffusion experiment in-volves acquiring 2 sets of images, one at low orzero b value and another at high b value, to derivea voxelwise diffusion coefficient. Because multipletissues and compartments may be present in asingle voxel, the derived coefficient is referred toas an ADC that reflects a sum of exponential de-cays rather than a single, pure diffusion constant.In addition to the ADC, which assumes monoex-ponential decay of diffusion signal, there aremodels that account for the more complex proper-ties of tissues:

� Flowing blood contributes to diffusion signaland leads to measureable effects on the diffu-sion decay. The intravoxel incoherent motion(IVIM) approach11 integrates these effects ina biexponential model whereby a faster de-caying exponential, reflecting perfusion ef-fects at low b values, can be separated fromslower exponential decay reflecting true waterdiffusion. This approach is especially appro-priate in highly perfused organs, such as theliver.12–18

� Because of restricted diffusion in tissues, thegaussian model of Equation 1 is not validanymore, and diffusion weighting has a lesstrivial form. A successful approach has beendiffusion kurtosis imaging (DKI) that evaluatesrestricted diffusion by analyzing the nongaus-sian diffusion distribution, using an additionalconstant derived from acquisitions at a higherb value.19 Recently, DKI has been investi-gated in liver explant studies, whereby resultshave been correlated with hepatocellular car-cinoma (HCC) tumor cellularity.20

� Biologic tissues can be anisotropic and,therefore, so can diffusion decay. The ADCmeasured using different diffusion gradient di-rections may differ in an anisotropic sample;therefore, diffusion tensor imaging21 (DTI)has been proposed in order to extractdirection-specific information using at least 6

Diffusion-Weighted Imaging of the Liver 375

different diffusion gradient directions. In theliver, acquisition protocols typically involveacquiring the diffusion trace, which is anaverage of all 3 principal diffusion directions.DTI has been investigated in abdominal or-gans,22–25 but there are limited data on aniso-tropic properties in the liver.26–28

Quality controlThere are sources of variability for diffusion quan-tification across different platforms or centers.First, hardware variations exist between differentplatforms, leading to different estimates for ADC.Sasaki and colleagues29 reported up to 9% vari-ability in brain ADC in a multicenter study in 12healthy subjects. Chenevert and colleagues30

measured the ADC of a temperature-controlledice-water sample in several platforms and re-ported 5% variability across 3 vendors at 1.5 Tand 3.0 T. Another source for variability is thechoice of b values used to derive the ADC. Previ-ous studies in liver DWI have used a b-value rangeof 0 to 1000 s/mm2. Because of the confoundingeffects of perfusion, computing the ADC using alower range of b values (eg, 0–400) may lead tohigher values than using an extended range of bvalues (eg, 0–1000) and to different accuracy ofthe technique for disease detection.31 Accordingto Xing and colleagues,32 the ADC can be reliablyestimated using 2 b values chosen outside ofthe perfusion range, with a magnitude such thatbhigh� bloww 1.1/D. In order to measure perfusioneffects using the IVIM model, more b values areneeded; the b-value distributions must be appro-priately chosen to sample the fast pseudodiffusion(b<100) and slower true diffusion (100<b<1000)decays.33,34

Reproducibility and repeatabilityReproducibility and repeatability are important con-cepts for quantitative diffusion evaluation, particu-larly for longitudinal and multicenter studies. In theliver, Braithwaite and colleagues35 reported amean coefficient of variation (CV) (which equalsthe standard deviation divided by the mean) of14%. In hepatic tumors,36 similar results were re-ported for short-term reproducibility (95% limits ofagreement: 30%); ADC reproducibility was foundto depend on the lesion location (lower reproduc-ibility in the left lobe) and size (lower reproducibilityin smaller lesions). The reproducibility of liver IVIMwas also reported,13,14,17 showing higher reproduc-ibility for D and ADC (CV<20%) compared withperfusion parameters perfusion fraction (f) andpseudodiffusion coefficient (D*) (CV>20%). IVIMreproducibility was further investigated in lesionsin a recent study showing lower reproducibility of

IVIM parameters in liver metastases13 and HCC37

compared with liver parenchyma. These findingsmay be caused by higher variability of physiologicprocesses in HCC, or to limitations of DWI for mea-surement of finite size lesions (lesions DWI signal ismore sensitive to motion artifacts and to limitedresolution).

Diffusion Acquisition

Imaging strategyThe most widely used strategy for DWI is EPI,which allows acquisition of a full slice in a singleshot. A typical protocol involves using fat-saturated single shot diffusion weighted EPIplayed in an interleaved multislice fashion to allowvolume coverage. However, the EPI readout isalso subject to ghosting and susceptibility arti-facts.38 Some alternative techniques to EPIinclude the following:

� Segmented EPI39 is performed by distributingthe EPI acquisition over repeated cycles. Thisway, susceptibility artifacts are reduced bydecreasing the EPI readout length. Comparedwith single-shot EPI, segmented EPI is moresensitive to the motion occurring betweenconsecutive shots, leading to signal inconsis-tency and artifacts.

� Turbo spin echo40 and steady-state free-pre-cession41 (SSFP) may offer reduced B0 andsusceptibility artifacts compared with EPIreadouts. However, this usually leads toincreased echo trains and to higher-powerdeposition of the radiofrequency pulse train.

� Periodically Rotated Overlapping ParallELLines with Enhanced Reconstruction (PRO-PELLER)42 is a modified segmented EPI tech-nique whereby successive segments areacquired in a radial fashion. Compared withsegmented EPI, PROPELLER may be moreimmune to motion artifacts, with options forrobust motion correction.43

Control of physiologic motionPhysiologic motion is inherent to any liver imagingprotocol, with breathing and cardiac motion result-ing in subject- and acquisition-dependent imagingartifacts. Single-shot EPI is robust to motionbecause its acquisition time is faster than physio-logic processes. However, in a typical DWI acqui-sition with multiple b values and signal averaging,residual fluctuations exist between successivelyacquired EPI images. Breath hold (BH), freebreathing (FB), or respiratory triggering (RT) tech-nique may be used. BH results in the shortest im-age acquisition time, with a limitation in thenumber of b values that can be used. FB protocols

Lewis et al376

typically result in blurred images. RT acquisition(using navigator echoes or bellows) may improveDWI data quality14,36,44,45 at the cost of increasedimaging time. Cardiac motion artifacts usuallyappear in the left liver lobe as signal loss at ahigh b value as a result of strong dephasing ofthe coherently moving spins under the influenceof diffusion gradients. This artifact can be over-come by performing DWI acquisitions at diastoleusing an electrocardiogram or pulse trigger,46 butthis approach results in significantly increasedscan time. Another promising approach to mitigatesignal loss caused by cardiac motion is to usemotion-compensated diffusion gradients47,48 tocancel the dephasing of coherently moving tissueswhile maintaining diffusion weighting.

Parallel imagingUsing multiple receiver radiofrequency coils, k-space acquisition can be accelerated by skippinglines and reconstructing the image using well-established algorithms.49–51 Applied to single-shot EPI, parallel imaging allows for a shorter EPItrain, hence, lower susceptibility artifacts andachievable higher resolution.52 In order to avoid re-sidual artifacts and severe signal loss, accelerationfactors are typically limited to 3, depending on thenumber of coil elements. Another more recentapplication of parallel imaging is multiband imag-ing that allows for simultaneous excitation andacquisition of multiple slices.53,54 Multiband imag-ing leads to increased slice coverage comparedwith regular DWI acquisition, with minimal signaldegradation.

Effect of magnetic field strength: 1.5 T versus3.0 TBecause of the strong signal attenuation imposedby diffusion gradients, DWI suffers from low signallevels, as quantified by the signal-to-noise ratio(SNR). Low SNR, especially at high b values, canbe overcome by signal averaging or by increasedstatic magnetic field. In this regard, a magneticfield strength of 3.0 T might result in up to a2-fold SNR increase compared with 1.5 T (4-foldreduction in signal averaging). However, Rozenk-rantz and colleagues55 reported worse imagequality for abdominal DWI at 3.0 T when comparedwith 1.5 T. This study was using a first-generation3 T system. Physical limitations at 3 T may explainthe moderate advantages of 3 T DWI. First, higherEPI distortions may arise from stronger B0 andsusceptibility variations. Second, the radiofre-quency wavelength at 3 T, of the same order asthe human body, results in inhomogeneity of theexcitation field B1. This inhomogeneity, in turn,may lead to large signal variations across the liver

and ghosting artifacts caused by incomplete fatsaturation. In addition to image quality degrada-tion, previous reports have shown significant dif-ferences in the ADC of abdominal organs at1.5 T and 3.0 T.56 Fig. 1 illustrates DWI artifactsat 3.0 T compared with 1.5 T.

LIVER APPLICATIONSLesion Detection

Detection of malignant liver lesions is predicated onthe belief that tumors have increased cellularitycompared with background liver parenchyma and,consequentially, higher signal intensity on DWI.DWI is increasingly used in oncologic imaging forthe detection of liver metastases, HCC, and cholan-giocarcinoma, the application of which is discussedindetail later. Lowb-valuediffusion imagesusingSSEPI (ie, b<100) aremost valuable for lesiondetectiongiven the suppression of vascular flow, producingblack-blood images, resulting in improved conspi-cuity for lesions located near vessels.57,58 The sup-pression of intrahepatic vascular flow also avoidsconfusing a small lesion with a vessel on the T2-weighted imaging (T2WI) sequence.57 Comparedwithhighb-value images (ie, b1000), lowb-value im-ages have relatively higher contrast-to-noise ratiosand SNRs.58–61 Low b-value images are also lessaffected by signal loss caused by cardiac motionand geometric distortions induced by eddy cur-rents.62 Prior work has comparedDWI lesion detec-tion rates with conventional sequences, includingT2WIs and postcontrast-enhanced T1-weightedimaging (CE T1WI) using extracellular gadolinium-based contrast agents (GBCAs), liver-specificGBCAs and non-GBCAs liver-specific agents, asdiscussed later. Studies comparingDWIwith super-paramagnetic iron oxide or mangafodipir trisodiumare not discussed in this review as these agentsare either no longer available or not widely used.

Liver metastasesThe accurate diagnosis of liver metastases is crit-ical for treatment planning and for identifying pa-tients suitable for liver resection, locoregionaltherapy or systemic chemotherapy.63–65 MR imag-ing provides valuable and accurate informationregarding lesion size, segmental location, and dis-tribution of liver metastases.4,66 DWI has outper-formed T2WI in several studies for liver lesiondetection (including metastases), with DWI detec-tion rates as high as 71.0% to 87.7% versus55.6% to 70.1% for T2WI.57,67–71 A current chal-lenge remains the ability to detect small lesions,especially metastases less than 2 cm in size. Ina prospective study of 129 lesions including40 metastases, Coenegrachts and colleagues57

Fig. 1. Example of increased single shot (SS) EPI diffusion artifacts at 3.0 T (A, B) compared with 1.5 T (C, D) in thesame patients imaged with both systems (b400 images are shown). Image distortion on coronal SS EPI imagecaused by higher B0 inhomogeneity (A) is observed in a 62-year-old male patient with hepatitis C virus andfibrosis. Signal drop caused by B1 field inhomogeneity at 3.0 T (arrow on B) is observed in a 24-year-old femalehealthy subject, compared with 1.5 T (D).


demonstrated that 100% of metastases less than1 cm (40 of 40) were detected on low b-valueDWI (b 5 20). Similarly, Parikh and colleagues71

showed an excellent metastatic lesion detectionrate of 78.5% (57 of 72) for DWI compared with45.8% (33 of 72) for T2WI for lesions 1 to 2 cm insize. In a retrospective series of 118 hepatic me-tastases, Bruegel and colleagues68 reported DWIdetection rates of 88% to 91% versus 45% to62% for T2WI, a result that was also impressivefor small lesions less than 1 cm, with a DWI detec-tion rate of 85% versus 26% to 44% for T2WIsequences. The superior detection rates of opti-mized RT DWI may be explained by improved im-age quality compared with T2WI with a reductionin eddy current–induced artifacts and image blur-ring.6 Furthermore, diffusion images also haveimproved background fat saturation, suppressionof intrahepatic vascular flow, and liver to lesioncontrast compared with T2WI.58,67

Routine liver MR imaging protocols typicallyinclude postcontrast T1WI for the detection of livermetastases. Both extracellular and combinedextracellular-hepatocyte–specific GBCAs, suchas gadoxetic acid, are widely used. DWI is notsignificantly affected by gadoxetic acid injection;in order to minimize the table time, DWI may beobtained between the dynamic and hepatobiliary(HBP) phases.72 Results comparing DWI with

dynamic CE T1WI and/or HBP for the detectionof liver metastases have been mixed, with moststudies evaluating hypovascular metastases.Overall, DWI performs reasonably well comparedwith CE-MR imaging, with equivocal or slightlylower detection rates reported in most series. Arecent meta-analysis by Wu and colleagues62

examined the overall diagnostic value of DWI indetecting liver metastases and to compared itsperformance with CE-MR imaging. The authorsexamined 11 studies (9 retrospective, 2 pro-spective) with 537 patients using either avariety of GBCAs or liver-specific agents for com-parison with DWI. These results show no majordifference in the performance of DWI comparedwith either extracellular GBCAs or combinedGBCAs.68,70,73–81 DWI used in combination withCE-MR imaging or HBP imaging demonstrateshigher sensitivity compared with DWI alone, withpooled sensitivities across multiple studies of97% for the combined set versus 86% for DWIalone.62 The sensitivity of DWI was statisticallyequivalent when compared with CE-MR imagingin several recent studies (66.3%–84.4% vs76.0%–79.7%, respectively).70,75,79 Subgroupanalysis in several of these studies also showedthat DWI is highly accurate in detecting small le-sions. For example, Lowenthal and colleagues80

demonstrated 92% sensitivity of DWI for detecting

Lewis et al378

lesions less than 1 cm. In a recent retrospectivestudy of 144 pathologically proven metastases,the investigators noted that assessment of DWIwas very important in identifying small lesions,which may be missed or misinterpreted as peri-pheral vessels or imaging artifacts on HBP.82

Per-lesion analysis in the study by Hardie andcolleagues75 revealed a trend toward an improvedDWI detection rate for the less experiencedobserver for both lesions less than 1 cm andgreater than 1 cm, without reaching statistical sig-nificance. It was thought that this finding wascaused by the simplicity of the interpretation ofDWI (Figs. 2 and 3). Recent publications haveevaluated DWI for the detection of hypervascularliver metastases. DWI outperformed CE-MR imag-ing in a retrospective series of 41 patients with 162hypervascular neuroendocrine liver metastases,with sensitivities of 71.6% to 71.0% versus47.5% to 48.1% (P 5 .001).69 Further work isnecessary to compare the role of DWI for hyper-vascular and hypovascular liver metastases.

In summary, DWI is highly sensitive for thedetection of liver metastases, especially forsmall lesions less than 2 cm. Combined DWIand CE T1WI yields the best performance forlesion detection compared with each sequencealone. In addition, DWI is a reasonable alterna-tive to CE MR imaging in patients who cannotreceive GBCAs.

HCCHCC is the most common primary liver cancer andthe third most common cause of cancer-relateddeath in the United States.83 Early diagnosis isessential given the improved prognosis of earlystage HCC and advances in surgical and locore-gional therapies (LRT).84 The diagnosis of HCC ismade at CE-computed tomography (CT) or CE-MR imaging for lesions greater than 1 cm thatdemonstrate arterial hyperenhancement and por-tal venous/late venous phase washout.5 Of note,the sensitivity for detecting small HCC less than

Fig. 2. A 51-year-old man with pancreatic cancer and liver50 and 800 s/mm2 (A, B), ADC map (C), fast spin echo T2Wafter injection of extracellular contrast medium (E). Both limor with secondary pancreatic duct distention (dashedhyperintense on high b-value image (B). ADC value of the

2 cm is considerably lower using CE-CT or CE-MR imaging.85,86 Several researchers have investi-gated DWI as a tool to improve noninvasive HCCdetection. DWI is moderately sensitive forthe detection of HCC, with detection rates rangingfrom 45.0% to 91.2%.71,87–89 The wide range ofreported sensitivities likely reflects differences instudy technique and selection bias; thus, compar-isons of reported results are somewhat chal-lenging. Studies with very high reportedsensitivities included either larger HCCs or caseswith a high pretest probability of HCC. Overall,DWI was better for HCC detection using lowb-value DWI (b0, b50) compared with conventionalBH T2WI sequences (80.5% vs 54.0%, P<.001)and especially for small HCCs less than 2 cm(79% vs 46%, P<.001).71 An excellent detectionrate of HCCs less than 2 cm of 91.2% was shownin a series of 58 HCCs using high b-value DWIcompared with postcontrast and T2WI.88 Overall,DWI has shown mixed results compared withCE-MR imaging. In the study by Piana and col-leagues,85 there was better sensitivity for HCCdetection using DWI compared with CE T1WI in91 patients with 109 HCCs greater than 1 cm insize (72.5%–81.7% vs 59.6%–59.6%). In contrast,CE-MR imaging was superior for HCC detection in52 patients with 72 HCCs in the only publishedliver explant DWI correlation study (87.9% vs75.8%, P 5 .02).90 In another study, DWI and ga-doxetic acid images showed equivalent pooledsensitivity for HCC detection (87.1% vs 91.1%,P>.05) in 130 patients with 179 surgicallyconfirmed HCCs in a separate series.91 Fig. 4demonstrates the classic imaging characteristicsof HCC with DWI and CE-MR imaging. Finally,published data have shown that the combinationof DWI, CE-MR imaging, and/or HBP imagesyields the best sensitivity for HCC detectioncompared with either sequence alone.90,91 Sensi-tivities as high as 97.9% may be achieved usingthis the combination of sequences.92 The addedbenefit of the combination of sequences is thatthe DWI and CE MR imaging provide complemen-tary information: DWI is particularly useful in

metastases. Axial RT SS EPI DWI at 3 Twith b values ofI (D), and axial late venous phase gradient echo T1WIver metastases (arrows) and the primary pancreatic tu-arrow) are identified on b50 images (A) and remainlargest confluent metastasis was 0.97 � 10�3 mm2/s.

Fig. 3. An 80-year-oldmanwith diffusemetastaticmelanoma. Axial T1 in-phase (A) and opposed-phase images (B),axial RT SS EPI DWI with b value of 1000 s/mm2 (C), and ADC map (D). Innumerable non–fat-containing T1 hyper-intense lesions are present throughout the liver, the largest confluent lesion measuring 5.2 cm in the caudate lobe(arrow). Lesions demonstrate diffusion restriction with low signal on ADC map (ADC value 0.87 � 10�3 mm2/s).


identifying lesions near vessels, and CE MR imag-ing is useful for identifying subcapsular lesions orlesions in the left lobe. DWI may also be very help-ful in detecting infiltrative HCC, which often doesnot demonstrate typical arterial hypervascularityand a delayed washout appearance (Fig. 5).93

There are several possible explanations for thelower DWI detection rate for HCC. A recognizedlimitation of DWI sequences for HCC detection isthat not all HCCs are hyperintense. Approximately8.8% to 38.9% of HCC are either isointense orhypointense on DWI sequences, and smaller le-sions (<2 cm) are frequently inconspicuous(Fig. 6).85,94,95 Well-differentiated HCCs may be

Fig. 4. A 66-year-old man with HCC. Axial RT SS EPI DWI amap (C), CE (using extracellular contrast medium) T1WI onleft hepatic lobe HCC (arrow), which is hyperintense on10�3 mm2/s). The lesion demonstrates typical hypervasculaand pseudocapsular enhancement (E).

more frequently isointense than HCCs with highertumor grade; however, visual analysis of DW im-ages alone is not reliable for this distinction.94,96

The conspicuity and signal intensity of an HCC ina cirrhotic liver ultimately depend on a complexinteraction between biologic and technical factors.HCCs are also more difficult to identify in the back-ground of cirrhosis as the distorted liver paren-chyma demonstrates diffusion restriction inrelation with background fibrosis.97,98 A retrospec-tive study by Kim and colleagues99 showed a trendtoward decreased sensitivity for HCC detectionwith increasing severity of liver cirrhosis for 2 ob-servers (Child-Pugh A 93.9%–95.6%, Child-Pugh

t 1.5 T with b values of 50 and 1000 s/mm2 (A, B), ADCarterial phase (D), and late venous phase (E). There is ab50 (A) and b1000 (B) images with low ADC (1.31 �r hyperenhancement (D) with delayed-phase washout

Fig. 5. A 59-year-old woman with infiltrative HCC invading the portal vein. Axial FB SS EPI DWI with b values of 50and 1000 s/mm2 (A, B) ADC map (C), fat-suppressed fast spin echo T2WI (D), and CE (using extracellular contrastmedium) T1WI at portal venous phase (E). Tumor thrombus (arrows) demonstrates T2 hyperintensity andrestricted diffusion on both b50 (A) and b1000 (B) images with low corresponding ADC value 1.45 �10�3 mm2/s. The extent of tumor is better delineated on the DWI and T2WIs (arrows, A–D) compared with post-contrast image (E).

Lewis et al380

B 77.1%–83.0%, Child-Pugh C 60.0%–60.6%).Based on the data presented and the establishedbenefits and limitations of each individual se-quences, the optimal MR imaging protocol for de-tecting HCCs in the cirrhotic liver should includethe combination of DWI and CE/HBP MR imaging,thus yielding a maximized diagnostic accuracy.

In summary, DWI demonstrates moderate sensi-tivity for HCC detection. DWI should be used incombination with conventional MR imaging se-quences, including postcontrast sequences, forHCC detection. Currently, DWI in isolation isinsufficient for detecting HCCs in cirrhosis.

CholangiocarcinomaIntrahepatic cholangiocarcinoma (ICC) is the sec-ond most common primary malignant tumor ofthe liver and is increasing in incidence.100 Similar

Fig. 6. A 63-year-old man with hepatitis C cirrhosis and HClayed HBP phase (C), axial FB SS EPI DWI with b values of 50of gadoxetic acid. A 2.4-cm left hepatic lobe lesion (arrowdelayed pseudocapsule enhancement and demonstrates hyis isointense on both low and high b-value DWI images anwell-differentiated HCC was found at histopathology.

to HCC, ICC may arise in the context of hepatitisC and chronic liver disease.101 There are 3 typesof ICC classified based onmorphology and growthpattern, including mass forming, periductal infil-trating type, and intraductal type.102 The massforming is the most common type and must bedifferentiated from HCC given the differences intreatment and prognosis. Recent reports haveshown that small ICCs (<3 cm) may have a similarimaging appearance and enhancement pattern asHCCs.103,104 There is a paucity of data evaluatingDWI imaging features and detection rates forICC. Very few ICCs were included in the previouswork comparing DWI with T2 and CE-MR imaging.A recent study by Park and colleagues105 de-scribes a target appearance of ICC on DWI: a pe-ripheral hyperintense rim with a central area ofhypointensity on high b-value images. The centralareas of hypointensity corresponded to fibrosis onhistopathology. This finding was noted in 75.0% of

C. Axial arterial phase (A), portal venous phase (B), de-(D) and 500 s/mm2 (E) and ADC map (F) after injection) is hypervascular with portal venous-phase washout/perintensity on delayed HBP phase images. The lesiond ADC map, with ADC value of 1.49 � 10�3 mm2/s. A


ICCs (24 of 32) versus 3.1% of HCC (1 of 32) andmay be a reliable imaging feature to distinguish asmall ICC from HCC. Further work is necessaryto evaluate the role of DWI in the detection ofICC given its growing incidence and relativelypoor prognosis.

In summary, there are limited data regardingthe sensitivity of DWI for ICC detection. DWImay be of value in detecting ICC and shows acharacteristic morphologic target appearance,which should be verified in further studies.

In summary, the assessment of qualitative DWimages allow for the characterization of liver le-sions as benign or malignant with a high degreeof accuracy. DWI may play an added role to con-ventional sequences in characterizing indeter-minate lesions in the cirrhotic liver.

Lesion Characterization

Qualitative assessmentThe noninvasive characterization of liver lesionswithout using GBCAs is increasingly attractive inthe era of nephrogenic systemic fibrosis.106 Visualassessment of liver lesion morphology, signalintensity, and changes in signal intensity atincreasing b values on DWI is useful for lesioncharacterization.6 The changes in signal intensityobserved on DWI over a range of b values dependon true lesion water diffusivity, vascular microper-fusion, and T2 relaxation time. It is essential toseparate true diffusion restriction from T2 shinethrough. T2 shine through is the phenomenon pre-sent in lesions with long T2 relaxation times, suchas cysts or gallbladder content, causing persistenthyperintensity on high b-value images and corre-sponding high ADC values; thus, the high signalobserved is caused by the long T2 of the lesioncontents (fluid).3 The observed behavior of a lesionon qualitative images must, therefore, be inter-preted along with the corresponding ADC map,which is the graphical representation of the ratioof the DW signal intensities.107 Mean ADC values(square millimeter per second) are obtaineddirectly from the ADC map by placing a region ofinterest (ROI). ADC maps are calculated with alinear regression analysis of the function. TheROI used must be large enough for sufficientSNR and to avoid the effects of volume averaging.

Fig. 7. A 58-year-old woman with metastatic colon cancerat 1.5 Twith b values of 0, 50, 500, and 800 s/mm2 (A–D) animages (B) with excellent suppression of intrahepatic vashyperintense on high b value with low ADC (1.1 � 10�3

drop on intermediate b-value images and is not seen on

Previous work suggested that it is possible tocharacterize liver lesions as benign ormalignant us-ing DWI sequences.71,106,108–111 Liver lesions areconsidered benign when demonstrating hyperin-tensity on low b-value DWI, attenuation of signalon higher b-value DWI (b>500), and a correspond-ing high ADC value. Lesions are considered malig-nant when remaining hyperintense on higherb-value DW with a low corresponding ADC value(Fig. 7).71,108 Performing visual assessment oflesion characterization has been shown to be highlyaccurate.71,108,112 In a study of 185 focal liver le-sions including 76 metastases and 11 HCCs, Hol-zapfel and colleagues113 demonstrated accuracyof 93% for characterizing small lesions less than1 cm. However, most of the benign lesions werecysts or hemangiomas. Of note, limitations of thesereports include the relatively small sample sizes andlimited types of liver lesions evaluated.106 DWI mayprovide addedbenefit to conventional sequences inthe characterization of indeterminate lesions in thecirrhotic liver. For example, distinguishing well-differentiatedHCCfrombenigncirrhotic nodules re-mains challenging in many cases. In a recent studywith histopathologic correlation, high b-value DWIshowed higher accuracy (79%) than did hypointen-sity on HBP for distinguishing well-differentiatedHCCs from benign cirrhotic nodules.114 DWI wasalso useful in distinguishing hypervascular HCCfromhypervascularpseudolesions,asnopseudole-sionswere hyperintense at DWI in a series byMoto-sugi and colleagues.115 In addition, indeterminatehypovascular lesions with diffusion restriction oftengo on to develop hypervascularity and overt imag-ing features of HCC; thus, DWImaybe a strongpre-dictor for progression to hypervascular HCC.116

and hepatic cyst. Axial respiratory-triggered SS EPI DWId ADC map (E). Both lesions are easily detected on b50cular flow. The liver metastasis (solid arrow) remainsmm2/s). The cyst (dashed arrow) shows strong signalhigh b value with high ADC (3.2 � 10�3 mm2/s).

In summary, ADC quantification can be helpfulin separating liver cysts and hemangiomasfrom malignant liver lesions. However, DWIwith ADC quantification cannot reliably makethe distinction between solid benign andmalig-nant lesions and between different malignantlesions.

Lewis et al382

Quantitative assessmentThe ADC has been investigated as a tool to char-acterize liver lesions. As described previously inthis text, the ADC is calculated from the diffusionsequence using multiple (at least 2) b values.ADC quantification requires minimum acceptableSNR at higher b values and a minimum lesionsize of 1.5 to 2.0 times the in-plane resolution inorder to avoid partial volume averaging.6 ADC im-ages must be interpreted in conjunction with qual-itative DW images and other sequences. Extensiveprior work has shown that benign lesions generallyhave a significantly higher ADC value comparedwith malignant lesions.71,108,110,111,117–119 Cystsand hemangiomas have statistically higher meanADC values compared with all solid liver lesions(benign and malignant).106 In a large study of 542lesions in 382 patients, the mean ADC values ofspecific liver lesions were as follows (measuredas � 10�3 mm2/s): cysts (3.40 � 0.48), hemangi-omas (2.26 � 0.70), focal nodular hyperplasias(FNHs) (1.79 � 0.39), hepatocellular adenomas(1.49 � 0.39), metastases (1.50 � 0.65), andHCC (1.54 � 0.44).106 ADC cutoff values be-tween 1.4 and 1.6 � 10�3 mm2/s have been re-ported in the literature for diagnosing malignantliver lesions, with reported sensitivities of 57.1%to 100% and specificities of 77% to 100%(Fig. 8).71,106,108,110,118 Of note, these ADC cutoffvalues vary depending on the patient populationstudied and DWI acquisition parameters. There issignificant overlap in the ADC measurement be-tween solid benign and malignant lesions, thus,making this distinction with DWI challenging.There is no statistically significant difference in re-ported ADC values for FNHs, hepatocellular ade-nomas, metastases, and HCC.106,120,121 There isalso a paucity of literature examining the differ-ences in the ADC value for HCC, dysplastic nod-ules (DN), and regenerative nodules. A report byXu and colleagues122 did find that the ADC valuesof HCC were significantly lower than DN (1.28 �

Fig. 8. A 79-year-old woman with subcapsular ovarian canc1.5 Twith b values of 50 and 1000 s/mm2 (A, B), ADC map (a metastatic lesion that is T2 hyperintense with diffusion

10�3 mm2/s vs 1.53 � 10�3 mm2/s, respectively,P 5 .0082) in a series of 54 patients with 40HCCs and 19 DNs, with histopathologicconfirmation.

Common pitfalls in using DWI for lesioncharacterizationThere are several important limitations to the DWIcharacterization of focal liver lesions. As stated,DWI and ADC characterization cannot distinguishbetween benign and malignant solid liver lesionsgiven the overlap in signal intensity and ADCvalues (Fig. 9). Furthermore, malignant lesionsmay be heterogeneous and may contain cystic,mucinous, or necrotic components (Fig. 10). TheT2 prolongation of these complex components re-sults in elevated ADC values, allowing for mischar-acterization of a malignant lesion as benign(Fig. 11).3,6 In addition, the ADC of liver abscessescan demonstrate significant overlap with bothhemangiomas and solid lesions (Fig. 12).106

Thus, it is essential to rely on clinical history inaddition to the imaging features on conventionalsequences in order to correctly diagnose inhomo-geneous metastasis and hepatic abscess.Issues related to image quality may also affect

the interpretation of DWI and ADC. Patients withchronic liver disease and cirrhosis may haveconcomitant hepatic iron deposition, which maycause a significant T2 shortening effect in the liverthat may result in decreased SNR on DW images,most pronounced on higher b-value images, withfalsely decreased ADC.123–125 For example, ADC

er metastasis. Axial respiratory triggered SS EPI DWI atC), and fat-suppressed fast spin echo T2WI (D). There isrestriction (arrow). ADC value was 1.0 � 10�3 mm2/s.

Fig. 9. A 29-year-old man with biopsy-proven necrotizing granuloma. Axial RT SS EPI DWI with b values of 50 and1000 s/mm2 (A, B), ADC map (C), fat-suppressed fast spin echo T2WI (D), and gadobutrol-enhanced gradient echoT1WI on arterial and equilibrium phases (E, F). There is a T2-weighted hypointense, enhancing lesion in segment 2with foci of restricted diffusion and adjacent capsular retraction (arrow). ADC value was 1.19 � 10�3 mm2/s.


values were noted to be significantly lower in pa-tients with hepatic siderosis in a series of 52 pa-tients with cirrhosis.124 Finally, the interpretationof absolute ADC quantification has yet to be clari-fied as multiple factors influence the ADC value (ie,instrumental, sequencing, biologic). DW imagesare often of limited quality with relatively poorSNR, spatial resolution, and EPI-related artifacts.DWI is still an imaging technique requiring imageoptimization to ensure consistent high-qualityperformance.126

In summary, qualitative and quantitative infor-mation provided by DWI sequence is best usedfor characterizing cystic/necrotic lesions andhighly cellular solid lesions. One must be awareof the potential pitfalls and limitations when in-terpreting DW images and ADC maps.

Tumor Treatment Response

DWI has been investigated as a tool to evaluatethe tumor response to therapy and potentially pre-dict which lesions will respond to treatment byproviding information regarding tumor viability,cellularity, and vascularity.127 Changes in DWIsignal intensity and ADC following treatment canprecede changes in lesion size or enhancementand reflect tumor necrosis.128 The ability to detectearly tumor treatment response or lack of

Fig. 10. A 61-year-old woman with cystic metastasis (arrow500 (B–D); and ADC map (E). There is a 1.2-cm cystic lesionSSFSE. The lesion is hyperintense on low b-value image whigh ADC signal (ADC value 3.4 � 10�3 mm2/s) (arrow).cyst; however, this lesion represented a cystic neuroendoc

response has a significant impact on clinical man-agement and treatment planning. Making thisdistinction would enable repeat treatment or alter-native therapy if necessary. Furthermore, given thesignificant impact and adverse effects of currenttreatments, including chemotherapy and LRT,the knowledge of how likely a lesion is to respondto therapy will also facilitate clinical decision mak-ing in terms of how aggressively to pursuetreatment.8

Metastatic lesions to the liver have shown a sig-nificant increase in ADC values following systemicchemotherapy that precedes changes in lesionsize.129,130 A recent study by Marugami and col-leagues131 showed significant overlap in ADCvalues between responding and nonrespondingcolorectal metastatic lesions following intra-arterial chemotherapy. Using a threshold changein ADC value of 3.5%, receiver operating charac-teristic analysis showed higher sensitivity andspecificity values for the percentage of minimumADC (100% and 92.6%, respectively) than for thepercentage of mean ADC (66.7% and 74.1%,respectively) to identify responding lesions. Thepercentage of min ADC indicated the mostdiffusion-restricting voxel within the ROI. Further-more, liver metastases with low pretreatmentADC values have responded better to systemicchemotherapy compared with lesions with highpretreatment ADC values in several series. High

). Axial T2 SSFSE (A); BH axial SS EPI DWI with b 50, 100,in segment 6, which is markedly hyperintense on T2

ith attenuation of signal on high b-value images andImaging characteristics are consistent with a benignrine metastasis.

Fig. 11. A 57-year-oldmanwith indeterminate liver lesion. There is a 6.9-cm segment 6 lesion (arrow) that is hyper-intense on axial half-Fourier acquisition single-shot turbo spin-echo (A), FB SS EPI DWI b50 (B), and b500 (C), withcorresponding high signal on ADCmap (D). ADC valuemeasured as 1.6� 10�3 mm2/s. The central hypointense arearepresents susceptibility artifact from intralesional hemorrhage (dashed arrow) with corresponding T2 black outon DWI. Given atypical findings on postcontrast T1WI (not shown), lesion was considered indeterminate and bi-opsy revealed epithelioid angiosarcoma.

Lewis et al384

pretreatment ADC value of liver metastases is,therefore, a potential predictor of a poor responseto chemotherapy.8,129,132 Metastatic lesions withhigher ADC values may, in fact, have diminishedblood supply and, consequentially, more baselinecellular hypoxia and necrosis, which may limit theeffectiveness of therapy. Another hypothesis toexplain a poor treatment response in lesions witha high ADC value is the variation in local immuneresponse: local host immune response results inincreased tissue cellularity and cellular swelling,resulting in a lower ADC.8 Fig. 13 demonstrates tu-mor recurrence in a patient with colon cancer whounderwent wedge resection of the liver.The development of LRT for HCC, including

thermal radiofrequency ablation (RFA), transarte-rial chemoembolization (TACE), transarterial radio-embolization (TARE), and external beam radiation,has revolutionized HCC treatment in nonoperativepatients or patients with unresectable disease.Lesions treated with RFA or TACE typically un-dergo coagulative hemorrhagic necrosis withheterogeneous hyperintensity on unenhancedT1WI, making the evaluation of CE-MR imaging

Fig. 12. A 67-year-old man after chemotherapy for diffuseEPI DWI with b values of 50 and 800 s/mm2 (A, B), ADCgadobutrol-enhanced gradient echo T1WI on late venoupresent in the right hepatic lobe (arrow) demonstrating pdiffusion on both b 50 (A) and b750 (B). fine-needle aspirpatible with hepatic abscess.

challenging.133 Several researchers have investi-gated DWI to evaluate the HCC response toTACE and TARE.7,128,134–138 LRT causes disrup-tion of cellular membranes, cell death, and tumornecrosis with the resultant increase in water diffu-sivity. An increase in ADC values has been re-ported following TACE in the early posttreatmentperiod (a few days up to 2 weeks) with measurabledifferences before and after treatment(Fig. 14).7,128,135 ADC values also increase afterTARE according to several preliminary series, butthe treatment effect has been noted at 1 to3 months after treatment.137,139,140 ADC hasdemonstrated a significant correlation with tumornecrosis, especially complete tumor necrosis as-sessed with histopathology.141 Furthermore,studies have demonstrated differences in ADCvalues between viable and necrotic portions ofHCCs.7,141 In theory, the viable portion of the tu-mor restricts diffusion (hyperintense with lowADC value), whereas the necrotic portions willshow relatively unimpeded water diffusion (hyper-intense with high ADC value).142 Fig. 15 demon-strates an example of a patient who underwent

large B-cell lymphoma. Axial respiratory-triggered SSmap (C), fat-suppressed fast spin echo T2WI (D), ands phase (E). A thick-walled T2 hyperintense lesion iseripheral rim enhancement (E) and centrally restrictedation–guided biopsy of the lesion revealed pus, com-

Fig. 13. A 51-year-old man after partial right wedge resection for metastatic colon cancer with tumor recurrence.Axial respiratory-triggered SS EPI DWI with b value of 500 s/mm2 (A), axial BH fat-suppressed T2WI (B), and axialfat-suppressed gradient echo T1WI after gadoxetic acid injection at portal venous phase (C) show a hypovascularmetastatic lesion at the resection margin (arrow), which is hyperintense on DWI and T2WI.


TACE and RFA to an index HCCwith partial necro-sis, tumors. Recent publications have also investi-gated the role of the pretreatment ADC value inpredicting the response to TACE with discordantresults. Mannelli and colleagues138 showed thatHCCs with a low pre-TACE ADC tended to beless responsive to TACE than those with a higherADC, whereas Yuan and colleagues143 showedthat HCCs with a higher pre-TACE ADC wereless responsive than those with a lower ADC toTACE. Thus, DWI and ADC quantification mayplay a very important role in the noninvasive, non-contrast assessment of the posttreatmentresponse and guiding clinical treatment planningin a select population of patients with HCC.144

Pre-LRT ADC may be useful in predicting theresponse to therapy; however, large prospectivestudies are needed.

In summary, DWI is a potential useful noninva-sive tool in evaluating the response to therapyfor liver metastases and HCC. Changes in theADC value precede changes in lesion size orenhancement and correspond with tumor ne-crosis. Pretreatment ADC values may also be ofvalue in predicting the response to treatment.Therefore, DWI and ADC quantification mayprovide early information valuable for treat-ment planning and clinical decision making.However, large prospective studies are needed.

Liver Fibrosis and Cirrhosis

Liver fibrosis and cirrhosis affect the parenchymalstructure and function because of excessive extra-cellular matrix deposition. These microscopicchanges may lead to reduced diffusion becauseof a larger proportion of macromolecules, suchas collagen (inducing restricted motion), andmay also lead to reduced perfusion. Previous

studies16,17,98,145,146 have reported correlationsbetween liver ADC and the fibrosis stage assessedby histopathology. In most studies, a decrease inADC was observed with an increasing fibrosisstage and in cirrhosis. Lewin and colleagues145 re-ported area under the curve (AUC) values of 0.79and 0.92 for the detection of stages F2-F4 andF3-F4, using ADC cutoff values of 1.24 and 1.21� 10�3 mm2/s, respectively. Taouli and col-leagues98 observed AUCs of 0.896 and 0.896 forthe detection of stages F2-F4 and F3-F4, usingcutoff values of 1.54 and 1.53 � 10�3 mm2/s.Wang and colleagues,146 in a study comparingDWI and MR elastography (MRE) for liver fibrosisdetection, reported AUCs of 0.78 to 0.88 usingADC for the detection of any stage of fibrosisversus AUC of 0.92 to 0.99 for MRE. BeyondADC measurements, the IVIM technique hasbeen tested recently for liver fibrosis and cirrhosisdetection14,16,17 with variable differencesobserved in fibrotic liver for both diffusion andperfusion metrics (Fig. 16). Fibrosis is, however,not the only source of altered diffusion propertiesin the noncirrhotic and cirrhotic liver. The influenceof inflammatory response on ADC has been stud-ied in lesions147,148 and in the liver26 showingrestricted diffusion possibly because of increasedviscosity. Increased fat content may also lead toreduced ADC.149

LIMITATIONS

DWI sequence has several important limitations interms of SS EPI image quality and ADC reproduc-ibility. The SS EPI DWI sequence is limited given itsrelatively low spatial resolution and reduced SNR.6

EPI artifacts, including ghosting and distortion,can result in image degradation, which is worseat higher field strength and in the left hepaticlobe. Despite gains in SNR, the susceptibility arti-facts caused by B1 inhomogeneities and dielec-tric/conductive artifacts may further impair 3-T

Fig. 14. A 66-year-old man with hepatitis C–related cirrhosis and HCC. A 3.2-cm HCC in segment 8 (arrow).Pretreatment MR imaging (top) and posttreatment MR imaging (bottom row). Respiratory-triggered SS EPIDWI at b50, 1000 (A, B), ADC (C), and arterial-phase CE T1WI (gadobutrol) (D) demonstrate HCC in the rightlobe with restricted diffusion (ADC value of 1.0 � 10�3 mm2/s). Patient underwent right hepatic lobe Y-90radio-embolization. Follow-up MR imaging using FB SS EPI DWI at b50, b1000 (E, F) and ADC (G) demonstratea significant increase in ADC value of 1.7 � 10�3 mm2/s. Postcontrast subtracted (H) image demonstrates completeinterval necrosis.

Lewis et al386

SS EPI images.55,58 Technical improvements,including the application of strong gradients,multichannel coils, high magnetic fields, andadvanced software platforms, are being used tooptimize DWI.6,126

Significant variability in ADC values has been re-ported, likely as a result of a combination of hard-ware and biologic factors.6 The CV of ADC valuesmeasured in liver parenchyma ranges from 3.0%to 16.2% in several reports.17,35,55,56 Several re-searchers have investigated ADC reproducibilityfor malignant liver lesions. The DWI acquisitiontechnique (BH, RT, or FB), ADC calculationmethod, and selection of b values all potentially

Fig. 15. A 79-year-old woman after TACE/RFA of left hepat500 s/mm2 (A, B), ADC map (C), fat-suppressed fast spin echon arterial phase (E), and late-phase subtraction images (Fand adjacent arterial enhancing nodule (E) with venous waT2-weighted hyperintensity (D) and restricted diffusion (A,additional HCCs in the left and caudate lobes (dashed arr

influence the ADC measurement and the repro-ducibility for malignant liver lesions.150 ADCscalculated from BH DWI were more reproduciblethan fromRT DWI in a recent study by Kim and col-leagues.150 An additional recent study reports95% limits of agreement of repeated ADC mea-surements of malignant liver lesions of 28.7%–31.3% of the mean.36 Greater reproducibility ofADC measurement for larger lesions and forlesions in the right hepatic lobe was also reportedin this study36 as well as in a recent studyusing IVIM.37 Further improvements and modifica-tions to the DWI sequence and ADC calculationare needed to optimize ADC measurement

ic lobe HCC. Axial FB SS EPI DWI with b values of 50 ando T2WI (D), gadoxetate-enhanced gradient echo T1WI). A left lobe approximately 50% necrotic treated HCCshout (F) are present (arrow). The viable nodule showsB) with low ADC value of 0.70 � 10�3 mm2/s. There areows).

Fig. 16. Parametric IVIM diffusion maps obtained with respiratory-triggered SS EPI sequence and 16 b values (0–800) in a 26-year-old healthy woman and a 55-year-old woman with hepatitis C virus (HCV) and stage F3 fibrosis.(A, E) Coronal single-shot echo-planar DW image (b200). (B, F) True diffusion coefficient (D) map. (C, G) PF map.(D, H) D* map. D and PF maps show lower values in a patient with HCV than in healthy subject. There is no visibledifference in D* maps. Corresponding parameter values were as follows: for D, 0.81 10�3 mm2/s and 1.0610�3 mm2/s for the patient with HCV and the healthy volunteer, respectively; for PF, 5.1% and 9.8%, respectively;and for D*, 50.5 � 10�5 mm2/s and 49.4 � 10�5 mm2/s, respectively. (From Dyvorne HA, Galea N, Nevers T, et al.Diffusion-weighted imaging of the liver with multiple b values: effect of diffusion gradient polarity and breath-ing acquisition on image quality and intravoxel incoherent motion parameters–a pilot study. Radiology2013;266:924; with permission.)


reproducibility and accuracy, thus, enabling a reli-able comparison of research studies and the eval-uation of the posttreatment tumor response.Standardization of imaging technique and ADCcalculation methods across imaging platformsmay serve to address these issues.

In summary, DWI sequences are limited becauseof the relative poor SNR- and EPI-related arti-facts, including ghosting and distortion. ADCquantification is limited by poor measurementreproducibility for both liver parenchyma andlesions. Future DWI sequence improvementsand modifications are needed to address theselimitations.

FUTURE DIRECTIONSIVIM in the Clinic

Emerging research has investigated the roleof IVIM for the following potential clinicalapplications.

Noninvasive detection of liver fibrosisGiven the experience with DWI and monoexpo-nential ADC calculation, preliminary work has

shown that IVIM can distinguish the individualcontributions of cellular restricted diffusion andtissue perfusion, as defined by true D,perfusion-related pseudodiffusion (PF), andD*.11,14,16,17,35,151 In a study comparing RT andFB IVIM using 9 b values, there was a significantdecrease in liver true D, PF, D*, and ADC incirrhotic livers compared with healthy livers.17

Similarly, a recent study using triexponentialIVIM analysis also demonstrated a significantreduction in D*, D, and PF in cirrhosis.151 In a pro-spective pilot study by Dyvorne and colleagues14

evaluating the effect of diffusion gradient polarityand breathing acquisition on image quality, PFand D were reduced in patients with hepatitis Cvirus–related liver fibrosis. Overall, these findingsreflect the physiologic and microscopic cellularchanges to the liver with fibrosis and cirrhosis,namely, the decrease in hepatic blood flow(reduction in portal flow) and the increased depo-sition of proteins and macromolecules (collagen).Current challenges in the role of IVIM for liverfibrosis quantification remains the optimizationof image quality and standardization of the tech-nique and interplatform reproducibility for whichfurther research is necessary.44

Lewis et al388

Liver lesion characterizationIVIM is being investigated for liver lesion charac-terization in hopes of addressing the existing chal-lenges for ADC characterization, including thesignificant overlap in ADC values for benign andmalignant lesions and the variability in ADC mea-surement across imaging platforms. The ability todiscriminate and measure diffusion and perfusioncontributions may be of value for highly cellularand hypervascular lesions.11 A recent study byYoon and colleagues152 retrospectively examined142 patients with 169 focal liver lesions (120 malig-nant lesions) and found that the diagnostic perfor-mance of true diffusion (D) was better than that ofADCtotal in the diagnosis of malignancy (AUC 0.971using a cutoff value 1.278 � 10�3 mm2/s). In addi-tion, the researchers found that the PF and D*were significantly higher in hypervascular lesionscompared with hypovascular lesions. Microvas-cular flow in hypervascular lesions, thus, contami-nated the ADCtotal value; therefore, the removal ofperfusion contamination may be useful in revealingthe increased cellularity in malignant hypervascu-lar lesions, thus, allowing improved characteriza-tion.152 The exciting new application of IVIM islimited by measurement reproducibility, for whichfurther investigation is warranted.13,37

Combination and Comparison with PositronEmission Tomography/MR Imaging

Recent technologic advancements have led tothe integration of DWI with fludeoxyglucose(FDG)-positron emission tomography (PET), awell-established functional modality for variousoncologic and nononcologic applications, allow-ing for simultaneous data acquisition. PET-MR im-aging has the potential to combine excellent softtissue contrast resolution, anatomic resolution,and functional information with reduction in overallradiation exposure. DWI and PET are predicatedon entirely different fundamental principles andmechanisms, DWI assessing random motion ofwater and PET assessing cell glucose metabolismwithin biologic tissues. Maximum standardizeduptake value (SUVmax) is a quantitative measureof cellular glucose metabolism and has beenshown to correlate with histopathologic tumorgrade and the presence of necrosis.153,154 Thesemodalities can provide complementary informa-tion. DWI may be of value for lesions that are asso-ciated with low 18F-FDG accumulation and for theassessment of organs with normal high levels ofbackground 18F-FDG accumulation. PET may beof value for lesions close to the heart or spleen.155

Both DWI and PET can be acquired using the FBtechnique with the ability to perform whole-body

imaging and multiplanar and 3-dimensional refor-mats. DWI and PET images are visually similar inappearance. Diagnostic performance of PET-MRimaging can be improved with the acquisition ofanatomic MR imaging sequences.156 PET-MR im-aging T1-weighted gradient-echo images andT2WI have more accurate spatial registrationthan PET/CT images, although the DWI sequencecan result in significant image misregistration fromEPI-related geometric distortions, which is a po-tential weakness of the DWI-PET integration.157

Studies on the correlation between PET and DWIare limited, and results of preliminary work haveshown an inverse relationship between ADC andSUVmax.

158–160 A recent prospective study wasthe first to examine the correlation between SUVand ADC in neoplastic lesions using a simultaneoushybrid PET/MR imaging system. In this study, therewas a significant negative correlation betweenlesion SUV ratio (SUVmax/mean liver SUV) andADCmin (correlation coefficient [R] 5 �0.46,P<.04) for nonosseous soft tissue lesions.161 Well-designed prospective studies are needed toprovide more insight and experience into the PET-DWI integration, especially in terms of patient toler-ance, safety, diagnostic and prognostic capability,and cost-effectiveness.155 Currently, there are nopublished clinical studies investigating the role ofDWI-PET in liver malignancy.

New DWI Sequences

The SS EPI sequence is the standard pulsesequence for liver MR imaging given its rapidacquisition time and relative insensitivity to patientmotion.162 Given enhancements in gradient sys-tems and the emergence of parallel imaging, addi-tional pulse sequences have been investigated foruse in abdominal imaging, including spin-echo,stimulated-echo, fast-spin-echo, and SSFP se-quences.163 The most important alternativesequences used in the abdomen include thesingle-shot and PROPELLER fast-spin-echo andSSFP sequences.163 Modifications of imaging pa-rameters including the selection of b values isnecessary when using these pulse sequences inorder to obtain sufficient image quality and reliablequantification of diffusion parameters.163

DTI, which uses additional sensitizing gradientsto plot the relative degree of diffusion in multiple di-rections, has been studied in anatomy wherebydiffusion is preferentially restricted in one direc-tion.21 DTI has been investigated in the central ner-vous system for the evaluation of nonisotropictissue, such as white matter and spinal cord dis-eases. Limited data are available for DTI in theliver.25,52 The use of multiple diffusion gradient


directions with DTI results in a stronger appliedgradient and shorter time to echo values.52

Because more dimensions are plotted in DTIcompared with conventional DWI, there is a theo-retical advantage of more precise ADC calculationand information regarding tissue anisotropy. Pre-liminary work has investigated DTI in both healthyand fibrotic livers, with mixed results regarding im-age quality of DTI compared with conventionalDWI.26 Taouli and colleagues52 showed nocompromise in hepatic ADC values using DTI.Further work is necessary to determine the role ofDTI, if any, in the assessment of diffuse liver dis-ease or focal liver lesion.

In summary, emerging applications of DWI tech-nology, including IVIM, DWI/PET integration,and new DWI sequences, may further expandthe role and ability of DWI to accurately detectand characterize diffuse and focal liver disease.

SUMMARY

In this review, the authors describe their currentunderstanding, applications, and potential of DWIfor the evaluation of focal and diffuse liver disease.DWI has many potential evolving applications,including focal liver lesion detection and charac-terization, the assessment of tumor response,and the evaluation of diffuse liver disease. In theauthors’ opinion, DWI should be incorporatedinto routine liver MR imaging protocols. DWI isbest used in combination with conventional se-quences, especially CE sequences, for detectionand characterization of malignant liver lesions.DWI currently has a limited role as an isolatedtechnique for lesion characterization. The readermust be cognizant of the potential pitfalls and lim-itations of this technique. For patients withimpaired renal function or GBCA allergy, DWI is areasonable noninvasive alternative to CE MR im-aging. Further sequence improvements and wide-spread standardization of technique, b values, andimaging parameters are necessary for DWI toreach its full potential. Future applicationsincluding IVIM and DWI/PET MR imaging showpromise and warrant further investigation.

REFERENCES

1. Le Bihan D. Molecular diffusion nuclear mag-

netic resonance imaging. Magn Reson Q 1991;

7(1):1–30.

2. Bammer R. Basic principles of diffusion-weighted

imaging. Eur J Radiol 2003;45(3):169–84.

3. Taouli B, Koh DM. Diffusion-weighted MR imaging

of the liver. Radiology 2010;254(1):47–66.

4. Charnsangavej C, Clary B, Fong Y, et al. Selection

of patients for resection of hepatic colorectal me-

tastases: expert consensus statement. Ann Surg

Oncol 2006;13(10):1261–8.

5. Bruix J, Sherman M, Llovet JM, et al. Clinical man-

agement of hepatocellular carcinoma. Conclusions

of the Barcelona-2000 EASL conference. European

Association for the Study of the Liver. J Hepatol

2001;35(3):421–30.

6. Galea N, Cantisani V, Taouli B. Liver lesion detec-

tion and characterization: role of diffusion-

weighted imaging. J Magn Reson Imaging 2013;

37(6):1260–76.

7. Kamel IR, Bluemke DA, Eng J, et al. The role of

functional MR imaging in the assessment of tumor

response after chemoembolization in patients with

hepatocellular carcinoma. J Vasc Interv Radiol

2006;17(3):505–12.

8. Tam HH, Collins DJ, Brown G, et al. The role of pre-

treatment diffusion-weighted MRI in predicting

long-term outcome of colorectal liver metastasis.

Br J Radiol 2013;86(1030):20130281.

9. Holz M, Heil SR, Sacco A. Temperature-dependent

self-diffusion coefficients of water and six selected

molecular liquids for calibration in accurate 1H

NMR PFG measurements. Phys Chem Chem

Phys 2000;2(20):4740–2.

10. Stejskal EO, Tanner JE. Spin diffusion measure-

ments: spin echoes in the presence of a time-

dependent field gradient. J Chem Phys 1965;42:

288–92.

11. Le Bihan D, Breton E, LallemandD, et al. Separation

of diffusion and perfusion in intravoxel incoherent

motion MR imaging. Radiology 1988;168(2):

497–505.

12. KohDM,CollinsDJ,OrtonMR. Intravoxel incoherent

motion in body diffusion-weighted MRI: reality and

challenges. AJR Am J Roentgenol 2011;196(6):

1351–61.

13. Andreou A, Koh DM, Collins DJ, et al. Measure-

ment reproducibility of perfusion fraction and pseu-

dodiffusion coefficient derived by intravoxel

incoherent motion diffusion-weighted MR imaging

in normal liver and metastases. Eur Radiol 2013;

23(2):428–34.

14. Dyvorne HA, Galea N, Nevers T, et al. Diffusion-

weighted imaging of the liver with multiple B

values: effect of diffusion gradient polarity and

breathing acquisition on image quality and intra-

voxel incoherent motion parameters–a pilot study.

Radiology 2013;266(3):920–9.

15. Kang KM, Lee JM, Yoon JH, et al. Intravoxel inco-

herent motion diffusion-weighted MR imaging for

characterization of focal pancreatic lesions. Radi-

ology 2014;270(2):444–53.

http://refhub.elsevier.com/S1064-9689(14)00042-7/sref1






























































Lewis et al390

16. Luciani A, Vignaud A, Cavet M, et al. Liver

cirrhosis: intravoxel incoherent motion MR imag-

ing–pilot study. Radiology 2008;249(3):891–9.

17. Patel J, Sigmund EE, Rusinek H, et al. Diagnosis of

cirrhosis with intravoxel incoherent motion diffusion

MRI and dynamic contrast-enhanced MRI alone

and in combination: preliminary experience.

J Magn Reson Imaging 2010;31(3):589–600.

18. Yamada I, Aung W, Himeno Y, et al. Diffusion coef-

ficients in abdominal organs and hepatic lesions:

evaluation with intravoxel incoherent motion

echo-planar MR imaging. Radiology 1999;210(3):

617–23.

19. Jensen JH, Helpern JA, Ramani A, et al. Diffusional

kurtosis imaging: the quantification of non-

gaussian water diffusion by means of magnetic

resonance imaging. Magn Reson Med 2005;

53(6):1432–40.

20. Rosenkrantz AB, Sigmund EE, Winnick A, et al.

Assessment of hepatocellular carcinoma using

apparent diffusion coefficient and diffusion kurto-

sis indices: preliminary experience in fresh liver

explants. Magn Reson Imaging 2012;30(10):

1534–40.

21. Le Bihan D, Mangin JF, Poupon C, et al. Diffusion

tensor imaging: concepts and applications.


22. Gurses B, Kabakci N, Kovanlikaya A, et al. Diffu-

sion tensor imaging of the normal prostate at 3

Tesla. Eur Radiol 2008;18(4):716–21.

23. Kataoka M, Kido A, Yamamoto A, et al. Diffusion

tensor imaging of kidneys with respiratory trig-

gering: optimization of parameters to demonstrate

anisotropic structures on fraction anisotropy

maps. J Magn Reson Imaging 2009;29(3):736–44.

24. Manenti G, Carlani M, Mancino S, et al. Diffusion

tensor magnetic resonance imaging of prostate

cancer. Invest Radiol 2007;42(6):412–9.

25. Sigmund EE, Vivier PH, Sui D, et al. Intravoxel inco-

herent motion and diffusion-tensor imaging in renal

tissue under hydration and furosemide flow chal-

lenges. Radiology 2012;263(3):758–69.

26. Taouli B, Chouli M, Martin AJ, et al. Chronic hepa-

titis: role of diffusion-weighted imaging and diffu-

sion tensor imaging for the diagnosis of liver

fibrosis and inflammation. J Magn Reson Imaging

2008;28(1):89–95.

27. Cheung JS, Fan SJ, Gao DS, et al. Diffusion tensor

imaging of liver fibrosis in an experimental model.


28. Taouli B, Vilgrain V, Dumont E, et al. Evaluation of

liver diffusion isotropy and characterization of focal

hepatic lesions with two single-shot echo-planar

MR imaging sequences: prospective study in 66

patients. Radiology 2003;226(1):71–8.

29. Sasaki M, Yamada K, Watanabe Y, et al. Variability

in absolute apparent diffusion coefficient values

across different platforms may be substantial: a

multivendor, multi-institutional comparison study.

Radiology 2008;249(2):624–30.

30. Chenevert TL, Galban CJ, Ivancevic MK, et al.

Diffusion coefficient measurement using a

temperature-controlled fluid for quality control in

multicenter studies. J Magn Reson Imaging 2011;

34(4):983–7.

31. Girometti R, Furlan A, Esposito G, et al. Relevance of

b-values in evaluating liver fibrosis: a study in healthy

and cirrhotic subjects using two single-shot spin-

echo echo-planar diffusion-weighted sequences.


32. Xing D, Papadakis NG, Huang CL, et al. Optimised

diffusion-weighting for measurement of apparent

diffusion coefficient (ADC) in human brain. Magn

Reson Imaging 1997;15(7):771–84.

33. Lemke A, Stieltjes B, Schad LR, et al. Toward an

optimal distribution of b values for intravoxel inco-

herent motion imaging. Magn Reson Imaging

2011;29(6):766–76.

34. Zhang JL, Sigmund EE, Rusinek H, et al. Optimiza-

tion of b-value sampling for diffusion-weighted

imaging of the kidney. Magn Reson 2012;67(1):

89–97.

35. Braithwaite AC, Dale BM, Boll DT, et al. Short- and

midterm reproducibility of apparent diffusion coef-

ficient measurements at 3.0-T diffusion-weighted

imaging of the abdomen. Radiology 2009;250(2):

459–65.

36. Kim SY, Lee SS, Byun JH, et al. Malignant hepatic

tumors: short-term reproducibility of apparent diffu-

sion coefficients with breath-hold and respiratory-

triggereddiffusion-weightedMR imaging.Radiology

2010;255(3):815–23.

37. Kakite S, Dyvorne HA, Besa C, et al. Hepatocellular

carcinoma: short-term reproducibility of ADC and

IVIM parameters at 3.0T. J Magn Reson Imag

2014, in press.

38. Le Bihan D, Poupon C, Amadon A, et al. Artifacts

and pitfalls in diffusion MRI. J Magn Reson Imaging

2006;24(3):478–88.

39. Porter DA, Heidemann RM. High resolution

diffusion-weighted imaging using readout-

segmented echo-planar imaging, parallel imaging

and a two-dimensional navigator-based reacquisi-

tion. Magn Reson 2009;62(2):468–75.

40. Lovblad KO, Jakob PM, Chen Q, et al. Turbo spin-

echo diffusion-weighted MR of ischemic stroke.

AJNR Am J Neuroradiol 1998;19(2):201–8 [discus-

sion: 209].

41. Lu L, Erokwu B, Lee G, et al. Diffusion-prepared

fast imaging with steady-state free precession

(DP-FISP): a rapid diffusion MRI technique at 7 T.

Magn Reson 2012;68(3):868–73.

42. Deng J, Miller FH, Salem R, et al. Multishot diffusion-

weighted PROPELLER magnetic resonance




















































































































imaging of the abdomen. Invest Radiol 2006;41(10):

769–75.

43. Pipe JG. Motion correction with PROPELLER MRI:

application to head motion and free-breathing car-

diac imaging. Magn Reson 1999;42(5):963–9.

44. Kwee TC, Takahara T, Koh DM, et al. Comparison

and reproducibility of ADCmeasurements in breath-

hold, respiratory triggered, and free-breathing diffu-

sion-weighted MR imaging of the liver. J Magn

Reson Imaging 2008;28(5):1141–8.

45. Taouli B, Sandberg A, Stemmer A, et al. Diffusion-

weighted imaging of the liver: comparison of navi-

gator triggered and breathhold acquisitions.


46. MurtzP, FlackeS,TraberF, et al. Abdomen:diffusion-

weighted MR imaging with pulse-triggered single-

shot sequences. Radiology 2002;224(1):258–64.

47. Johnson GA, Maki JH. In vivo measurement of pro-

ton diffusion in the presence of coherent motion.

Invest Radiol 1991;26(6):540–5.

48. Ozaki M, Inoue Y, Miyati T, et al. Motion artifact

reduction of diffusion-weighted MRI of the liver:

use of velocity-compensated diffusion gradients

combined with tetrahedral gradients. J Magn Re-

son Imaging 2013;37(1):172–8.

49. GriswoldMA,JakobPM,HeidemannRM,etal.Gener-

alized autocalibrating partially parallel acquisitions

(GRAPPA). Magn Reson Med 2002;47(6):1202–10.

50. Pruessmann KP, Weiger M, Scheidegger MB, et al.

SENSE: sensitivity encoding for fast MRI. Magn Re-

son 1999;42(5):952–62.

51. Sodickson DK, Manning WJ. Simultaneous acquisi-

tion of spatial harmonics (SMASH): fast imaging

with radiofrequency coil arrays. Magn Reson Med

1997;38(4):591–603.

52. Taouli B, Martin AJ, Qayyum A, et al. Parallel imag-

ing and diffusion tensor imaging for diffusion-

weighted MRI of the liver: preliminary experience

in healthy volunteers. AJR Am J Roentgenol 2004;

183(3):677–80.

53. Setsompop K, Gagoski BA, Polimeni JR, et al. Blip-

ped-controlled aliasing in parallel imaging for

simultaneous multislice echo planar imaging with

reduced g-factor penalty. Magn Reson Med 2012;

67(5):1210–24.

54. Setsompop K, Cohen-Adad J, Gagoski BA, et al.

Improving diffusion MRI using simultaneous multi-

slice echo planar imaging. Neuroimage 2012;

63(1):569–80.

55. Rosenkrantz AB, Oei M, Babb JS, et al. Diffusion-

weighted imaging of the abdomen at 3.0 Tesla: im-

age quality and apparent diffusion coefficient

reproducibility compared with 1.5 Tesla. J Magn

Reson Imaging 2011;33(1):128–35.

56. Dale BM, Braithwaite AC, Boll DT, et al. Field

strength and diffusion encoding technique affect

the apparent diffusion coefficient measurements

in diffusion-weighted imaging of the abdomen.

Invest Radiol 2010;45(2):104–8.

57. Coenegrachts K, Delanote J, Ter Beek L, et al.

Improved focal liver lesion detection: comparison

of single-shot diffusion-weighted echoplanar and

single-shot T2 weighted turbo spin echo tech-

niques. Br J Radiol 2007;80(955):524–31.

58. van den Bos IC, Hussain SM, Krestin GP, et al.

Liver imaging at 3.0 T: diffusion-induced black-

blood echo-planar imaging with large anatomic

volumetric coverage as an alternative for specific

absorption rate-intensive echo-train spin-echo se-

quences: feasibility study. Radiology 2008;248(1):

264–71.

59. Okada Y, Ohtomo K, Kiryu S, et al. Breath-hold T2-

weighted MRI of hepatic tumors: value of echo

planar imaging with diffusion-sensitizing gradient.

J Comput Assist Tomogr 1998;22(3):364–71.

60. Hussain SM, De Becker J, Hop WC, et al. Can a

single-shot black-blood T2-weighted spin-echo

echo-planar imaging sequence with sensitivity en-

coding replace the respiratory-triggered turbo

spin-echo sequence for the liver? An optimization

and feasibility study. J Magn Reson Imaging

2005;21(3):219–29.

61. Moteki T, Horikoshi H. Evaluation of hepatic le-

sions and hepatic parenchyma using diffusion-

weighted echo-planar MR with three values of

gradient b-factor. J Magn Reson Imaging 2006;

24(3):637–45.

62. Wu LM, Hu J, Gu HY, et al. Can diffusion-weighted

magnetic resonance imaging (DW-MRI) alone be

used as a reliable sequence for the preoperative

detection and characterisation of hepatic metasta-

ses? A meta-analysis. Eur J Cancer 2013;49(3):

572–84.

63. Cummings LC, Payes JD, Cooper GS. Survival

after hepatic resection in metastatic colorectal can-

cer: a population-based study. Cancer 2007;

109(4):718–26.

64. Benoist S, Nordlinger B. The role of preoperative

chemotherapy in patients with resectable colo-

rectal liver metastases. Ann Surg Oncol 2009;

16(9):2385–90.

65. Gillams AR, Lees WR. Five-year survival in 309 pa-

tients with colorectal liver metastases treated with

radiofrequency ablation. Eur Radiol 2009;19(5):

1206–13.

66. Koh DM, Collins DJ, Wallace T, et al. Combining

diffusion-weighted MRI with Gd-EOB-DTPA-

enhanced MRI improves the detection of colorectal

liver metastases. Br J Radiol 2012;85(1015):980–9.

67. Zech CJ, Herrmann KA, Dietrich O, et al. Black-

blood diffusion-weighted EPI acquisition of the liver

with parallel imaging: comparison with a standard

T2-weighted sequence for detection of focal liver

lesions. Invest Radiol 2008;43(4):261–6.



















































































































Lewis et al392

68. Bruegel M, Gaa J, Waldt S, et al. Diagnosis of

hepatic metastasis: comparison of respiration-

triggered diffusion-weighted echo-planar MRI and

five t2-weighted turbo spin-echo sequences. AJR

Am J Roentgenol 2008;191(5):1421–9.

69. d’Assignies G, Fina P, Bruno O, et al. High sensi-

tivity of diffusion-weighted MR imaging for the

detection of liver metastases from neuroendocrine

tumors: comparison with T2-weighted and dynamic

gadolinium-enhanced MR imaging. Radiology

2013;268(2):390–9.

70. Soyer P, Boudiaf M, Place V, et al. Preoperative

detection of hepatic metastases: comparison of

diffusion-weighted, T2-weighted fast spin echo

and gadolinium-enhanced MR imaging using sur-

gical and histopathologic findings as standard of

reference. Eur J Radiol 2011;80(2):245–52.

71. Parikh T, Drew SJ, Lee VS, et al. Focal liver lesion

detection and characterization with diffusion-

weighted MR imaging: comparison with standard

breath-hold T2-weighted imaging. Radiology

2008;246(3):812–22.

72. Choi JS, Kim MJ, Choi JY, et al. Diffusion-weighted

MR imaging of liver on 3.0-Tesla system: effect of

intravenous administration of gadoxetic acid diso-

dium. Eur Radiol 2010;20(5):1052–60.

73. Nasu K, Kuroki Y, Nawano S, et al. Hepatic metas-

tases: diffusion-weighted sensitivity-encoding

versus SPIO-enhanced MR imaging. Radiology

2006;239(1):122–30.

74. Koh DM, Brown G, Riddell AM, et al. Detection of

colorectal hepatic metastases using MnDPDP MR

imaging and diffusion-weighted imaging (DWI)

alone and in combination. Eur Radiol 2008;18(5):

903–10.

75. Hardie AD, Naik M, Hecht EM, et al. Diagnosis of

liver metastases: value of diffusion-weighted MRI

compared with gadolinium-enhanced MRI. Eur Ra-

diol 2010;20(6):1431–41.

76. Shimada K, Isoda H, Hirokawa Y, et al. Comparison

of gadolinium-EOB-DTPA-enhanced and diffusion-

weighted liver MRI for detection of small hepatic

metastases. Eur Radiol 2010;20(11):2690–8.

77. Chung WS, Kim MJ, Chung YE, et al. Comparison

of gadoxetic acid-enhanced dynamic imaging

and diffusion-weighted imaging for the preopera-

tive evaluation of colorectal liver metastases.


78. Holzapfel K, Reiser-Erkan C, Fingerle AA, et al.

Comparison of diffusion-weighted MR imaging

and multidetector-row CT in the detection of liver

metastases in patients operated for pancreatic

cancer. Abdom Imaging 2011;36(2):179–84.

79. Kenis C, Deckers F, De Foer B, et al. Diagnosis of

liver metastases: can diffusion-weighted imaging

(DWI) be used as a stand alone sequence? Eur J

Radiol 2012;81(5):1016–23.

80. Lowenthal D, Zeile M, Lim WY, et al. Detection and

characterisation of focal liver lesions in colorectal

carcinoma patients: comparison of diffusion-

weighted and Gd-EOB-DTPA enhanced MR imag-

ing. Eur Radiol 2011;21(4):832–40.

81. Kim YK, Lee MW, Lee WJ, et al. Diagnostic accu-

racy and sensitivity of diffusion-weighted and of

gadoxetic acid-enhanced 3-T MR imaging alone

or in combination in the detection of small liver

metastasis (</5 1.5 cm in diameter). Invest Radiol

2012;47(3):159–66.

82. Macera A, Lario C, Petracchini M, et al. Staging of

colorectal liver metastases after preoperative

chemotherapy. Diffusion-weighted imaging in com-

bination with Gd-EOB-DTPA MRI sequences in-

creases sensitivity and diagnostic accuracy. Eur

Radiol 2013;23(3):739–47.

83. Parkin DM, Bray F, Ferlay J, et al. Estimating the

world cancer burden: Globocan 2000. Int J Cancer

2001;94(2):153–6.

84. Bruix J, Llovet JM. Major achievements in hepato-

cellular carcinoma. Lancet 2009;373(9664):614–6.

85. Piana G, Trinquart L, Meskine N, et al. New MR im-

aging criteria with a diffusion-weighted sequence

for the diagnosis of hepatocellular carcinoma

in chronic liver diseases. J Hepatol 2011;55(1):

126–32.

86. Forner A, Vilana R, Ayuso C, et al. Diagnosis of he-

patic nodules 20 mm or smaller in cirrhosis: pro-

spective validation of the noninvasive diagnostic

criteria for hepatocellular carcinoma. Hepatology

2008;47(1):97–104.

87. Kim JE, Kim SH, Lee SJ, et al. Hypervascular hepa-

tocellular carcinoma 1 cm or smaller in patients

with chronic liver disease: characterization with ga-

doxetic acid-enhanced MRI that includes diffusion-

weighted imaging. AJR Am J Roentgenol 2011;

196(6):W758–65.

88. Vandecaveye V, De Keyzer F, Verslype C, et al.

Diffusion-weighted MRI provides additional value

to conventional dynamic contrast-enhanced MRI

for detection of hepatocellular carcinoma. Eur Ra-

diol 2009;19(10):2456–66.

89. Hardie AD, Kizziah MK, Boulter DJ. Diagnostic ac-

curacy of diffusion-weighted MRI for identifying

hepatocellular carcinoma with liver explant correla-

tion. J Med Imaging Radiat Oncol 2011;55(4):

362–7.

90. Park MS, Kim S, Patel J, et al. Hepatocellular carci-

noma: detection with diffusion-weighted versus

contrast-enhanced magnetic resonance imaging

in pretransplant patients. Hepatology 2012;56(1):

140–8.

91. ParkMJ, KimYK, LeeMW, et al. Small hepatocellular

carcinomas: improved sensitivity by combining ga-

doxetic acid-enhanced and diffusion-weighted MR

imaging patterns. Radiology 2012;264(3):761–70.




















































































































92. Xu PJ, Yan FH, Wang JH, et al. Added value of

breathhold diffusion-weighted MRI in detection

of small hepatocellular carcinoma lesions

compared with dynamic contrast-enhanced MRI

alone using receiver operating characteristic

curve analysis. J Magn Reson Imaging 2009;

29(2):341–9.

93. Rosenkrantz AB, Lee L, Matza BW, et al. Infiltrative

hepatocellular carcinoma: comparison of MRI se-

quences for lesion conspicuity. Clin Radiol 2012;

67(12):e105–11.

94. Nasu K, Kuroki Y, Tsukamoto T, et al. Diffusion-

weighted imaging of surgically resected hepato-

cellular carcinoma: imaging characteristics and

relationship among signal intensity, apparent diffu-

sion coefficient, and histopathologic grade. AJR

Am J Roentgenol 2009;193(2):438–44.

95. Kim YK, Kim CS, Han YM, et al. Detection of liver

malignancy with gadoxetic acid-enhanced MRI: is

addition of diffusion-weighted MRI beneficial? Clin

Radiol 2011;66(6):489–96.

96. Sandrasegaran K, Tahir B, Patel A, et al. The use-

fulness of diffusion-weighted imaging in the char-

acterization of liver lesions in patients with

cirrhosis. Clin Radiol 2013;68(7):708–15.

97. Sandrasegaran K, Akisik FM, Lin C, et al. Value of

diffusion-weighted MRI for assessing liver fibrosis

and cirrhosis. AJR Am J Roentgenol 2009;193(6):

1556–60.

98. Taouli B, Tolia AJ, Losada M, et al. Diffusion-

weighted MRI for quantification of liver fibrosis:

preliminary experience. AJR Am J Roentgenol

2007;189(4):799–806.

99. Kim AY, Kim YK, Lee MW, et al. Detection of hepa-

tocellular carcinoma in gadoxetic acid-enhanced

MRI and diffusion-weighted MRI with respect to

the severity of liver cirrhosis. Acta Radiol 2012;

53(8):830–8.

100. Patel T. Increasing incidence and mortality of pri-

mary intrahepatic cholangiocarcinoma in the

United States. Hepatology 2001;33(6):1353–7.

101. Kim SJ, Lee JM, Han JK, et al. Peripheral mass-

forming cholangiocarcinoma in cirrhotic liver. AJR

Am J Roentgenol 2007;189(6):1428–34.

102. Lim JH. Cholangiocarcinoma: morphologic clas-

sification according to growth pattern and imag-

ing findings. AJR Am J Roentgenol 2003;181(3):

819–27.

103. Vilana R, Forner A, Bianchi L, et al. Intrahepatic pe-

ripheral cholangiocarcinoma in cirrhosis patients

may display a vascular pattern similar to hepato-

cellular carcinoma on contrast-enhanced ultra-

sound. Hepatology 2010;51(6):2020–9.

104. Kim SA, Lee JM, Lee KB, et al. Intrahepatic mass-

forming cholangiocarcinomas: enhancement pat-

terns at multiphasic CT, with special emphasis on

arterial enhancement pattern–correlation with

clinicopathologic findings. Radiology 2011;260(1):

148–57.

105. Park HJ, Kim YK, Park MJ, et al. Small intrahepatic

mass-forming cholangiocarcinoma: target sign on

diffusion-weighted imaging for differentiation from

hepatocellular carcinoma. Abdom Imaging 2013;

38(4):793–801.

106. Miller FH, Hammond N, Siddiqi AJ, et al. Utility of

diffusion-weighted MRI in distinguishing benign

and malignant hepatic lesions. J Magn Reson Im-

aging 2010;32(1):138–47.

107. Koh DM, Collins DJ. Diffusion-weighted MRI in the

body: applications and challenges in oncology.

AJR Am J Roentgenol 2007;188(6):1622–35.

108. Mohr A, Priebe M, Taouli B, et al. Selective water

excitation for faster MR imaging of articular carti-

lage defects: initial clinical results. Eur Radiol

2003;13(4):686–9.

109. Demir OI, Obuz F, Sagol O, et al. Contribution of

diffusion-weighted MRI to the differential diagnosis

of hepatic masses. Diagn Interv Radiol 2007;13(2):

81–6.

110. Ichikawa T, Haradome H, Hachiya J, et al. Charac-

terization of hepatic lesions by perfusion-weighted

MR imaging with an echoplanar sequence. AJR

Am J Roentgenol 1998;170(4):1029–34.

111. Namimoto T, Yamashita Y, Sumi S, et al. Focal liver

masses: characterization with diffusion-weighted

echo-planar MR imaging. Radiology 1997;204(3):

739–44.

112. Ichikawa T, Haradome H, Hachiya J, et al. Diffu-

sion-weighted MR imaging with a single-shot echo-

planar sequence: detection and characterization of

focal hepatic lesions. AJR Am J Roentgenol 1998;

170(2):397–402.

113. Holzapfel K, Bruegel M, Eiber M, et al. Character-

ization of small (</510mm) focal liver lesions: value

of respiratory-triggered echo-planar diffusion-

weighted MR imaging. Eur J Radiol 2010;76(1):

89–95.

114. Lee MH, Kim SH, Park MJ, et al. Gadoxetic acid-

enhanced hepatobiliary phase MRI and high-b-

value diffusion-weighted imaging to distinguish

well-differentiated hepatocellular carcinomas from

benign nodules in patients with chronic liver dis-

ease. AJR Am J Roentgenol 2011;197(5):W868–75.

115. Motosugi U, Ichikawa T, Sou H, et al. Distinguishing

hypervascular pseudolesions of the liver from hy-

pervascular hepatocellular carcinomas with gadox-

etic acid-enhanced MR imaging. Radiology 2010;

256(1):151–8.

116. Kim YK, Lee WJ, Park MJ, et al. Hypovascular hy-

pointense nodules on hepatobiliary phase gadox-

etic acid-enhanced MR images in patients with

cirrhosis: potential of DW imaging in predicting

progression to hypervascular HCC. Radiology

2012;265(1):104–14.



















































































































Lewis et al394

117. Papanikolaou N, Gourtsoyianni S, Yarmenitis S,

et al. Comparison between two-point and four-

point methods for quantification of apparent diffu-

sion coefficient of normal liver parenchyma and

focal lesions. Value of normalization with spleen.

Eur J Radiol 2010;73(2):305–9.

118. Bruegel M, Holzapfel K, Gaa J, et al. Characteriza-

tion of focal liver lesions by ADC measurements us-

ing a respiratory triggered diffusion-weighted

single-shot echo-planar MR imaging technique.

Eur Radiol 2008;18(3):477–85.

119. Gourtsoyianni S, Papanikolaou N, Yarmenitis S,

et al. Respiratory gated diffusion-weighted imaging

of the liver: value of apparent diffusion coefficient

measurements in the differentiation between most

commonly encountered benign and malignant

focal liver lesions. Eur Radiol 2008;18(3):486–92.

120. Sandrasegaran K, Akisik FM, Lin C, et al. The value

of diffusion-weighted imaging in characterizing focal

liver masses. Acad Radiol 2009;16(10):1208–14.

121. Agnello F, Ronot M, Valla DC, et al. High-b-value

diffusion-weighted MR imaging of benign hepato-

cellular lesions: quantitative and qualitative anal-

ysis. Radiology 2012;262(2):511–9.

122. Xu PJ, Yan FH, Wang JH, et al. Contribution of

diffusion-weighted magnetic resonance imaging

in the characterization of hepatocellular carci-

nomas and dysplastic nodules in cirrhotic liver.

J Comput Assist Tomogr 2010;34(4):506–12.

123. Sener RN. Echo-planar and gradient-echo diffusion

MRI of normal brain iron in the globus pallidus. Clin

Imaging 2002;26(6):371–4.

124. Chandarana H, Do RK, Mussi TC, et al. The effect

of liver iron deposition on hepatic apparent diffu-

sion coefficient values in cirrhosis. AJR Am J

Roentgenol 2012;199(4):803–8.

125. Metwally MA, Zein CO, Zein NN. Clinical signifi-

cance of hepatic iron deposition and serum iron

values in patients with chronic hepatitis C infection.

Am J Gastroenterol 2004;99(2):286–91.

126. Chandarana H, Taouli B. Diffusion and perfusion

imaging of the liver. Eur J Radiol 2010;76(3):

348–58.

127. Alonzi R, Hoskin P. Functional imaging in clinical

oncology: magnetic resonance imaging- and com-

puterised tomography-based techniques. Clin On-

col (R Coll Radiol) 2006;18(7):555–70.

128. Kamel IR, Liapi E, Reyes DK, et al. Unresectable

hepatocellular carcinoma: serial early vascular

and cellular changes after transarterial chemoem-

bolization as detected with MR imaging. Radiology

2009;250(2):466–73.

129. Cui Y, Zhang XP, Sun YS, et al. Apparent diffusion

coefficient: potential imaging biomarker for predic-

tion and early detection of response to chemo-

therapy in hepatic metastases. Radiology 2008;

248(3):894–900.

130. Anzidei M, Napoli A, Zaccagna F, et al. Liver me-

tastases from colorectal cancer treated with con-

ventional and antiangiogenetic chemotherapy:

evaluation with liver computed tomography perfu-

sion and magnetic resonance diffusion-weighted

imaging. J Comput Assist Tomogr 2011;35(6):

690–6.

131. MarugamiN,TanakaT,KitanoS,et al. Early detection

of therapeutic response to hepatic arterial infusion

chemotherapy of liver metastases from colorectal

cancer usingdiffusion-weightedMR imaging.Cardi-

ovasc Intervent Radiol 2009;32(4):638–46.

132. Koh DM, Scurr E, Collins D, et al. Predicting

response of colorectal hepatic metastasis: value

of pretreatment apparent diffusion coefficients.


133. Kierans AS, Elazzazi M, Braga L, et al. Thermoa-

blative treatments for malignant liver lesions:

10-year experience of MRI appearances of treat-

ment response. AJR Am J Roentgenol 2010;

194(2):523–9.

134. Kamel IR, Bluemke DA, Ramsey D, et al. Role of

diffusion-weighted imaging in estimating tumor ne-

crosis after chemoembolization of hepatocellular

carcinoma. AJR Am J Roentgenol 2003;181(3):

708–10.

135. Chen CY, Li CW, Kuo YT, et al. Early response of

hepatocellular carcinoma to transcatheter arterial

chemoembolization: choline levels and MR diffu-

sion constants–initial experience. Radiology 2006;

239(2):448–56.

136. Rhee TK, Omary RA, Gates V, et al. The effect of

catheter-directed CT angiography on yttrium-90 ra-

dioembolization treatment of hepatocellular carci-

noma. J Vasc Interv Radiol 2005;16(8):1085–91.

137. Kamel IR, Reyes DK, Liapi E, et al. Functional MR

imaging assessment of tumor response after 90Y

microsphere treatment in patients with unresect-

able hepatocellular carcinoma. J Vasc Interv Radiol

2007;18(1 Pt 1):49–56.

138. Mannelli L, Kim S, Hajdu CH, et al. Serial diffusion-

weighted MRI in patients with hepatocellular carci-

noma: prediction and assessment of response to

transarterial chemoembolization. Preliminary expe-

rience. Eur J Radiol 2013;82(4):577–82.

139. Deng J,Miller FH, Rhee TK, et al. Diffusion-weighted

MR imaging for determination of hepatocellular car-

cinoma response to yttrium-90 radioembolization.

J Vasc Interv Radiol 2006;17(7):1195–200.

140. Rhee TK, Naik NK, Deng J, et al. Tumor response

after yttrium-90 radioembolization for hepatocellu-

lar carcinoma: comparison of diffusion-weighted

functional MR imaging with anatomic MR imaging.

J Vasc Interv Radiol 2008;19(8):1180–6.

141. Mannelli L, Kim S, Hajdu CH, et al. Assessment

of tumor necrosis of hepatocellular carcinoma

after chemoembolization: diffusion-weighted and




















































































































contrast-enhanced MRI with histopathologic corre-

lation of the explanted liver. AJR Am J Roentgenol

2009;193(4):1044–52.

142. Lim KS. Diffusion-weighted MRI of hepatocellu-

lar carcinoma in cirrhosis. Clin Radiol 2014;

69(1):1–10.

143. Yuan Z, Ye XD, Dong S, et al. Role of magnetic

resonance diffusion-weighted imaging in evalu-

ating response after chemoembolization of hepato-

cellular carcinoma. Eur J Radiol 2010;75(1):e9–14.

144. Yaghmai V, Besa C, Kim E, et al. Imaging assess-

ment of hepatocellular carcinoma response to lo-

coregional and systemic therapy. AJR Am J

Roentgenol 2013;201(1):80–96.

145. Lewin M, Poujol-Robert A, Boelle PY, et al. Diffu-

sion-weighted magnetic resonance imaging for

the assessment of fibrosis in chronic hepatitis C.

Hepatology 2007;46(3):658–65.

146. Wang Y, Ganger DR, Levitsky J, et al. Assessment

of chronic hepatitis and fibrosis: comparison of MR

elastography and diffusion-weighted imaging. AJR

Am J Roentgenol 2011;196(3):553–61.

147. Goyal A, Sharma R, Bhalla AS, et al. Diffusion-

weighted MRI in inflammatory renal lesions: all

that glitters is not RCC! Eur Radiol 2013;23(1):

272–9.

148. Kiryu S, Dodanuki K, Takao H, et al. Free-breathing

diffusion-weighted imaging for the assessment of

inflammatory activity in Crohn’s disease. J Magn

Reson Imaging 2009;29(4):880–6.

149. Poyraz AK, Onur MR, Kocakoc E, et al. Diffusion-

weighted MRI of fatty liver. J Magn Reson Imaging

2012;35(5):1108–11.

150. Kim SY, Lee SS, Park B, et al. Reproducibility of

measurement of apparent diffusion coefficients of

malignant hepatic tumors: effect of DWI techniques

and calculation methods. J Magn Reson Imaging

2012;36(5):1131–8.

151. Hayashi T, Miyati T, Takahashi J, et al. Diffusion

analysis with triexponential function in liver

cirrhosis. J Magn Reson Imaging 2013;38(1):

148–53.

152. Yoon JH, Lee JM, Yu MH, et al. Evaluation of hepat-

ic focal lesions using diffusion-weighted MR imag-

ing: comparison of apparent diffusion coefficient

and intravoxel incoherent motion-derived parame-

ters. J Magn Reson Imaging 2014;39(2):276–85.

153. Folpe AL, Lyles RH, Sprouse JT, et al. (F-18) fluoro-

deoxyglucose positron emission tomography as a

predictor of pathologic grade and other prognostic

variables in bone and soft tissue sarcoma. Clin

Cancer Res 2000;6(4):1279–87.

154. Rakheja R, Makis W, Skamene S, et al. Correlating

metabolic activity on18F-FDGPET/CTwith histopath-

ologic characteristics of osseous and soft-tissue

sarcomas: a retrospective review of 136 patients.


155. Kwee TC, Takahara T, Ochiai R, et al. Complemen-

tary roles of whole-body diffusion-weighted MRI

and 18F-FDG PET: the state of the art and potential

applications. J Nucl Med 2010;51(10):1549–58.

156. Tsushima Y, Takano A, Taketomi-Takahashi A, et al.

Body diffusion-weighted MR imaging using high

b-value for malignant tumor screening: usefulness

and necessity of referring to T2-weighted images

and creating fusion images. Acad Radiol 2007;

14(6):643–50.

157. Rakheja R, Demello L, Chandarana H, et al. Com-

parison of the accuracy of PET/CT and PET/MRI

spatial registration of multiple metastatic lesions.


158. Ho KC, Lin G, Wang JJ, et al. Correlation of

apparent diffusion coefficients measured by 3T

diffusion-weighted MRI and SUV from FDG PET/

CT in primary cervical cancer. Eur J Nucl Med

Mol Imaging 2009;36(2):200–8.

159. Wu X, Korkola P, Pertovaara H, et al. No correlation

between glucose metabolism and apparent diffu-

sion coefficient in diffuse large B-cell lymphoma:

a PET/CT and DW-MRI study. Eur J Radiol 2011;

79(2):e117–21.

160. Gu J, Khong PL, Wang S, et al. Quantitative

assessment of diffusion-weighted MR imaging in

patients with primary rectal cancer: correlation

with FDG-PET/CT. Mol Imaging Biol 2011;13(5):

1020–8.

161. Rakheja R, Chandarana H, Demello L, et al. Corre-

lation between standardized uptake value and

apparent diffusion coefficient of neoplastic lesions

evaluated with whole-body simultaneous hybrid

PET/MRI. AJR Am J Roentgenol 2013;201(5):

1115–9.

162. Turner R, Le Bihan D, Maier J, et al. Echo-planar

imaging of intravoxel incoherent motion. Radiology

1990;177(2):407–14.

163. Dietrich O, Biffar A, Baur-Melnyk A, et al. Technical

aspects of MR diffusion imaging of the body. Eur J

Radiol 2010;76(3):314–22.



































































































Documents

Diffusion-Weighted Imaging of the Liver