Quantitative Elastography Methods in Liver Diseases Kennedy Et Al

Reviews and Commentary n STATE OF THE ART 738 Bachir Taouli,MD Claude B. Sirlin, MD Curtis L. Johnson, PhD Cheng William Hong, MD Laurent Castéra, MD Mathilde Wagner, MD, PhD Paul Kennedy, PhD activity foramaximumof1.0 for physicians. thisjournal-basedSA-CME The RSNAdesignates (ACCME)toprovidecontinuingmedicaleducation Medical Education The RSNAisaccreditedbythe CouncilforContinuing Accreditation extent of their participation intheactivity.extent oftheirparticipation Physicians shouldclaim withthe onlythecreditcommensurate q and DigestiveKidneyDiseases(R01DK087877). InstituteofDiabetes Cancer Institute(U01CA172320)andNational R01DK088925, R01DK106419). P.K. andB.T. supportedbyNational of DiabetesandDigestiveKidneyDiseases(R01DK087877, Institute andBioengineering(T32EB005970)National Imaging C.W.H. andC.B.S. InstituteofBiomedical supportedbyNational (e-mail: version acceptedSeptember7. May 3;finalrevisionreceivedJuly20;accepted August2;final Newark, Del(C.L.J.). ReceivedMarch29, 2017;revisionrequested Department ofBiomedicalEngineering, UniversityofDelaware, University ofCalifornia–SanDiego, SanDiego, Calif(C.W.H., C.B.S.); Clichy, France Group, (L.C.);LiverImaging DepartmentofRadiology, Department ofHepatology, UniversityParis-VII, HôpitalBeaujon, Assistance Publique-HôpitauxdeParis, Paris, France (M.W.); of Radiology, SorbonneUniversités, UPMC, HôpitalPitié-Salpêtrière, Mount Sinai, 1470Madison Ave, New York, NY10029;Department (B.T.),and DepartmentofRadiology Icahn School ofMedicineat 1 Accreditation and Designation Statement andDesignation Accreditation n n n able to: After readingthearticle andtakingthetest, thereaderwillbe Learning Objectives: activity, authordisclosuresarelistedattheendofthisarticle. itors, andreviewersforthisactivity.Forjournal-basedCME CME, obtainsigneddisclosurestatementsfromtheauthors,ed- The ACCMErequiresthattheRSNA,asanaccreditedproviderof Disclosure Statement

From the Translational Institute(P.K., andMolecularImaging B.T.)

RSNA, 2018 RSNA, See www.rsna.org/education/search/ry Online SA-CME

measurements usingUSandMRimaging Identify pitfallsandconfoundersofliverstiffness disease inliver Describe thecurrentperformanceofelastography methods Describe thebasicprinciplesofelastography [email protected] AMA PRACategory1Credit Address correspondenceto ). This copyisforpersonaluseonly.Toorderprintedcopies,contact ™ . B.T. Quantitative Evidence andFutureDirections M ethods inLiver through observation of shear-wave propagation in the invasive measurement of tissue mechanical properties search interest inthefield.Elastography enablesthenon limitations, noninvasive methodshave becomeakey re of choice for fibrosis evaluation. Becauseofliver biopsy relevant. Traditionally, liver biopsy has been the method ity ofliver fibrosis, thusliver fibrosis stagingisclinically egies andprognosis differgreatly dependingonthesever liver fibrosis andultimately, cirrhosis. Treatment strat Chronic liver diseasesoftenresult inthedevelopment of Online supplementalmaterialisavailable forthisarticle. q emerging advances willbediscussed. presented. Reliability, reproducibility, failure rate, and most commonetiologiesofchronic liver diseasewillbe elastography techniques. Diagnosticperformanceinthe niques, andresults andlimitationsofUS-MR-based the authors review the technical basis, acquisition tech each with their respective strengths and limitations. Here, aging elastographic methods are commercially available, Ultrasonographic (US)andmagneticresonance (MR)im feature thatcan beexploited by elastographic methods. with increased liver stiffness, providing adiscriminatory tissue ofinterest. Increasing fibrosis stageisassociated

SA 2018 RSNA, E lastography radiology.rsna.org [email protected] D isease: n

Radiology: Volume 286: Number 3—March 2018 Current 1 ------STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

hronic liver diseases are a major fibrosis may be beneficial in monitoring composite scores, such as FibroTest/Fi- cause of morbidity and mortality treatment efficacy, disease progression, broSure (BioPredictive, Paris, France) Cin the United States and world- and in establishing prognosis. (21), the enhanced liver fibrosis, or wide. The most prevalent etiologies of Until recently, liver fibrosis was pri- ELF (Siemens Healthineers, Erlangen, chronic liver diseases include chronic marily assessed with liver biopsy, which Germany), test (22), and FibroMeter hepatitis B virus (HBV) infection, is the reference standard for staging of (Echosens, Paris, France) (23). Though chronic hepatitis C virus (HCV) infec- liver fibrosis and grading of necroin- simple to perform, these tests have tion, nonalcoholic fatty liver disease flammatory changes, by using various limited accuracy in intermediate levels (NAFLD), and alcohol abuse (1). In the semiquantitative scoring systems (8– of fibrosis (24), and they are generally United States, chronic liver diseases 15). The most commonly used scoring considered less accurate than elasto- affect approximately 360 per 100 000 systems include the METAVIR score in graphic methods (25). persons and are the 12th leading cause chronic HBV or HCV infections and the of death, with an estimated mortality Brunt score in nonalcoholic steatohepa- of 12 deaths per 100 000 persons (2,3) titis (NASH) (8,9). All scoring systems Principles of Elastography and a projected financial burden ex- (except the Ishak score) range from F0 Elastography, first coined by Ophir et al ceeding $100 billion annually, mostly to F4, where F0 indicates no fibrosis; (26), describes the noninvasive assess- from NAFLD and chronic HCV infection F1, mild fibrosis; F2, moderate fibrosis; ment of tissue mechanical properties (4,5). Chronic liver diseases can lead F3, advanced fibrosis, and F4, cirrho- such as elasticity, which describes the to liver fibrosis, which is the result of sis. Despite its strengths, liver biopsy resistance to deformation of a tissue to chronic liver injury (6). The end-stage has several drawbacks: liver biopsy is an applied stress. In quantitative elas- of liver fibrosis is cirrhosis, which has relatively invasive and associated with tography methods, the stress is applied potential complications including portal a low rate of complications (approxi- via shear-wave propagation, generated hypertension, liver failure, and hepato- mately 3%) such as pain and bleeding, transiently, for example, via single me- cellular carcinoma (HCC). There is ev- which reduce patient acceptance and chanical impulse, or dynamically, for idence that when the underlying cause limit its suitability for repeated mea- example, via continuous application of is removed, liver fibrosis may regress or surements and disease monitoring. acoustic waves. US and MR elastograph- stabilize (7). Accurate staging of liver Also, biopsy enables the analysis of ic methods are available clinically and only a small portion of the liver, about are summarized in Figure 1 and Table 1. 1/50 000th of the total parenchyma, A more comprehensive overview of the Essentials introducing sampling variability and elastographic methods to be reviewed is nn Elastographic methods are accu- possible diagnostic errors (16,17). In- included in the Appendix E1 (online). rate tools for diagnosing liver ter- and intraobserver variability have Quantitative US elastography fibrosis and cirrhosis in a wide also been suggested as limitations to methods include transient elastogra- range of etiologies. biopsy, which may be linked to the in- phy (TE) and acoustic radiation force nn US-based transient elastography consistency in the definition of some impulse (ARFI) techniques such as is the most validated elasto- pathologic features (8,16,18). Finally, graphic technique. liver biopsy lacks dynamic information nn Acoustic radiation force impulse about the speed of disease progression. https://doi.org/10.1148/radiol.2018170601 All of these limitations make liver bi- elastographic methods are inte- Content codes: grated into clinical US systems opsy an imperfect reference standard. Radiology 2018; 286:738–763 allowing elastography to be per- An important ramification of these limitations is the difficulty in validating formed in routine clinical Abbreviations: examinations. noninvasive tests with use of biopsy as ARFI = acoustic radiation force impulse the reference standard, as the inherent AUC = area under the receiver operating characteristic n n MR elastography offers excellent flaws of biopsy can cause misinterpreta- curve diagnostic accuracy, although it is tion of results. GRE = gradient recalled echo less well validated and less avail- Alternative, noninvasive methods HBV = hepatitis B virus HCC = hepatocellular carcinoma able than US elastographic of evaluating liver health are being de- methods. HCV = hepatitis C virus veloped, such as serum markers and ICC = intraclass correlation coefficient nn Liver stiffness measurement can ultrasonographically (US) and mag- NAFLD = nonalcoholic fatty liver disease be affected by a variety of con- netic resonance (MR) imaging–based NASH = nonalcoholic steatohepatitis founding factors, such as hepatic elastography. Serum markers include pSWE = point SWE inflammation, congestion, chole- simple markers—such as platelet SWE = shear-wave elastography stasis (for all elastographic count, aspartate aminotransferase-to- TE = transient elastography 2D = two-dimensional methods), and steatosis (at least platelet radio index, or APRI (19), FIB- for transient elastography). 4 (20)—and more complex, patented Conflicts of interest are listed at the end of this article.

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 739 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 1

Figure 1: Illustrations of US elastography techniques, including TE (FibroScan, Echosens), pSWE (Virtual Touch Quantification, Siemens Acuson S2000), 2D SWE (Aixplorer, Supersonic Imagine), and MR elastography. Sampling area for each method is depicted by enclosed green area. TE and pSWE have a fixed sampling area size, though pSWE allows the depth and location to be chosen. Two-dimensional SWE has the ability of pSWE sampling area placement with the additional ability to change the size. MR elastography offers (near) full organ coverage. Corresponding example images for each method are also shown. TE = transient elastography, pSWE = point shear-wave elastography, 2D SWE = two-dimensional shear-wave elastography.

Table 1 Comparison of Quantitative Elastography Methods Main Reasons for Failure Method Availability Cost Evidence Liver-sampling Area Region of Interest Placement Reported Parameter or Unreliable Results

Transient Widespread Low Excellent Small Restricted, no guidance Young modulus (kPa) High body mass index elastography validation (M probe), ascites ARFI Moderate Low Moderate, Small (pSWE); Flexible with US guidance; Young modulus (kPa) or High body mass index good medium (2D SWE) recommended 1 cm below wave speed (m/sec) validation liver capsule and , 5 cm from transducer MR elastograhy Limited High Limited Large Large organ coverage Complex shear modulus Liver iron deposition, large validation (kPa) ascites, body mass index*, 3 T (2D GRE)

Note.— ARFI = acoustic radiation force impulse, GRE = gradient recalled echo, pSWE = point shear-wave elastography, SWE = shear-wave elastography, 2D = two-dimensional. * Conflicting evidence reported regarding MR elastography failure in patients with high body mass index.

point shear-wave elastography (pSWE) of the generated shear wave (Figs 1, 2). techniques use focused US “push” and two-dimensional (2D) shear-wave There are several probes available, with pulses to deform internal tissue and elastography (SWE). The FibroScan the M probe being used for standard generate shear waves (27). The nomen- system (Echosens, Paris, France) was examinations and the XL probe intro- clature for ARFI elastography in the the first commercially available TE duced to increase the reliability of TE literature is not standardized. Though system, introduced in Europe in 2003 measurements in overweight patients. pSWE and 2D SWE both utilize ARFI and approved in the United States by The XL probe records the measure- to generate shear waves, pSWE is of- the Food and Drug Administration in ment at a greater depth than does the ten referred to as ARFI elastography in 2013. The FibroScan probe delivers a M probe (35–75 mm vs 25–65 mm), published studies. To avoid confusion, 50-Hz mechanical impulse to the skin with a lower operating frequency (2.5 in this review we use ARFI to describe surface and then measures the velocity MHz vs 3.5 MHz). ARFI elastography the method of wave generation and

740 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 2

Figure 2: Transient elastography images. Left: image in a 39-year-old woman with chronic hepatitis C virus infection with no fibrosis (stage F0) (liver stiffness, 3.2 kPa; M probe). Middle: image in a 59-year-old man with chronic hepatitis B virus infection with stage F2 fibrosis (liver stiffness, 8.7 kPa; M probe), Right: Image in a 57-year-old man with nonalcoholic fatty liver disease with cirrhosis (liver stiffness, 27.0 kPa; XL probe). Liver stiffness measurement (Young modulus, median value of 10 measurements), interquartile range, and median value percentage are automatically calculated. An elastographic image (red box) shows axial displacement in terms of depth (y-axis) and time (x-axis). In stiffer tissues, the shear wave propagates more quickly and produces a steeper time-depth gradient (arrows).

refer to the respective implementations which generate shear waves and image the speed of a shear wave. In MR elas- as pSWE and 2D SWE. them with the same probe, MR elastography, increased wavelength is evi- Originally available clinically with tography requires external hardware to dent in stiffer tissues. An obstacle to Siemens (pSWE, Virtual Touch Quan- generate shear waves in the tissue of direct comparison between techniques tification) and Supersonic Imagine (2D interest. Tissue mechanical properties is the frequency dependence of biologic SWE) systems, ARFI methods are now are quantified through inversion of the tissue. Higher frequency shear waves integrated into clinical systems by other visualized “wave field” into a map of the produce higher stress and strain rates, major vendors such as Philips (28,29), mechanical parameter of interest with- resulting in higher stiffness measure- GE (30), Hitachi (31), Toshiba (32), Es- out the intermediate step of measuring ments (38). This can be problematic aote (33), and Samsung (34) (Figs 3, 4). shear-wave speed (Fig 5). Commercial when comparing US elastographic tech- Two-dimensional SWE has now been MR elastography implementations re- niques, as TE operates at 50 Hz whereas incorporated into Siemens clinical port the shear stiffness of tissue, which ARFI methods operate at frequencies of US systems to complement the pSWE is the magnitude of the complex shear 100–500 Hz (39). Thus, the frequency method (Fig 4). Clinical US elastography modulus, |G*|. MR elastography was dependence, method of measurement, systems report “stiffness” values in terms first described in 1995 by Muthupillai et and parameter reported (wave speed, of the Young modulus (E, in kilopascals), al (36) and was approved by the Food E, or G*) should be considered when others as shear-wave speed (in meters and Drug Administration in 2009. Ini- comparing elastography techniques. per second), and others with options for tially introduced with GE systems, the both. Under simplifying assumptions of technique has since become available incompressibility, shear-wave speed and with Siemens and Philips MR systems. Reliability and Failure Rates of E are related by the following mathemat- Care must be taken when comparing re- Elastographic Methods ical equation: E = 3rc2, where c is the sults between US and MR elastography shear-wave speed and r is the density of due to the different output parameters TE Method tissue, assumed to be that of water. The reported. The failure rate and reliability of TE 2017 European Federation of Societies Liver fibrosis leads to increased were assessed in a study of 13 369 ex- for Ultrasound in Medicine and Biology, stiffness. As shear waves travel through aminations by using the M probe (40). or EFSUMB, guidelines recommend that a tissue, the speed of the wave de- The technique failed in 3.1% of cases; pSWE and 2D SWE measurements be pends on the tissue stiffness (37). In however, unreliable measurements were performed at least 1 cm below the liver stiffer tissues, the shear-wave speed acquired in a further 15.8% of cases. capsule to obtain the best results (35). is greater, enabling estimation of the Body mass index was identified as a In contrast to US elastographic systems, degree of liver fibrosis from measuring significant contributory factor to failed

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 741 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 3

Figure 3: A, Successful and, B, unsuccessful point shear-wave elastographic acquisition (Siemens Acuson S3000) in a 58-year-old man with chronic hepatitis C virus infection and stage F2 liver fibrosis. Unsuccessful measurement (displaying as X.XX m/s) related to poor breath hold. In the successful measurement, wave speed was measured at 1.10 m/sec.

Figure 4

Figure 4: Images obtained with the same system (Siemens Acuson S3000) in a 50-year-old woman with grade 2 steatosis without fibrosis.A, Point shear-wave elastographic image demonstrates placement of fixed-size region of interest in the right hepatic lobe, with measured wave speed of 1.17 m/sec. B, During the same examination, two-dimensional shear-wave elastographic image shows placement of larger size region of interest in the same area, with color elasticity map, and measured wave speed of 1.33 m/sec with interquartile range of 0.21 m/sec. and/or unreliable measurements. The intraclass correlation coefficient (ICC) radiologists (48). Failure rate was low introduction of the XL probe has im- of 0.98 in a cohort of 188 patients with for both methods (5% for 2D SWE proved the reliability of TE in patients chronic HCV in whom two measure- and 1% for pSWE) and intraobserver with NAFLD (41–46). For example, in a ments were performed by two opera- agreement was higher for pSWE than study of 276 patients, reliable measure- tors on the same day (47). 2D SWE (0.915 vs 0.829). Scan-rescan ments were obtained in 73% of patients repeatability of 2D SWE measurements with the XL probe compared with only pSWE and 2D SWE Methods performed on the same day by the same 50% of patients with the M probe (41). The reliability of both pSWE and 2D operator has been shown to be excel- Excellent interobserver variabil- SWE was compared in 79 patients with lent, with ICC of 0.95 and 0.93 for an ity has been reported for TE, with measurements performed by three expert and novice operator, respectively

742 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 5

Figure 5: MR elastography performed by using a two-dimensional gradient-recalled-echo sequence and a two-dimensional inversion algorithm in a 52-year-old woman with advanced liver fibrosis (stage F3) secondary to nonalcoholic steatohepatitis.A, Transverse magnitude image with intravoxel phase dispersion (arrows) present under the actuator (which is not visible on MR images). B, Transverse image with waves visible in liver parenchyma. C, Transverse colorized wave image shows wave propagation through liver parenchyma. D, Transverse gray-scale elastogram. E, Transverse colorized elastogram (0–8-kPa scale). F, Transverse colorized elastogram (0–8-kPa scale) with 95% confidence grid overlaid highlighting areas of reliable liver stiffness measurement. Liver stiffness was increased (5.3 KPa).

(49). However, intraobserver repro- failure rate at 1.5 T (3.5%); however, reproducibility. Similarly, with precise ducibility between measurements per- at 3T the rate of failure was higher definition of region of interest place- formed in the same subject on different (15.3%) (53). In the same study, fail- ment, including all the pixels with con- days revealed ICC values of 0.84 and ure was also significantly associated fidence index higher than 95%, Yasar 0.65 for an expert and novice operator, with iron deposition, the presence of et al (62) found almost perfect inter- respectively. There is further evidence large ascites, and increased body mass and intraobserver reproducibility, with to suggest that operator experience has index (see below) (53). The use of MR ICC higher than 0.97 and Bland-Atman an effect on pSWE measurements (50), elastography in children is also reliable, limits of agreement range lower than thus operators are recommended to be with a recent study reporting a failure 15.9% and 4.2%, respectively. suitably trained. With the introduction rate of 4% with use of the 2D GRE To become a widely accepted of pSWE and 2D SWE into commercial sequence (54). The test-retest repeat- method for diagnosis and staging of fi- US systems by many manufacturers, ability of MR elastography is high (55– brosis, MR elastography must produce interplatform variability may be an 59), with reported ICC of 0.95 (56). consistent results regardless of the issue. The Quantitative Imaging Bio- A recent meta-analysis of 274 patients MR system used. Good interplatform markers Alliance, or QIBA, has formed concluded that a change in stiffness of reproducibility was reported with use a committee tasked with establishing 22% or greater measured at the same of a 2D GRE sequence at 1.5 T, with reproducibility across US elastography site by using the same equipment sig- ICC between 0.82 and 0.99 (62–64). systems (51). nified a true change in stiffness with At 3 T, between-vendor ICC was 0.71 95% confidence (60). Lee et al (61) for the 2D GRE sequence and 0.69 for MR Elastography showed that a large region of interest the 2D spin-echo echo-planar imag- The failure rate of MR elastography is representing approximately 70% of the ing sequence (64). In the same study, low, with the largest series to date re- liver, including the greatest part of the the variance in liver stiffness based porting a failure rate of only 5.6% when liver parenchyma excluding hepatic hi- on technical factors was reported to using a 2D GRE sequence (52). The lar vessels, increased the interobserver be only 0.042 kPa, with correspond- majority of these failures (71%) were reproducibility, while a placement of ing coefficient of variation of 10.7%. attributed to iron deposition. Another a 1-cm region of interest in each liver The reproducibility of spleen stiffness large cohort study found a similarly low segment optimized the intraobserver measurement using MR elastography

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 743 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

has been less well studied, with a study reporting ICC greater than 0.88 (62). The QIBA has formed a committee (65) 0.84 0.92 Specificity 0.87 0.90/0.94 0.96 0.85 0.91 0.80 0.90 0.93 0.83/0.70 0.90 … dedicated to standardizing MR elastog- 0.63 raphy through identifying bias in measurements and identifying a suitable 0.90 0.92 Sensitivity 0.93 0.75/0.84 0.86 0.76 0.87 0.87 0.72 0.94 0.81/0.92 0.77 … phantom for characterization of data 0.92 acquired from different MR elastogra- Stage F4 9.6 phy systems. 9.5 11.3 11.4 Cutoff (kPa) 11.0 11/12.5 14.6 12.8 12.5 12.9 13.2 10.3/7.2 12.9 … In summary, MR elastography and 2D SWE appear to produce the highest rate of successful measurements; how- 0.91 0.92 AUC 0.93 0.95/0.94 0.97 0.89 0.90 0.95 0.90 0.97 0.89/0.91 0.93 0.92 0.87 ever, the introduction of the XL probe has improved the applicability of TE in overweight patients. Reproducibility 0.91 0.92 Specificity 0.85 0.81/0.93 0.85 … 0.80 0.91 … 0.88 0.68/0.54 … 0.76 0.45 is good to excellent among all elastographic techniques. 0.88 0.81 Sensitivity 0.86 0.88/0.68 0.86 … 0.72 0.73 … 0.89 0.88/0.91 … 0.98 0.90

Diagnostic Performance of Stage F3–F4 8.2 8.1 8.1/9.5 9.6 9.6 9.5 9.1 7.9/5.7 7.9 8.2 10.4 Cutoff (kPa) … … Elastographic Methods for Staging … Liver Fibrosis Diagnostic accuracy for liver fibrosis 0.95 0.83 AUC 0.93 0.90/0.89 0.91 … 0.83 0.90 … 0.94 0.87/0.85 … 0.94 staging in chronic liver diseases, includ- 0.86 ing chronic HBV or HCV infections, and NAFLD for select publications across 0.90 0.81 Specificity 0.83 0.88/0.89 0.91 0.63 0.75 0.89 0.34 0.84 0.42/0.37 0.35 0.42 Europe, Asia, and the United States, 0.45 are presented for TE (43,46,66–77) (Table 2), ARFI elastographic methods 0.78 0.78 Sensitivity 0.70 0.74/0.68 0.56 0.68 0.58 0.67 0.90 0.80 0.94/0.92 0.97 0.93 (46,72,77–83) (Table 3), and MR elas- 0.90 tography (38,81,84–94) (Table 4). Pub- Stage F2–F4 7.8 6.9 Cutoff (kPa) 7.2 7.2/7.1 8.8 7.1 8.4 7.1 5.2 7.4 5.8/4.8 5.2 5.8 lications have been selected on the ba- 6.2 sis of reporting of diagnostic accuracy, historical primacy, cohort size, and 0.91 0.78 AUC 0.81 0.87/0.87 0.79 0.76 0.73 0.83 0.76 0.89 0.83/0.80 0.82 0.85 geographic distribution. The TE tech- 0.82 nique has been thoroughly researched and validated for diagnosis of liver fi- † ‡ Success (%) 93 92 92* 88 88 95 74 89 90 67/75 78 85 92 brosis in large cohort studies, mostly in 77 Europe. In the following sections, diagnos- No. of No. Patients 187 613 274 372 643 193 1773 1202 454 193 512 253 102 291 tic performance of US and MR elastographic methods will be discussed for Etiology HBV HBV/HCV HCV HBV HBV/HCV HCV HBV/HCV HCV HBV NAFLD HCV NAFLD HCV the major etiologies of liver disease. NAFLD Summary statements are included to Probe M M M M M M M M M M/XL M M XL provide condensed conclusions. M

Chronic HBV and HCV Infections Region Europe Europe Europe Europe United States Europe Europe Europe Asia Europe/ Asia Europe Asia United States Knowledge of liver fibrosis stage in Europe chronic HBV and HCV infections is beneficial for prognosis, follow-up, and treatment decisions. The combination of powerful direct-acting antivirals recently developed in chronic HCV infec- (68) Marcellin 2009 (70) Cardoso 2012 (73) Ziol 2005 (66) Castera 2011 (71) Afdhal 2015 (74) Castera 2005 (67) Degos 2010 (75) Lupsor Platon 2013 Leung 2013 (72) 2012 (43) Wong Zarski 2012 (69) Loong 2016 (76) Yoneda 2015 (46) Yoneda Cassinotto 2016 (77) Combined failure rate for two-dimensional shear-wave elastography and transient elastography. Transient elastography only data not available. Transient elastography and transient elastography. Combined failure rate for two-dimensional shear-wave for 226 patients reported. Performance Reported Diagnostic Performance of Prospective Transient Elastography Studies for Liver Fibrosis Staging in Chronic HBV or HCV and NAFLD Elastography Transient Reported Diagnostic Performance of Prospective Note.— AUC = area under the receiver operating characteristic curve, HBV = hepatitis B virus, HCV = hepatitis C virus, NAFLD = nonalcoholic fatty liver disease. Successful measurements exclude failures and unreliable results. Successful measurements exclude NAFLD = nonalcoholic fatty liver disease. HCV = hepatitis C virus, HBV = hepatitisAUC = area under the receiver operating B virus, characteristic curve, Note.— for 188 patients reported. * Performance † ‡ tion (95) and the recent increased use First Author and Year of noninvasive tests for liver fibrosis 2 Table

744 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

staging as proposed in the most recent European Association for the Study of the Liver, or EASL, guidelines 0.76 0.92 0.93 0.97 0.83 0.93 0.95 0.72/0.67 Specificity 0.90 on chronic HCV infection (96) have decreased the use of liver biopsy in chronic HCV infection (97). The Ameri- 0.90 0.96 0.78 0.88 0.84 0.97 0.95 0.90/0.90 Sensitivity 0.91 can Association for the Study of Liver

Stage F4 Diseases, or AASLD, guidelines on chronic HBV infection state that staging 1.41 1.88 2.48 10.4 11.9 10.1 10.4 10.5/1.3 Cutoff* 1.82 of liver disease is important to inform therapy decisions and cite the utility of 0.89 0.97 0.86 0.98 0.91 0.98 0.98 0.88/0.84 AUC 0.94 TE for noninvasive staging of fibrosis, especially in excluding advanced fibrosis (98). — 0.995 0.74 0.95 0.89 0.90 0.97 0.71/0.63 Specificity 0.94 TE Technique Several early studies reported excellent — 0.94 0.95 0.97 0.91 0.90 0.92 0.91/0.90 Sensitivity 0.84 diagnostic performance of TE for the

Stage F3–F4 detection of advanced fibrosis and cir- — 1.69 1.34 8.7 9.3 7.9 9.2 8.3/1.15 Cutoff* 1.58 rhosis in chronic HCV infection, with areas under the receiver operating characteristic curve (AUCs) of 0.88– 0.88 0.99 0.90 0.98 0.95 0.93 0.96 0.89/0.84 AUC 0.91 0.99 (37,66,67) (Table 2, Fig 2). Sim- ilar results were subsequently reported by other studies in chronic HCV and 0.25 — 0.78 0.88 0.96 0.92 0.90 0.50/0.36 Specificity 0.91 HBV infections (68,70,71,73,75,99–101), though in some cases the performance of TE was decreased compared with se- 0.90 — 0.82 0.90 0.73 0.85 0.92 0.90/0.90 Sensitivity 0.84 rum markers owing to a high proportion

Stage F2–F4 of unreliable results (69). Several meta- 1.04 — 1.34 7.1 7.9 7.1 7.6 6.3/0.95 Cutoff* 1.21 analyses (102–109) have confirmed the excellent diagnostic accuracy of TE for diagnosing cirrhosis (AUC, 0.93–0.96), 0.81 — 0.85 0.92 0.87 0.88 0.97 0.86/0.77 AUC 0.89 better than that for detecting moderate fibrosis (F2–F4) (AUC, 0.83–0.88),

† ‡ § with cutoffs ranging from 7.0–7.9 kPa 99 98 92 98 98 96 94 80/81 100 Success (%) for the diagnosis of moderate fibrosis (F2–F4) and 11.3–15.6 kPa for the di- 264 128 121 102 454 549 291 No. of No. Patients 274 235 agnosis of cirrhosis (F4) (105,107–109). These results suggest that TE is better at ruling out rather than ruling in liver HBV NAFLD HCV HCV HBV HBV NAFLD Etiology HCV HCV cirrhosis, with negative predictive value greater than 90%. These results have been confirmed in a North American pSWE pSWE SWE SWE SWE SWE SWE/pSWE Method pSWE pSWE study (74). The EASL-Asociación Lati- noamericana para el Estudio del Híga- do guidelines (110) recommend that a Asia United States Europe United States Asia Asia Europe Region Europe Europe combination of TE and serum markers be used to diagnose moderate fibro-

|| sis (F2–F4) in chronic HCV infection. These guidelines were recently vali- = 304) results reported. Scheuer scoring system used. = 304) results reported. n dated in chronic HCV infection (111). The guidelines also recommend TE as a noninvasive method to diagnose fibro- Ye 2012 (80) Ye Cui 2016 (81) 2012 (82) Ferraioli 2015 (46) Yoneda Leung 2013 (72) Zhuang 2016 (83) Cassinotto 2016 (77) Sporea 2011 (78) 2015 (79) Friedrich-Rust Reported for 182. SWE only data not available. Two-dimensional TE. Combined failure rate for 2D SWE and Reported for 226. Training set ( Training Note.— AUC = area under the receiver operating characteristic curve, HBV = hepatitis B virus, HCV = hepatitis C virus, NAFLD = nonalcoholic fatty liver disease, pSWE = point shear-wave elastography, SWE = shear-wave elastography, 2D = two- elastography, SWE = shear-wave elastography, pSWE = point shear-wave NAFLD = nonalcoholic fatty liver disease, HCV = hepatitis C virus, HBV = hepatitisAUC = area under the receiver operating B virus, characteristic curve, Note.— dimensional. failures and unreliable results. Successful measurements exclude * Cutoff values in kilopascals for 2D SWE and meters per second pSWE. † ‡ § || Reported Diagnostic Performance of Prospective pSWE and 2D SWE Studies for Liver Fibrosis Staging in Chronic HBV or HCV NAFLD First Author and Year sis in treatment-naïve chronically HBV- Table 3 Table infected patients (110).

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 745 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

pSWE Technique 0.93 0.9 0.94 0.78 0.95 Spec 0.90 0.96 0.91 0.87 — 0.91 0.86 0.99 ARFI methods have become available more recently than TE, thus are less 0.8 0.97 0.8 0.59 0.91 Sens 0.89 1 1 0.95 — 0.82 1 1 well investigated, with data in chronic HBV (112–118) and HCV (78,79,119– Stage F4 3.67 3.57 4.67 4.15 6.7 Cutoff (kPa) 6.47 4.13 4.21 5.97 — 4.52 4.6 6.87 127) infections showing high accuracy for liver fibrosis staging (Table 3, Fig 3). 0.92 0.98 0.89 0.88 0.97 AUC 0.92 0.998 0.99 0.95 — 0.95 0.97 0.998 For example, in a study of 274 patients with chronic HCV (78), AUCs of 0.91 0.85 0.91 0.91 0.95 0.87 Spec 0.96 0.97 0.87 0.95 0.92 0.83 1 1 and 0.94 were reported for diagnosing stage F3–F4 and cirrhosis, respectively. 0.93 1 0.86 0.91 0.75 Sens 0.78 0.91 0.97 0.92 0.9 0.84 0.87 1 A study in chronic HBV infection (83) also reported excellent diagnostic per- Stage F3–F4 2.78 3.28 3.64 3.62 4.80 Cutoff (kPa) 6.47 3.13 3.32 5.97 4.07 3.60 4.00 5.45 formance, with AUC greater than 0.95 for Scheuer stage S3–S4 and S4. A re- 0.94 0.99 0.92 0.93 0.89 AUC 0.92 0.96 0.97 0.99 0.94 0.92 0.97 1 cent meta-analysis (128) comprising 21 studies (2691 patients) with chronic 0.96 0.91 0.92 0.96 0.85 Spec 0.85 0.91 0.97 0.97 1 0.9 1 0.95 HBV or HCV infections reported AUCs of 0.88 and 0.91 for stage F2–F4 and 0.89 0.95 0.66 0.67 0.87 Sens 0.86 1 0.77 0.91 0.6 0.82 0.89 0.95 F4, respectively. Recent guidelines

Stage F2–F4 from the National Health Service in the 2.57 2.79 3.58 3.62 3.4 Cutoff (kPa) 4.89 2.49 3.18 5.37 3.9 3.5 3.2 4.07 United Kingdom (129) recommend the adoption of pSWE for the diagnosis and 0.97 0.97 0.86 0.89 0.89 AUC 0.92 0.99 0.92 0.98 0.78 0.93 0.99 0.99 monitoring of liver fibrosis in patients with chronic HBV or HCV infections. The National Health Service report sug- † 98 100 99 100 Success (%) 98 94* 99 97 90 96 96 91 91 gested that pSWE has similar or higher performance than TE in diagnosing liver fibrosis. 179 117 126 142 No. of No. Patients 85 141 74 78 42 110 119 132 539 Two-dimensional SWE Two-dimensional SWE is also a highly accurate method in chronically HBV- or HBV/HCV NAFLD NAFLD NAFLD Etiology Mixed Mixed Mixed Mixed Mixed Mixed/obese HCV HBV HBV HCV-infected populations (72,82,130– 132) (Table 3); however, less well P P P P Design P P P P P P R P R studied than pSWE and TE. Two-dimensional SWE has been found to be an equivalent, if not better, diagnostic 3 T 3 T 3 T 3 T Field Strength 1.5 T 1.5 T 1.5 T 1.5 T 1.5 T/3 T 1.5 T 1.5 T 3 T 1.5 T tool than TE in chronic HCV cohorts (82,133). A meta-analysis based on seven 2D SWE studies reported AUC values of 0.91 for stage F2–F4 and 0.95 for cirrhosis (134). Thus, at present, 2D SWE can be used with equivalent GE (3D SE EPI) GE (2D GRE) GE (2D GRE) GE (2D GRE) Manufacturer and Sequence GE (2D GRE) Philips (3D SE) Siemens (2D SE EPI) Siemens (2D GRE) GE/S (2D GRE) GE (2D GRE) GE (2D GRE) GE (2D GRE) GE (2D GRE) diagnostic results to TE; however, further validation is required to establish cutoffs for HCV and HBV populations. Asia United States United States Asia Region United States Europe Europe United States United States United States Asia Asia Asia MR Elastography Given the limited availability and recent clinical use of MR elastography, less published data are available compared with TE and pSWE, with a smaller number of prospective studies, many Shi 2016 (92) Loomba 2014 (93) Cui 2016 (81) Imajo 2016 (94) Yin 2007 (84) Yin Huwart 2008 (85) Asbach 2010 (38) 2011 (86) Wang Dyvorne 2016 (87) Chen 2016 (88) Ichikawa 2012 (89) Shi 2014 (90) Chang 2016 (91) Reported for 270. First Author and Year Reported Diagnostic Performance of MR Elastography for Liver Fibrosis Staging in Mixed-Etiology Cohorts, Chronic HBV or HCV, and NAFLD Chronic HBV or HCV, Reported Diagnostic Performance for Liver Fibrosis Staging in Mixed-Etiology Cohorts, of MR Elastography R = retrospective, P = prospective, NAFLD = nonalcoholic fatty liver disease, HCV = hepatitis C virus, HBV = hepatitis B virus, GRE = gradient recalled echo, GE = General Electric, Note.—AUC = area under the receiver operating characteristic curve, 2D = two-dimensional. 3D = three-dimensional, Spec = specificity, Sens = sensitivity, SE = spin echo, * Reported for 96. † in cohorts of mixed etiologies of liver Table 4 Table disease (52,55,84,86–88,135–138), and

746 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 6

Figure 6: Transverse T2-weighted half-Fourier acquisition single-shot turbo spin-echo, or HASTE, anatomic images (top) and transverse MR elastograms (bottom) depict increasing liver stiffness with increasing fibrosis in patients with chronic hepatitis C virus infection: stage F1 in a 51-year-old man, stage F2 in a 67-year-old man, stage F3 in a 46-year-old man, and stage F4 in a 65-year-old woman. Anatomic images depict no significant liver nodularity in patients with stage F3–F4 fibrosis, while MR elastograms reveal increasing stiffness (yellow and red colored areas). lack of studies including validation co- fibrosis and cirrhosis in chronic HCV interest for clinicians and in terms of horts (Table 4, Fig 6). From the pub- and HBV infections and has been incor- public health perspective. lished studies in chronic HCV or HBV porated in several guidelines for man- infections (58,89–91,139–144), 2D agement of chronic HBV and HCV infec- TE Method GRE MR elastography has shown excel- tions. Emerging data suggest that pSWE The current EASL guidelines for the lent accuracy in diagnosing liver fibrosis methods are equivalent to or possibly management of NAFLD recommend TE or cirrhosis, with AUCs for the diagno- superior to TE in viral hepatitis, with as a noninvasive method for liver fibrosis sis of fibrosis stages F2–F4, F3–F4, and the integration of pSWE in the National assessment and monitoring, while liver F4 of 0.95–0.99, 0.94–1, and 0.92–1, Health Service guidelines in the United biopsy is still recommended to confirm respectively (89–91,140–143). A me- Kingdom. Two-dimensional SWE and advanced fibrosis and cirrhosis (152). ta-analysis of data from patients with MR elastography show promising results The use of TE in patients with NAFLD is chronic HCV and chronic HBV report in chronic HBV and HCV infections; challenging owing to the poor reliability accuracy equivalent to, or slightly lower however, the available data are limited. of the technique in overweight or obese than (in early fibrosis stages), that in In the case of MR elastography, studies patients when the standard M probe is published studies, with AUCs for stage have often included mixed etiologies, used. The range of unreliable measures F2–F4, F3–F4, and F4 of 0.88, 0.94, and with a lack of prospective studies with is large, with reported unreliable and/ 0.92 in HCV infection and 0.94 (stage validation cohorts. All elastographic or failed measurements in 3.8%–50% of F2–F4) and 0.97 (stage F3–F4) in HBV methods have higher accuracy for diag- patients (41,76,153). A meta-analysis of infection (145). Several studies also nosing advanced fibrosis and cirrhosis the TE performance using the M probe showed that necroinflammation may than lower fibrosis stages. in NAFLD (n = 854) (154) reported increase liver stiffness (90,146,147). pooled sensitivity and specificity of 79% Currently, there are limited data on and 75% for F2–F4, 85% and 82% for the performance of 2D spin-echo echo- NAFLD and NASH F3–F4, and 92% and 92% for stage F4. planar imaging and three-dimensional NASH is becoming a widespread prob- Pooled AUC was not reported, though spin-echo echo-planar imaging MR lem in the United States due to the AUC ranges for the included studies elastography in chronic HBV and HCV; increasing prevalence of obesity and were 0.79–0.87 for F2–F4, 0.76–0.98 for however, similar diagnostic accuracy as NAFLD (149,150). Liver fibrosis has F3–F4, and 0.91–0.99 for stage F4. As that with TE and 2D GRE MR elastog- been shown to be the strongest predic- in chronic HBV or HCV infections, TE is raphy has been reported in a few stud- tor of complications in NAFLD patients more accurate in higher fibrosis stages. ies (92,148). (151), which motivates the need for re- The introduction of the XL probe has led In summary, TE is the most vali- liable noninvasive techniques for detec- to more reliable results than with the M dated technique for diagnosing liver tion of liver fibrosis and will be of major probe in overweight or obese patients

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 747 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 7

Figure 7: Top row, images in 43-year-old woman with nonalcoholic steatohepatitis and advanced fibrosis (stage F3) at liver biopsy.A, Transverse PDFF image demonstrates mild steatosis (PDFF, 14.6%). B, Transverse wave image obtained with MR elastography demonstrates increased wavelength (thicker waves) in liver parenchyma. C, Transverse elastogram demonstrates increased liver stiffness (4.33 kPa). Bottom row, images in a 29-year-old woman with nonalcoholic fatty liver disease with no fibrosis (stage F0) at liver biopsy.D, Transverse PDFF image demonstrates mild steatosis (PDFF, 9.2%). E, Transverse wave image obtained with MR elastography demonstrates short wavelengths in the liver (thinner waves) parenchyma. F, Transverse elastogram demonstrates normal liver stiffness (2.22 kPa). PDFF = proton density fat fraction.

(155), with lower stiffness values com- patients (Fig 4). Recently, a prospective (165) reported AUCs of 0.90 or greater pared with the M probe which may ne- study in NAFLD (n = 291) evaluated for the diagnosis of fibrosis stages F3– cessitate revalidated cutoffs (42–45). 2D SWE, pSWE, and TE using the M F4 and F4, with associated cutoffs of probe (77). When accounting for un- 3.77 kPa and 4.09 kPa, respectively. As pSWE Method reliable results, all techniques had a they were derived from a meta-analysis, A preliminary study in NAFLD patients similar amount of successful measure- these cutoffs are probably the most ap- (156) found that pSWE performed very ments (80%, 77%, and 81% for 2D plicable at this time, but further valida- well when diagnosing fibrosis stage SWE, TE, and pSWE, respectively). tion studies are needed. There is also F3–F4 and F4, with AUC greater than Two-dimensional SWE performed bet- evidence that MR elastography may be 0.97 (Fig 4). Subsequent studies have ter than pSWE for the diagnosis of able to differentiate NASH and simple reported similar high accuracy in di- moderate fibrosis (stage F2–F4), with steatosis in NAFLD patients (161), with agnosing fibrosis and differentiating AUC of 0.85 versus 0.76. In a prospec- a reported AUC of 0.93, but this needs NASH from simple steatosis (157–159). tive study of overweight patients with further confirmation. In a comparative study of pSWE and mixed etiologies, 2D SWE of the right TE (with the M and XL probes) (160), lobe performed similarly to TE with the no significant difference was found, al- XL probe, with AUC greater than 0.90 Other Etiologies of Liver Disease though pSWE achieved a significantly for fibrosis stages F2 or greater (46). TE has also been applied in the study higher reliability rate. A more recent of autoimmune liver disease, with ex- study found that while pSWE reported MR Elastography cellent diagnostic performance (166), AUCs greater than 0.85 for discriminat- A small number of retrospective or though several studies have reported ing fibrosis stages F2 or greater, it was prospective studies have focused on that acute inflammation as a result of outperformed by MR elastography (81). NAFLD and/or NASH populations autoimmune hepatitis may affect liver (25,81,93,94,161–164), with reported stiffness (167,168). TE has also been Two-dimensional SWE AUCs greater than or equal to 0.86 shown to be an accurate method for Two-dimensional SWE is less well val- (Table 4; Figs 5, 7). A recent meta-anal- staging fibrosis in primary biliary cir- idated than pSWE and TE in NAFLD ysis of nine studies with 232 patients rhosis (169–171), primary sclerosing

748 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

cholangitis (172,173), and alcoholic higher risk of developing complications, (38.5% among patients with baseline liver disease (174–176). pSWE methods with one study estimating a 10-fold inliver stiffness values . 25 kPa, com- have been applied in autoimmune liver crease in complications (172). In a me- pared with 0.4% among patients with disease and alcoholic liver disease in ta-analysis of six studies that reported values  10 kPa). a small number of studies (177–181). hepatic decompensation as an outcome, pSWE.—So far, published data Two-dimensional SWE has also been each unit increase in liver stiffness was show similar prognostic ability between applied in alcoholic liver disease, with associated with a 7% increased de- pSWE and TE in predicting hepatic de- similar accuracy to TE reported (176). compensation risk (200). TE can also compensation (220). Due to the inte- Excellent diagnostic accuracy of MR help diagnose portal hypertension, with gration into conventional US systems, elastography has been reported in au- another meta-analysis reporting excel- pSWE methods are more suitable for toimmune hepatitis (182) and primary lent diagnostic performance of TE for measuring spleen stiffness than is TE. sclerosing cholangitis (183); however, diagnosing the presence of clinically A study of 393 patients with median the method has not been applied in the significant portal hypertension (defined follow up of 44 months found that liver study of alcoholic liver disease thus far. by hepatic venous pressure gradient  and spleen stiffness measured with In summary, patients with NAFLD 10 mm Hg), with an area under the hi- pSWE were associated with decom- and NASH are more likely to be over- erarchical summary receiver operating pensation (hazard ratios of 2.53 and weight, which is an important factor characteristic curve (HSROC) of 0.93 16.58 per unit increase in stiffness, to consider for elastography measure- (201). Scores combining liver stiffness respectively) (221). In patients with ments. The TE M probe is prone to un- with platelet count and spleen diameter chronic HCV infection, pSWE outper- reliable results and/or increased risk of at US, referred as LSPS (for liver stiff- formed serum markers in the predic- failure in these subjects, and this has ness spleen-diameter-to-platelet-ratio tion of esophageal varices (AUC, 0.89 been improved with the XL probe. In score) (202) or portal hypertension risk vs 0.75) (222). Diagnostic performance these subjects, ARFI methods are less score (203), have also been proposed was also excellent for the diagnosis of susceptible to failure than TE. MR elas- to increase the diagnostic accuracy. high-risk esophageal varices, with AUC tography is the most robust technique The performance of TE in predict- of 0.87 versus 0.74 for pSWE and se- in overweight or obese patients, with ing the presence and size of esophageal rum markers, respectively. Few pre- reported high accuracy for fibrosis stag- varices based on liver stiffness mea- liminary studies have evaluated pSWE ing, although published data are still surements is also promising. A meta- for the prediction of HCC development limited. analysis of 12 studies (2049 patients) (223,224), indicating that liver stiffness (201) found the HSROC of TE to be measured with pSWE is a significant 0.84 for diagnosing esophageal varices. predictor of HCC development. More Additional and Evolving Applications When evaluating the predictive value studies are needed to validate these of TE for diagnosing large esophageal results. Assessment of Degree of Portal varices (in nine studies comprising Two-dimensional SWE.—Liver and Hypertension, Risk of Hepatic 2168 patients), the HSROC was 0.78. spleen stiffness measured with 2D SWE Decompensation and HCC The Baveno VI consensus workshop have also been shown to correlate with Liver cirrhosis can be further catego- recommendations (204) state that liver hepatic venous pressure gradient mea- rized into compensated or decompen- stiffness measured with TE may be use- surement (225,226). In a cirrhotic pop- sated cirrhosis. Decompensated cirrho- ful for classifying patients with portal ulation (227), 2D SWE outperformed sis is primarily diagnosed on the basis hypertension. Spleen stiffness has been TE for diagnosing clinically significant of the presence of variceal bleeding and also suggested as a marker of severity portal hypertension (AUC, 0.87 vs 0.78, ascites and is associated with a signifi- of portal hypertension (205) with a rea- respectively). Two-dimensional SWE cantly increased risk of mortality (184). sonable accuracy (AUC, 0.78–0.90) in spleen stiffness measurements have Recent data suggest that liver and predicting the presence of esophageal been significantly associated with the spleen stiffness may represent potential varices (206–208). However, the mea- presence of esophageal varices (228), biomarkers for hepatic decompensa- surement of spleen stiffness is difficult though a high failure rate in spleen mea- tion and HCC risk. with TE and requires concomitant con- surements (~ 30%) is a limitation. There TE.—Many studies have shown the ventional US guidance. is little published data on the ability of ability of baseline liver stiffness mea- A correlation between liver stiff- 2D SWE to predict HCC development. surement (172,185–199) to help pre- ness measured by TE and the risk of In one retrospective study (229), the dict hepatic decompensation in patients developing HCC has been reported by authors noted significantly higher liver with chronic liver disease. Two studies several longitudinal prospective stud- stiffness in patients with HCC than those also looked at the evolution of liver stiff- ies (187,195,209–219). For example, without; however, prospective studies ness values over time (172,197) and a large prospective Japanese study (n are required for validation. found that patients with increasing liver = 866) (214) reported increased cumu- MR elastography.—To date, few stiffness (1–1.5 kPa per year) were at lative incidence of HCC within 3 years studies have investigated the role of

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 749 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 8 surveillance for complications such as HCC. Several studies have reported reduction in liver stiffness of up to 35% (243) following direct-acting antiviral therapy as measured by TE (243–248) and ARFI methods (249,250). The clinical significance of reduced liver stiffness following sustained virological response has not been established, as few longitudinal studies incorporating biopsy and noninvasive tests have been performed. One study investigated the longitudinal changes in liver stiffness with paired liver biopsies in HCV/HIV- coinfected patients treated with antiret- roviral therapy and anti-HCV drugs (in a portion of the population). Patients with progressing fibrosis were found to have significantly increased liver stiffness at 3 years after baseline, while liver stiffness was unchanged or reduced in Figure 8: A, Transverse T2-weighted half-Fourier acquisition single-shot turbo spin-echo, or stable patients (251). There is no pub- HASTE, MR anatomic image with arrows indicating actuator position and, B, transverse stiffness lished MR elastography study assessing map in a 27-year-old healthy woman with normal liver stiffness (2.1 kPa) and spleen stiffness changes in liver stiffness after antiviral measured at 4.3kPa. C, Transverse anatomic image with arrows indicating actuator position and, therapy. Further studies are required to D, transverse stiffness map in a 61-year-old female patient with cirrhosis (secondary to chronic establish the utility of noninvasive tests hepatitis C virus infection) and clinically significant portal hypertension (hepatic venous pressure in the longitudinal monitoring of HCV gradient of 15 mmHg) demonstrate elevated liver stiffness (7.5 kPa) and spleen stiffness (9.9 kPa). patients undergoing antiviral therapy.

HCC Characterization MR elastography in predicting hepatic beneficial in predicting risk of decom- decompensation (183,230,231). In a pensation and diagnosis of portal hy- Liver lesion characterization is beyond retrospective study of 266 patients with pertension, more studies are required the scope of this review. Briefly, there primary sclerosing cholangitis, Eaton et to confirm the findings. are some data assessing the role of US al (183) found that liver stiffness was A small number of studies to date and MR elastographic methods in quan- a significant predictor of risk of de- have evaluated MR elastography as a tifying tumor stiffness for the purpose of compensation (hazard ratio, 1.24–1.30 predictive tool for the development of liver lesion characterization (252–266), per unit increase in liver stiffness). MR HCC, with conflicting data (239,240). with a trend toward increased stiffness in elastography measurement of predic- While one retrospective study (240) re- malignant lesions, such as HCC (Fig 9). tive markers of portal hypertension is ported elevated liver stiffness in patients A recent study reported higher tumor an emerging field of research (232– with HCC compared with those with- stiffness values in well or moderately 236). The cross-sectional imaging vol- out HCC, another study did not (239). differentiated HCCs compared with ume available with MR elastography en- Thus, more data are needed to predict poorly differentiated HCCs (267). An- ables simultaneous acquisition of liver risk of HCC with MR elastography. other preliminary study has reported and spleen MR elastography data via a significant correlation between tumor the use of an additional actuator placed Monitoring of Chronic HCV Infection after stiffness and degree of tumor necrosis on the left side (62,237) and may allow Antiviral Therapy and enhancement in HCC after local- the development of composite diagnos- The introduction of direct-acting an- regional therapy (268). A drawback of tic tests by using both measurements as tivirals has revolutionized therapy in MR elastography investigation of tumors has been implemented with TE (238). chronically HCV-infected patients, is the limited spatial resolution and cov- Spleen stiffness measured by means of with excellent cure rates reported in erage of current 2D MR elastography MR elastography has been associated most genotypes (241) and subsequent implementations (269). Nonlinear inver- with the presence of esophageal vari- reduction in liver transplant waiting sion algorithms (270) paired with three- ces (233,235) and found to correlate lists (242). Determination of residual dimensional MR elastography (265) may with hepatic venous pressure gradient fibrotic burden in patients who have help to address the problem. (Fig 8) (234). Though the available data achieved sustained virological response In summary, the available data sug- suggest that MR elastography may be is important for both prognosis and gest that elastographic techniques may

750 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 9 example, liver stiffness values increase after meal intake (59,281–286); therefore, elastographic examinations should be performed after fasting for at least 2 hours (110), though fasting for 4–6 hours prior to measurement has also been recommended (39). Similarly, cholestasis has been shown to cause increased liver stiffness in TE (287), pSWE (288), and MR elastography (183). A further difficulty in establishing cutoff values for fibrosis staging is Figure 9: Transverse T2-weighted half-Fourier acquisition single-shot turbo spin-echo, or the influence of the underlying etiology HASTE, anatomic image (left) and transverse MR elastogram (right) in a 59-year-old man with of liver disease on measured stiffness chronic hepatitis C virus infection and infiltrative hepatocellular carcinoma in right hepatic values; for example, the cutoffs for pre- lobe (arrows). MR elastography demonstrates increased stiffness (7.7 kPa) compared with dicting esophageal varices in patients background liver parenchyma (3.2 kPa). Another hepatocellular carcinoma nodule is present in with cirrhosis using TE is higher in left lateral hepatic lobe (arrowheads), also demonstrating increased stiffness. those with alcoholic cirrhosis than in those with liver cirrhosis of viral etiol- be viable methods for prediction of (272). MR elastography has also been ogy (289). The cause of this variance portal hypertension and hepatic decom- compared with 2D SWE in a mixed- has not been conclusively established. pensation, although further research is etiology cohort (138), with comparable Factors such as alanine aminotransfer- needed to understand how the results diagnostic accuracy for both techniques ase levels have been suggested (290); should be used to inform patient care. in staging fibrosis. A recent study used however, the inherent morphologic Spleen stiffness appears to be a useful MR elastography as the reference stan- differences in collagen distribution biomarker for portal hypertension and dard and found the 2D SWE and MR caused by increases in myofibroblasts, prediction of esophageal varices. Fi- elastography measurements to be well which are associated with the various nally, TE has been applied in the mon- correlated (273). MR elastography has etiologies of fibrotic liver disease, could itoring of patients undergoing antiviral also been shown to outperform serum also be a reason for the discrepancy therapy in chronic HCV; however, the markers (25,87,89,141,143,162), mor- in cutoff values, even for livers with utility of the technique in this role has phologic features (137), and diffusion the same “stage” of fibrosis (291,292). yet to be established. measurements (86,87,266,274–276). Also, cutoff values are generally estab- When evaluating US elastographic lished on the basis of receiver operating methods alone, a meta-analysis of 13 characteristic analysis of a single-study Comparison of MR Elastography to US studies including 1163 patients found population and so are affected by the Elastography and Other Noninvasive that pSWE had a similar predictive prevalence and severity of fibrosis and Tests value as that of TE for advanced fibrosis cirrhosis in that population (35). Elas- A few comparative studies investi- and cirrhosis while producing a higher tography methods are useful as tools to gating the diagnostic accuracy of MR rate of reliable measurements (277). generally stratify patient risk; however, elastographic and US elastographic Studies comparing TE to pSWE (278) intermediate fibrosis stages can be dif- methods have been published. Though and 2D SWE (72,82) have found ARFI ficult to delineate. The Society of Radi- MR elastography was generally found methods to provide similar or superior ologists in Ultrasound consensus report to be superior to TE in diagnosing fi- diagnostic performance to TE. In com- of 2015 (39) highlighted the overlap of brosis in mixed cohorts (85,87,271) parisons of all three methods, pSWE, pSWE shear-wave speeds at intermedi- and NAFLD patients (94,164), other 2D SWE, and TE (77,279,280), 2D ate fibrosis levels (sourced from a meta- studies have found both techniques to SWE was the slightly superior method analysis [293]) and therefore suggests perform similarly (88,148). Less litera- for staging fibrosis (77), with vari- using thresholds to define population ture on the comparison of MR elastog- able reliability compared with pSWE groups, which are unlikely to require raphy with ARFI methods is available. (279,280). follow-up (stage F0–F2), those at high A meta-analysis assessing the diag- risk (some F3 and F4), and those in be- nostic accuracy of pSWE (15 studies, tween who may require further testing, Limitations of Elastographic 2128 patients) and MR elastography including MR elastography, to inform Techniques (11 studies, 982 patients) for staging treatment decisions. fibrosis found that MR elastography is Although each elastographic method Breathing motion may also affect more accurate than pSWE, particularly has its own limitations, some draw- US elastography and MR elastography in diagnosing early stages of fibrosis backs apply to all techniques. For measurements. In US elastography,

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 751 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

taking a deep breath or using the Val- decompensated liver cirrhosis for prog- There is conflicting evidence on salva maneuver can change liver stiff- nostic reasons. TE is not suited for the effect of body mass index on MR ness (294,295). The 2017 EFSUMB spleen measurements owing to the need elastography measurements: A recent guidelines recommend that measure- for external guidance from another US study found that body mass index was ment be performed during breath hold, system. The risk of overestimating liver not a contributing factor in failure (88), while avoiding deep inspiration (35). stiffness values has been reported, with but found waist circumference to be a Also, it has been recently suggested other confounding factors including significant factor of failure. In contrast, that inspiration statically deforms the alanine aminotransferase flares (298– a recent large retrospective study inves- liver, which is expected to alter the ob- 300), congestive heart failure (301), tigating the cause of MR elastography served stiffness due to nonlinearity in excessive alcohol intake (302–304), and failure using a 2D GRE sequence (53) elasticity of biologic tissues (296). In acute viral hepatitis (298,305). Some found that body mass index, iron de- 2D GRE MR elastography, a four-sec- work has been done to establish cut- position, massive ascites, and use of 3 tion acquisition generally requires four offs that account for these confounding T were significantly associated with MR breath holds of approximately 15 sec- factors (306), though further validation elastography failure (Fig 11). The over- onds each. Three-dimensional MR elas- is required. The influence of steatosis all failure rate was low (3.5%) at 1.5 tography acquisitions also require that is still a matter of debate with con- T though it increased to 15.3% at 3 T, multiple breath holds be performed. flicting results, some studies suggest a likely due to increased T2* relaxation For accurate determination of liver detrimental effect (307,308), whereas at higher field strength. Other potential stiffness over the liver volume, breath others do not (309,310). In summary, causes of failure include poor actuator holding should be performed in expi- US elastographic techniques need to placement, coupling to the body, or ration to minimize positional changes be performed by using a standardized tube disconnection, which require the between breath holds. protocol and with critically interpreted examination to be repeated, and ab- Because the liver is surrounded by results, taking confounding factors into normal physiology. MR elastography is a nonelastic envelope (Glisson capsule), account (110). also costlier and less available than US additional space-occupying tissue ab- elastography. normalities, such as edema, inflamma- Limitations of MR Elastography tion, or congestion, can interfere with Although considered a highly accurate measurements of liver stiffness, inde- technique, MR elastography has several New Technical Developments pendently of fibrosis. A further con- limitations. The primary drawback of sideration when utilizing elastographic liver MR elastography is the sensitivity Measurement of Liver Steatosis with TE methods is the additional cost of the of 2D GRE sequence to iron deposition. A more recent application of TE is the examination above standard clinical The short T2* time of the liver affected controlled attenuation parameter (CAP) examinations. In the United States, TE by iron deposition means that signal- (313). CAP, which is available on both has recently become a reimbursable to-noise ratio from a standard GRE M and XL probes, estimates the attenu- medical examination with the creation sequence is too low, and thus unable ation of the US signal in units of dBm21 of a Current Procedural Terminology to resolve wave propagation (311). This and is used as a method to grade steato- (CPT) code 91200 for the procedure has been addressed by the introduction sis. A recent meta-analysis of a mixed- (297). A CPT code 0346T for pSWE of spin-echo and spin-echo echo-planar etiology cohort using the M probe (314) and 2D SWE methods can be added to imaging–based sequences, which are reported excellent accuracy for detect- the regular US code. MR elastography, instead primarily sensitive to T2 relax- ing steatosis based on histopathologic the most expensive method, has not yet ation and thus provide higher signal-to- findings. A small pilot study using the XL been granted a CPT code, but its in- noise ratio even in slightly longer echo probe (315) found that performance was creasing use may motivate the introduc- times (312) (Fig 10). A study comparing similar between the M and XL probes tion of one. More specific limitations 2D GRE and spin-echo echo-planar se- for detecting liver fat. Further validation for US elastography and MR elastogra- quences found both sequences produced in large cohorts is required to determine phy are discussed below. consistent liver stiffness measurements, the performance of CAP, particularly with the spin-echo echo-planar sequence with the XL probe. A benefit of MR im- Limitations of US Elastography Methods producing reliable results in subjects in aging when assessing NAFLD patients is Two-dimensional SWE and pSWE can whom the GRE sequence failed due to the high accuracy of liver fat quantifica- be performed with one probe in all iron deposition and larger reliable re- tion using advanced confounder-correct- patients, independent of body weight, gions of interest (311). Another study ed chemical shift–encoded methods now as the region of interest can be posi- found 2D GRE and spin-echo echo- available with all scanner manufacturers tioned manually at different depths in planar MR acquisitions to produce rea- (316), which can be combined with liver the liver. As compared with TE, as- sonably consistent results at 1.5 T and stiffness measurement for a compre- cites is not a limitation for ARFI US 3.0 T with ICCs ranging 0.73–0.9 across hensive examination of liver health (Fig methods, enabling its performance in manufacturers (64). 6). In the study by Imajo et al (94), the

752 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 10

Figure 10: Images in a 61-year-old man with cirrhosis secondary to chronic hepatitis C virus infection and secondary hemosiderosis causing failure of two-dimensional gradient-recalled-echo (GRE) MR elastography (MRE) at 1.5 T. The shortened liver T2* (4.7 msec) due to iron deposition causes low signal-to-noise ratio, with disorganized wave propagation pattern and no areas of reliable stiffness measurement. Two- dimensional echo-planar imaging (EPI) sequence performed during the same MR imaging examination is less sensitive to T2* effects, allowing successful wave propagation and liver stiffness measurement (5.6 kPa). combination of MR elastography and which has been shown to be sensitive accuracy for diagnosing F3–F4 fibrosis in liver fat quantification outperformed to pressure-related changes (318) and 2D and three-dimensional MR elastog- TE and CAP for staging fibrosis and may have applications in the diagnosis of raphy at 60 Hz vibration frequency was fat quantification, respectively. Further portal hypertension, are still being eval- not significantly different. Three-dimen- comparison studies are required be- uated to establish clinical benefit. The sional MR elastography data acquired at tween MR-based fat quantification and acquisition of all three motion directions 40 Hz showed improved diagnostic accu- elastography and US-based TE and CAP also addresses the issue of artificially in- racy with a significantly higher AUC than for combined staging of fibrosis and creased wavelengths due to oblique 2D 2D measurements (AUC, 0.98 vs 0.92). steatosis. waves violating the planar wave assump- Three-dimensional spin-echo echo- tion (319). Further details on three-di- planar imaging MR elastography failure Three-dimensional MR Elastography mensional MR elastography are included rate has been reported as lower than Though the acquisition of all three di- in the Appendix E1 (online). that of 2D GRE MR elastography, as the rections of motion is not a new devel- A comparison of the diagnostic accu- spin-echo echo-planar imaging sequence opment (317), advances in inversion racies of 2D GRE and three-dimensional is expected to perform better in hepatic algorithms and the increasing availabil- spin-echo echo-planar imaging MR elas- iron deposition (92). Spleen stiffness ity of research three-dimensional MR tography in 73 patients with chronic has also been assessed with three-di- elastography imaging sequences has liver disease found 2D and three-dimen- mensional MR elastography (235), with made the technique more accessible. sional sequences to perform similarly liver stiffness and spleen stiffness signif- Three-dimensional MR elastography (320). However, three-dimensional MR icantly associated with the presence of enables the determination of additional elastography parameter results were esophageal varices. parameters compared with 2D owing significantly lower than those of 2D to the acquisition of the full wave field MR elastography. A similar result was Multifrequency MR Elastography and fewer assumptions about the ma- reported from a study of patients with Generally, MR elastography examina- terial model during inversion. These NAFLD performed at 60 Hz and 40 Hz tions are performed by imaging shear parameters, such as volumetric strain, vibration frequencies (163). Diagnostic waves at a single frequency (typically

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 753 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

Figure 11

Figure 11: Images in a 57-year-old man with decompensated cirrhosis (secondary to chronic hepatitis C virus infection) and large ascites causing two-dimensional gradient-recalled-echo MR elastography failure. A, Transverse T2-weighted half-Fourier acquisition single-shot turbo spin-echo, or HASTE, image shows cirrhotic liver (liver contour outlined in white) and large ascites (arrows). B, Transverse wave image shows propagation in subcutaneous fat and fluid but disrupted waves in liver parenchyma, resulting in no reliable areas of stiffness measurement on, C, transverse elastogram. at 60 Hz). The stiffness of tissue is de- of liver fibrosis in a wide range of eti- related to the present article: disclosed no relevant relationships. Activities not related to the pendent on the frequency of the imaged ologies. Interpretation of results should waves, so examinations at a frequency present article: consultancy fees from Olea Med- take into account potential confounding ical. Other relationships: disclosed no relevant other than 60 Hz will result in a dif- factors of liver stiffness measurements, relationships. L.C. Activities related to the pre- ferent stiffness measurement. The use pitfalls, and technical limitations. MR sent article: disclosed no relevant relationships. Activities not related to the present article: lec- of multifrequency MR elastography, elastography has equivalent to slightly where acquisitions acquired at multiple ture fees from Echosens. Other relationships: better diagnostic accuracy than TE and disclosed no relevant relationships. C.W.H. dis- frequencies are performed, may lead ARFI methods, while providing stiff- closed no relevant relationships. C.L.J. disclosed to the development of parameters that ness measurement over a larger area no relevant relationships. C.B.S. Activities related to the present article: disclosed no relevant are independent of frequency through of the liver; however, the method re- viscoelastic modeling (38,321,322) or relationships. Activities not related to the pre- quires wider validation, and the higher sent article: grants from Gilead, GE Healthcare, via analysis of the regression line of cost and limited availability may limit Siemens, Boehringer Ingelheim Pharma; speak- ing services for GE Healthcare, Bayer; member stiffness and frequency (323). Another adoption worldwide. In liver referral application of multifrequency data is of scientific committee/advisory board for Bayer centers performing a large number of and Advanced MR Analytics (AMRA); and lab combining the wave fields from each MR imaging examinations, it is feasible service agreements with Gilead, Shire, Virtu- frequency to improve the resulting elas- to incorporate MR elastography into alscopics, Intercept. Other relationships: disclosed no relevant relationships. B. T. Activities togram. This is achieved by accounting the standard imaging protocols to pro- for areas of low displacement and wave related to the present article: disclosed no rel- vide a fibrosis-staging tool. The weight evant relationships. Activities not related to the nodes that are present in each wave of published data on TE has allowed present article: grants from Guerbet and Bayer, field, but the location of which varies the establishment of measurement cut- equipment support from Siemens. Other relationships: disclosed no relevant relationships. depending on frequency (324–326). A offs for most etiologies. ARFI methods downside of the acquisition of multiple have shown similar diagnostic ability References frequencies is the increased imaging to TE, and it is reasonable to assume time, as each additional frequency in- that once sufficient data have been ac- 1. Tsochatzis EA, Bosch J, Burroughs AK. Liver crementally increases the imaging time cirrhosis. Lancet 2014;383(9930):1749– quired to fully validate ARFI methods 1761. limiting clinical adoption, and the asso- they will also become a recommended 2. Kim WR, Brown RS Jr, Terrault NA, El-Se- ciated challenges with increased wave noninvasive measurement tool for stag- attenuation at higher frequencies. Thus, rag H. Burden of liver disease in the United ing of liver fibrosis. The emergence of States: summary of a workshop. Hepatology the diagnostic benefit of multifrequency advanced techniques such as three-di- 2002;36(1):227–242. over single-frequency MR elastography mensional MR elastography– and US- must be established. 3. Centers for Disease Control and Prevention. based controlled attenuation parameter Chronic liver disease and cirrhosis. http:// measurement may increase the accuracy www.cdc.gov/nchs/fastats/liver-disease. Conclusion of fibrosis and steatosis staging in liver htm. Accessed February 24, 2017. disease, although more data are needed. 4. Razavi H, Elkhoury AC, Elbasha E, et al. US and MR elastographic techniques Chronic hepatitis C virus (HCV) disease have developed into accurate methods Disclosures of Conflicts of Interest: P.K. dis- burden and cost in the United States. Hepa- for quantitative, noninvasive diagnosis closed no relevant relationships. M.W. Activities tology 2013;57(6):2164–2170.

754 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

5. Younossi ZM, Blissett D, Blissett R, et al. of fibrosis in HCV infection. comparison Twice compared to 2D-shear wave elastog- The economic and clinical burden of non- with liver biopsy and fibrotest. Hepatology raphy with supersonic imaging. Ultraschall alcoholic fatty liver disease in the United 2007;46(1):32–36. Med 2016;37(S 01):PS4_08. States and Europe. Hepatology 2016;64(5): 21. Imbert-Bismut F, Ratziu V, Pieroni L, 34. Piscaglia F, Salvatore V, Mulazzani L, et al. 1577–1586. et al. Biochemical markers of liver fi- Differences in liver stiffness values obtained 6. Sebastiani G, Gkouvatsos K, Pantopoulos K. brosis in patients with hepatitis C virus with new ultrasound elastography machines Chronic hepatitis C and liver fibrosis. World infection: a prospective study. Lancet and Fibroscan: a comparative study. Dig J Gastroenterol 2014;20(32):11033–11053. 2001;357(9262):1069–1075. Liver Dis 2017;49(7):802–808. 7. Calvaruso V, Craxì A. Regression of fibrosis 22. Rosenberg WM, Voelker M, Thiel R, et al. 35. Dietrich CF, Bamber J, Berzigotti A, et al. after HBV antiviral therapy. Is cirrhosis re- Serum markers detect the presence of liver EFSUMB Guidelines and Recommendations versible? Liver Int 2014;34(Suppl 1):85–90. fibrosis: a cohort study. Gastroenterology on the Clinical Use of Liver Ultrasound 2004;127(6):1704–1713. Elastography, Update 2017 (Long Version). 8. Intraobserver and interobserver variations Ultraschall Med 2017;38(04):e16–e47. in liver biopsy interpretation in patients 23. Calès P, Oberti F, Michalak S, et al. A with chronic hepatitis C. The French META- novel panel of blood markers to assess 36. Muthupillai R, Lomas DJ, Rossman PJ, VIR Cooperative Study Group. Hepatology the degree of liver fibrosis. Hepatology Greenleaf JF, Manduca A, Ehman RL. Mag- 1994;20(1 Pt 1):15–20. 2005;42(6):1373–1381. netic resonance elastography by direct visual- ization of propagating acoustic strain waves. 9. Brunt EM, Janney CG, Di Bisceglie AM, 24. Parkes J, Guha IN, Roderick P, Rosenberg Science 1995;269(5232):1854–1857. Neuschwander-Tetri BA, Bacon BR. Nonalco- W. Performance of serum marker panels for holic steatohepatitis: a proposal for grading liver fibrosis in chronic hepatitis C. J Hepa- 37. Sandrin L, Fourquet B, Hasquenoph JM, et and staging the histological lesions. Am J tol 2006;44(3):462–474. al. Transient elastography: a new noninva- Gastroenterol 1999;94(9):2467–2474. sive method for assessment of hepatic fibro- 25. Cui J, Ang B, Haufe W, et al. Comparative sis. Ultrasound Med Biol 2003;29(12):1705– 10. Bravo A, Sheth S, Chopra S. Liver biopsy. N diagnostic accuracy of magnetic resonance 1713. Engl J Med 2001;344(7):465–500. elastography vs. eight clinical prediction rules for non-invasive diagnosis of advanced 38. Asbach P, Klatt D, Schlosser B, et al. Vis- 11. Rockey DC, Caldwell SH, Goodman ZD, fibrosis in biopsy-proven non-alcoholic fatty coelasticity-based staging of hepatic fibrosis Nelson RC, Smith AD; American Associa- liver disease: a prospective study. Aliment with multifrequency MR elastography. Radi- tion for the Study of Liver Diseases. Liver Pharmacol Ther 2015;41(12):1271–1280. ology 2010;257(1):80–86. biopsy. Hepatology 2009;49(3):1017–1044. 26. Ophir J, Céspedes I, Ponnekanti H, Yazdi 39. Barr RG, Ferraioli G, Palmeri ML, et al. 12. Bedossa P, Carrat F. Liver biopsy: the best, Y, Li X. Elastography: a quantitative method Elastography assessment of liver fibrosis: not the gold standard. J Hepatol 2009; for imaging the elasticity of biological tis- Society of Radiologists in Ultrasound Con- 50(1):1–3. sues. Ultrason Imaging 1991;13(2):111–134. sensus Conference Statement. Radiology 13. Knodell RG, Ishak KG, Black WC, et al. 2015;276(3):845–861. 27. Sarvazyan AP, Rudenko OV, Swanson SD, Formulation and application of a numerical Fowlkes JB, Emelianov SY. Shear wave elas- 40. Castéra L, Foucher J, Bernard PH, et al. scoring system for assessing histological ac- ticity imaging: a new ultrasonic technology Pitfalls of liver stiffness measurement: a tivity in asymptomatic chronic active hepati- of medical diagnostics. Ultrasound Med Biol 5-year prospective study of 13,369 exami- tis. Hepatology 1981;1(5):431–435. 1998;24(9):1419–1435. nations. Hepatology 2010;51(3):828–835. 14. Ishak KG. Chronic hepatitis: morphology 28. Ferraioli G, Maiocchi L, Lissandrin R, et 41. Myers RP, Pomier-Layrargues G, Kirsch R, and nomenclature. Mod Pathol 1994;7(6): al. Accuracy of the ElastPQ technique for et al. Feasibility and diagnostic performance 690–713. the assessment of Liver fibrosis in patients of the FibroScan XL probe for liver stiffness 15. Batts KP, Ludwig J. Chronic hepatitis: an with chronic hepatitis C: a “real life” sin- measurement in overweight and obese pa- update on terminology and reporting. Am J gle center study. J Gastrointestin Liver Dis tients. Hepatology 2012;55(1):199–208. Surg Pathol 1995;19(12):1409–1417. 2016;25(3):331–335. 42. Şirli R, Sporea I, Deleanu A, et al. Compar- 16. Regev A, Berho M, Jeffers LJ, et al. Sam- 29. Mare R, Sporea I, Lupuşoru R, et al. The value ison between the M and XL probes for liver pling error and intraobserver variation in of ElastPQ for the evaluation of liver stiffness fibrosis assessment by transient elastography. liver biopsy in patients with chronic HCV in patients with B and C chronic hepatopa- Med Ultrason 2014;16(2):119–122. infection. Am J Gastroenterol 2002;97(10): thies. Ultrasonics 2017;77:144–151. 43. Wong VW, Vergniol J, Wong GL, et al. Liver 2614–2618. 30. Sporea I, Bende F, Sirli R, et al. The per- stiffness measurement using XL probe in pa- 17. Bedossa P, Dargère D, Paradis V. Sampling formance of 2D SWE.GE compared to tients with nonalcoholic fatty liver disease. Am variability of liver fibrosis in chronic hepati- transient elastography for the evaluation of J Gastroenterol 2012;107(12):1862–1871. tis C. Hepatology 2003;38(6):1449–1457. liver stiffness. Ultraschall Med 2016;37(S 44. de Lédinghen V, Wong VW, Vergniol J, et al. 01):SL19_3. 18. Rousselet MC, Michalak S, Dupré F, et al. Diagnosis of liver fibrosis and cirrhosis us- Sources of variability in histological scor- 31. Dietrich CF, Dong Y. Shear wave elastogra- ing liver stiffness measurement: comparison ing of chronic viral hepatitis. Hepatology phy with a new reliability indicator. J Ultra- between M and XL probe of FibroScan®. J 2005;41(2):257–264. son 2016;16(66):281–287. Hepatol 2012;56(4):833–839. 19. Wai CT, Greenson JK, Fontana RJ, et al. 32. Yang YP, Xu XH, Guo LH, et al. Qualitative 45. Friedrich-Rust M, Hadji-Hosseini H, Kri- A simple noninvasive index can predict and quantitative analysis with a novel shear ener S, et al. Transient elastography with a both significant fibrosis and cirrhosis in -pa wave speed imaging for differential diagno- new probe for obese patients for non-inva- tients with chronic hepatitis C. Hepatology sis of breast lesions. Sci Rep 2017;7:40964. sive staging of non-alcoholic steatohepatitis. 2003;38(2):518–526. Eur Radiol 2010;20(10):2390–2396. 33. Mulazzani L, Salvatore V, Matassoni F, et al. 20. Vallet-Pichard A, Mallet V, Nalpas B, et al. Point shear wave ultrasound elastography to 46. Yoneda M, Thomas E, Sclair SN, Grant TT, FIB-4: an inexpensive and accurate marker quantify liver stiffness with Esaote MyLab Schiff ER. Supersonic shear imaging and

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 755 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

transient elastography with the XL probe 58. Shire NJ, Yin M, Chen J, et al. Test-retest 71. Castéra L, Bernard PH, Le Bail B, et al. accurately detect fibrosis in overweight or repeatability of MR elastography for noninva- Transient elastography and biomarkers for obese patients with chronic liver disease. sive liver fibrosis assessment in hepatitis C. J liver fibrosis assessment and follow-up of Clin Gastroenterol Hepatol 2015;13(8):1502– Magn Reson Imaging 2011;34(4):947–955. inactive hepatitis B carriers. Aliment Phar- 1509.e5. macol Ther 2011;33(4):455–465. 59. Jajamovich GH, Dyvorne H, Donnerhack 47. Neukam K, Recio E, Camacho A, et al. Inter- C, Taouli B. Quantitative liver MRI combin- 72. Leung VY, Shen J, Wong VW, et al. Quantita- observer concordance in the assessment of ing phase contrast imaging, elastography, tive elastography of liver fibrosis and spleen liver fibrosis in HIV/HCV-coinfected patients and DWI: assessment of reproducibility stiffness in chronic hepatitis B carriers: using transient elastometry. Eur J Gastroen- and postprandial effect at 3.0 T. PLoS One comparison of shear-wave elastography and terol Hepatol 2010;22(7):801–807. 2014;9(5):e97355. transient elastography with liver biopsy correlation. Radiology 2013;269(3):910–918. 48. Woo H, Lee JY, Yoon JH, Kim W, Cho B, 60. Serai SD, Obuchowski NA, Venkatesh SK, et Choi BI. Comparison of the reliability of al. Repeatability of MR elastography of liver: 73. Cardoso AC, Carvalho-Filho RJ, Stern C, acoustic radiation force impulse imaging and a meta-analysis. Radiology 2017;285(1):92– et al. Direct comparison of diagnostic per- supersonic shear imaging in measurement of 100. formance of transient elastography in pa- liver stiffness. Radiology 2015;277(3):881– tients with chronic hepatitis B and chronic 61. Lee DH, Lee JM, Han JK, Choi BI. MR elas- 886. hepatitis C. Liver Int 2012;32(4):612–621. tography of healthy liver parenchyma: nor- 74. Afdhal NH, Bacon BR, Patel K, et al. Ac- 49. Ferraioli G, Tinelli C, Zicchetti M, et al. Re- mal value and reliability of the liver stiffness curacy of fibroscan, compared with histol- producibility of real-time shear wave elas- value measurement. J Magn Reson Imaging ogy, in analysis of liver fibrosis in patients tography in the evaluation of liver elasticity. 2013;38(5):1215–1223. with hepatitis B or C: a United States mul- Eur J Radiol 2012;81(11):3102–3106. 62. Yasar TK, Wagner M, Bane O, et al. Interplat- ticenter study. Clin Gastroenterol Hepatol 50. Ferraioli G, Tinelli C, Lissandrin R, et al. form reproducibility of liver and spleen stiff- 2015;13(4):772–779.e1–e3. Ultrasound point shear wave elastography ness measured with MR elastography. J Magn 75. Degos F, Perez P, Roche B, et al. Diagnos- assessment of liver and spleen stiffness: ef- Reson Imaging 2016;43(5):1064–1072. tic accuracy of FibroScan and comparison fect of training on repeatability of measure- 63. Serai SD, Yin M, Wang H, Ehman RL, Pod- to liver fibrosis biomarkers in chronic vi- ments. Eur Radiol 2014;24(6):1283–1289. beresky DJ. Cross-vendor validation of liver ral hepatitis: a multicenter prospective 51. Quantitative Imaging Biomarkers Alliance. magnetic resonance elastography. Abdom study (the FIBROSTIC study). J Hepatol Ultrasound SWS Biomarker Ctte. http:// Imaging 2015;40(4):789–794. 2010;53(6):1013–1021. qibawiki.rsna.org/index.php/Ultrasound_ 64. Trout AT, Serai S, Mahley AD, et al. Liver 76. Loong TC, Wei JL, Leung JC, et al. Applica- SWS_Biomarker_Ctte. Accessed March 5, stiffness measurements with MR elastog- tion of the combined FibroMeter vibration- 2017. raphy: agreement and repeatability across controlled transient elastography algorithm 52. Yin M, Glaser KJ, Talwalkar JA, Chen J, imaging systems, field strengths, and pulse in Chinese patients with non-alcoholic Manduca A, Ehman RL. Hepatic MR elas- sequences. Radiology 2016;281(3):793–804. fatty liver disease. J Gastroenterol Hepatol 2017;32(7):1363–1369. tography: clinical performance in a series of 65. Quantitative Imaging Biomarkers Alliance. 1377 consecutive examinations. Radiology MRE Biomarker Ctte. http://qibawiki.rsna. 77. Cassinotto C, Boursier J, de Lédinghen V, et 2016;278(1):114–124. org/index.php/MRE_Biomarker_Ctte. Ac- al. Liver stiffness in nonalcoholic fatty liver 53. Wagner M, Corcuera-Solano I, Lo G, et al. cessed March 4, 2017. disease: A comparison of supersonic shear imaging, FibroScan, and ARFI with liver bi- Technical failure of MR elastography examina- 66. Ziol M, Handra-Luca A, Kettaneh A, et opsy. Hepatology 2016;63(6):1817–1827. tions of the liver: experience from a large sin- al. Noninvasive assessment of liver fibro- gle-center study. Radiology 2017;284(2):401– sis by measurement of stiffness in pa- 78. Sporea I, Sirli R, Bota S, et al. Is ARFI 412. tients with chronic hepatitis C. Hepatology elastography reliable for predicting fibrosis severity in chronic HCV hepatitis? World J 54. Joshi M, Dillman JR, Towbin AJ, Serai SD, 2005;41(1):48–54. Radiol 2011;3(7):188–193. Trout AT. MR elastography: high rate of tech- 67. Castéra L, Vergniol J, Foucher J, et al. Pro- nical success in pediatric and young adult spective comparison of transient elastogra- 79. Friedrich-Rust M, Lupsor M, de Knegt R, et patients. Pediatr Radiol 2017;47(7):838– phy, Fibrotest, APRI, and liver biopsy for the al. Point shear wave elastography by acous- 843. assessment of fibrosis in chronic hepatitis C. tic radiation force impulse quantification in Gastroenterology 2005;128(2):343–350. comparison to transient elastography for 55. Hines CD, Bley TA, Lindstrom MJ, Reed- the noninvasive assessment of liver fibrosis er SB. Repeatability of magnetic reso- 68. Lupsor Platon M, Stefanescu H, Feier D, Ma- in chronic hepatitis C: a prospective inter- nance elastography for quantification of niu A, Badea R. Performance of unidimen- national multicenter study. Ultraschall Med hepatic stiffness. J Magn Reson Imaging sional transient elastography in staging chronic 2015;36(3):239–247. 2010;31(3):725–731. hepatitis C: results from a cohort of 1,202 bi- 80. Ye XP, Ran HT, Cheng J, et al. Liver and opsied patients from one single center. J Gas- 56. Lee Yj, Lee JM, Lee JE, et al. MR elastog- spleen stiffness measured by acoustic radia- trointestin Liver Dis 2013;22(2):157–166. raphy for noninvasive assessment of hepatic tion force impulse elastography for noninva- fibrosis: reproducibility of the examination 69. Zarski JP, Sturm N, Guechot J, et al. Com- sive assessment of liver fibrosis and esopha- and reproducibility and repeatability of the parison of nine blood tests and transient geal varices in patients with chronic hepatitis liver stiffness value measurement. J Magn elastography for liver fibrosis in chronic B. J Ultrasound Med 2012;31(8):1245–1253. Reson Imaging 2014;39(2):326–331. hepatitis C: the ANRS HCEP-23 study. J 81. Cui J, Heba E, Hernandez C, et al. Magnetic Hepatol 2012;56(1):55–62. 57. Shi Y, Guo Q, Xia F, Sun J, Gao Y. Short- resonance elastography is superior to acous- and midterm repeatability of magnetic reso- 70. Marcellin P, Ziol M, Bedossa P, et al. Non- tic radiation force impulse for the diagnosis nance elastography in healthy volunteers at invasive assessment of liver fibrosis by stiff- of fibrosis in patients with biopsy-proven 3.0 T. Magn Reson Imaging 2014;32(6):665– ness measurement in patients with chronic nonalcoholic fatty liver disease: a prospective 670. hepatitis B. Liver Int 2009;29(2):242–247. study. Hepatology 2016;63(2):453–461.

756 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

82. Ferraioli G, Tinelli C, Dal Bello B, et al. fatty liver disease: a prospective study. Hep- 106. Stebbing J, Farouk L, Panos G, et al. A me- Accuracy of real-time shear wave elastog- atology 2014;60(6):1920–1928. ta-analysis of transient elastography for the raphy for assessing liver fibrosis in chronic detection of hepatic fibrosis. J Clin Gastro- 94. Imajo K, Kessoku T, Honda Y, et al. Mag- hepatitis C: a pilot study. Hepatology enterol 2010;44(3):214–219. netic resonance imaging more accurately 2012;56(6):2125–2133. classifies steatosis and fibrosis in patients 107. Friedrich-Rust M, Ong MF, Martens S, et al. 83. Zhuang Y, Ding H, Zhang Y, Sun H, Xu with nonalcoholic fatty liver disease than Performance of transient elastography for the C, Wang W. Two-dimensional shear-wave transient elastography. Gastroenterology staging of liver fibrosis: a meta-analysis. Gas- elastography performance in the nonin- 2016;150(3):626–637.e7. troenterology 2008;134(4):960–974. vasive evaluation of liver fibrosis in pa- 95. Suwanthawornkul T, Anothaisintawee T, 108. Talwalkar JA, Kurtz DM, Schoenleber SJ, tients with chronic hepatitis B: compari- Sobhonslidsuk A, Thakkinstian A, Teerawat- West CP, Montori VM. Ultrasound-based son with serum fibrosis indexes. Radiology tananon Y. Efficacy of second generation transient elastography for the detection 2017;283(3):873–882. direct-acting antiviral agents for treatment of hepatic fibrosis: systematic review and 84. Yin M, Talwalkar JA, Glaser KJ, et al. As- naïve hepatitis C genotype 1: a systematic meta-analysis. Clin Gastroenterol Hepatol sessment of hepatic fibrosis with magnetic review and network meta-analysis. PLoS 2007;5(10):1214–1220. resonance elastography. Clin Gastroenterol One 2015;10(12):e0145953. 109. Shaheen AA, Wan AF, Myers RP. FibroTest Hepatol 2007;5(10):1207–1213.e2. 96. European Association for the Study of and FibroScan for the prediction of hepatitis 85. Huwart L, Sempoux C, Vicaut E, et al. Mag- the Liver. Electronic address: easloffice@ C-related fibrosis: a systematic review of di- netic resonance elastography for the nonin- easloffice.eu. EASL recommendations on agnostic test accuracy. Am J Gastroenterol vasive staging of liver fibrosis. Gastroenter- treatment of hepatitis C 2016. J Hepatol 2007;102(11):2589–2600. ology 2008;135(1):32–40. 2017;66(1):153–194. 110. European Association for Study of Liver; 86. Wang Y, Ganger DR, Levitsky J, et al. As- 97. Gonzalez HC, Jafri SM, Gordon SC. Asociacion Latinoamericana para el Estudio sessment of chronic hepatitis and fibrosis: Role of liver biopsy in the era of direct- del Higado. EASL-ALEH clinical practice comparison of MR elastography and diffu- acting antivirals. Curr Gastroenterol Rep guidelines: non-invasive tests for evaluation sion-weighted imaging. AJR Am J Roent- 2013;15(2):307. of liver disease severity and prognosis. J genol 2011;196(3):553–561. Hepatol 2015;63(1):237–264. 98. Terrault NA, Bzowej NH, Chang KM, et al. 87. Dyvorne HA, Jajamovich GH, Bane O, et al. AASLD guidelines for treatment of chronic 111. Calès P, Boursier J, Lebigot J, et al. Liver Prospective comparison of magnetic reso- hepatitis B. Hepatology 2016;63(1):261–283. fibrosis diagnosis by blood test and elas- nance imaging to transient elastography and tography in chronic hepatitis C: agreement serum markers for liver fibrosis detection. 99. Saito H, Tada S, Nakamoto N, et al. Efficacy of or combination? Aliment Pharmacol Ther Liver Int 2016;36(5):659–666. non-invasive elastometry on staging of hepatic 2017;45(7):991–1003. fibrosis. Hepatol Res 2004;29(2):97–103. 88. Chen J, Yin M, Talwalkar JA, et al. Diagnos- 112. Dong DR, Hao MN, Li C, et al. Acoustic tic performance of MR elastography and vi- 100. Castera L. Noninvasive methods to assess radiation force impulse elastography, Fi- bration-controlled transient elastography in liver disease in patients with hepatitis B or C. broScan®, Forns’ index and their combi- the detection of hepatic fibrosis in patients Gastroenterology 2012;142(6):1293–1302.e4. nation in the assessment of liver fibrosis in with severe to morbid obesity. Radiology 101. Oliveri F, Coco B, Ciccorossi P, et al. Liver patients with chronic hepatitis B, and the 2017;283(2):418–428. stiffness in the hepatitis B virus carrier: a impact of inflammatory activity and steatosis on these diagnostic methods. Mol Med 89. Ichikawa S, Motosugi U, Ichikawa T, et al. non-invasive marker of liver disease influ- Rep 2015;11(6):4174–4182. Magnetic resonance elastography for staging enced by the pattern of transaminases. liver fibrosis in chronic hepatitis C. Magn World J Gastroenterol 2008;14(40):6154– 113. Friedrich-Rust M, Buggisch P, de Knegt RJ, Reson Med Sci 2012;11(4):291–297. 6162. et al. Acoustic radiation force impulse imaging for non-invasive assessment of liver 90. Shi Y, Guo Q, Xia F, et al. MR elastogra- 102. Li Y, Huang YS, Wang ZZ, et al. System- fibrosis in chronic hepatitis B. J Viral Hepat phy for the assessment of hepatic fibrosis in atic review with meta-analysis: the diag- 2013;20(4):240–247. patients with chronic hepatitis B infection: nostic accuracy of transient elastography does histologic necroinflammation influence for the staging of liver fibrosis in patients 114. Liu Y, Dong CF, Yang G, et al. Optimal the measurement of hepatic stiffness? Radi- with chronic hepatitis B. Aliment Pharmacol linear combination of ARFI, transient elas- ology 2014;273(1):88–98. Ther 2016;43(4):458–469. tography and APRI for the assessment of fibrosis in chronic hepatitis B. Liver Int 91. Chang W, Lee JM, Yoon JH, et al. Liver 103. Njei B, McCarty TR, Luk J, Ewelukwa O, Di- 2015;35(3):816–825. fibrosis staging with MR elastography: tah I, Lim JK. Use of transient elastography comparison of diagnostic performance be- in patients with HIV-HCV coinfection: a sys- 115. Park MS, Kim SW, Yoon KT, et al. Factors tween patients with chronic hepatitis B and tematic review and meta-analysis. J Gastro- influencing the diagnostic accuracy of acous- those with other etiologic causes. Radiology enterol Hepatol 2016;31(10):1684–1693. tic radiation force impulse elastography in 2016;280(1):88–97. 104. Chon YE, Choi EH, Song KJ, et al. Per- patients with chronic hepatitis B. Gut Liver. 2016;10(2):275–282. 92. Shi Y, Xia F, Li QJ, et al. Magnetic res- formance of transient elastography for the onance elastography for the evaluation of staging of liver fibrosis in patients with 116. Tai DI, Tsay PK, Jeng WJ, et al. Differ- liver fibrosis in chronic hepatitis B and chronic hepatitis B: a meta-analysis. PLoS ences in liver fibrosis between patients C by using both gradient-recalled echo One 2012;7(9):e44930. with chronic hepatitis B and C: evaluation and spin-echo echo planar imaging: a by acoustic radiation force impulse mea- 105. Tsochatzis EA, Gurusamy KS, Ntaoula S, prospective study. Am J Gastroenterol surements at 2 locations. J Ultrasound Med Cholongitas E, Davidson BR, Burroughs AK. 2016;111(6):823–833. 2015;34(5):813–821. Elastography for the diagnosis of severity 93. Loomba R, Wolfson T, Ang B, et al. Mag- of fibrosis in chronic liver disease: a meta- 117. Ye XP, Ran HT, Cheng J, et al. Liver and netic resonance elastography predicts ad- analysis of diagnostic accuracy. J Hepatol spleen stiffness measured by acoustic radi- vanced fibrosis in patients with nonalcoholic 2011;54(4):650–659. ation force impulse elastography for non-

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 757 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

invasive assessment of liver fibrosis and atic review and meta-analysis. Med Ultrason hepatic fibrosis in chronic hepatitis B. Gut esophageal varices in patients with chronic 2017;19(1):23–31. Liver 2017;11(3):401–408. hepatitis B. J Ultrasound Med 2012;31(8): 129. Summers JA, Radhakrishnan M, Morris E, 141. Wu WP, Chou CT, Chen RC, Lee CW, Lee 1245–1253. et al. Virtual Touch™ quantification to diag- KW, Wu HK. Non-invasive evaluation of he- 118. Zhang D, Chen M, Wang R, et al. Com- nose and monitor liver fibrosis in hepatitis B patic fibrosis: the diagnostic performance parison of acoustic radiation force impulse and hepatitis C: a NICE medical technology of magnetic resonance elastography in pa- imaging and transient elastography for non- guidance. Appl Health Econ Health Policy tients with viral hepatitis B or C. PLoS One invasive assessment of liver fibrosis in pa- 2017;15(2):139–154. 2015;10(10):e0140068. tients with chronic hepatitis B. Ultrasound 130. Verlinden W, Bourgeois S, Gigase P, et al. 142. Venkatesh SK, Xu S, Tai D, Yu H, Wee A. Med Biol 2015;41(1):7–14. Liver fibrosis evaluation using real-time shear Correlation of MR elastography with mor- 119. Conti F, Serra C, Vukotic R, et al. Accuracy wave elastography in hepatitis C-monoinfect- phometric quantification of liver fibrosis of elastography point quantification and ste- ed and human immunodeficiency virus/hepa- (Fibro-C-Index) in chronic hepatitis B. Magn atosis influence on assessing liver fibrosis in titis C-coinfected patients. J Ultrasound Med Reson Med 2014;72(4):1123–1129. patients with chronic hepatitis C. Liver Int 2016;35(6):1299–1308. 143. Venkatesh SK, Wang G, Lim SG, Wee A. 2017;37(2):187–195. 131. Tada T, Kumada T, Toyoda H, et al. Utility Magnetic resonance elastography for the de- 120. Chen SH, Peng CY, Lai HC, et al. Head-to- of real-time shear wave elastography for as- tection and staging of liver fibrosis in chronic head comparison between collagen propor- sessing liver fibrosis in patients with chronic hepatitis B. Eur Radiol 2014;24(1):70–78. tionate area and acoustic radiation force hepatitis C infection without cirrhosis: Com- 144. Lee VS, Miller FH, Omary RA, et al. Mag- impulse elastography in liver fibrosis quan- parison of liver fibrosis indices. Hepatol Res netic resonance elastography and bio- tification in chronic hepatitis C. PLoS One 2015;45(10):E122–E129. 2015;10(10):e0140554. markers to assess fibrosis from recurrent 132. Zeng J, Liu GJ, Huang ZP, et al. Diagnos- hepatitis C in liver transplant recipients. 121. Li SM, Li GX, Fu DM, Wang Y, Dang LQ. tic accuracy of two-dimensional shear wave Transplantation 2011;92(5):581–586. Liver fibrosis evaluation by ARFI and APRI elastography for the non-invasive staging of 145. Singh S, Venkatesh SK, Wang Z, et al. Diag- in chronic hepatitis C. World J Gastroen- hepatic fibrosis in chronic hepatitis B: a co- nostic performance of magnetic resonance terol 2014;20(28):9528–9533. hort study with internal validation. Eur Ra- elastography in staging liver fibrosis: a sys- diol 2014;24(10):2572–2581. 122. Nishikawa T, Hashimoto S, Kawabe N, et al. tematic review and meta-analysis of indi- Factors correlating with acoustic radiation 133. Deffieux T, Gennisson JL, Bousquet L, et vidual participant data. Clin Gastroenterol force impulse elastography in chronic hepatitis al. Investigating liver stiffness and viscosity Hepatol 2015;13(3):440–451.e6. C. World J Gastroenterol 2014;20(5):1289– for fibrosis, steatosis and activity staging 146. Ichikawa S, Motosugi U, Nakazawa T, et al. 1297. using shear wave elastography. J Hepatol Hepatitis activity should be considered a 2015;62(2):317–324. 123. Silva Junior RG, Schmillevitch J, Nasci- confounder of liver stiffness measured with mento MdeF, et al. Acoustic radiation force 134. Jiang T, Tian G, Zhao Q, et al. Diagnostic MR elastography. J Magn Reson Imaging impulse elastography and serum fibrosis accuracy of 2D-shear wave elastography for 2015;41(5):1203–1208. markers in chronic hepatitis C. Scand J Gas- liver fibrosis severity: a meta-analysis. PLoS 147. Yin M, Glaser KJ, Manduca A, et al. Distin- troenterol 2014;49(8):986–992. One 2016;11(6):e0157219. guishing between hepatic inflammation and 124. Yamada R, Hiramatsu N, Oze T, et al. Sig- 135. Godfrey EM, Patterson AJ, Priest AN, et al. fibrosis with MR elastography. Radiology nificance of liver stiffness measurement by A comparison of MR elastography and 31P 2017;284(3):694–705. acoustic radiation force impulse (ARFI) MR spectroscopy with histological staging of 148. Bohte AE, de Niet A, Jansen L, et al. Non- among hepatitis C patients. J Med Virol liver fibrosis. Eur Radiol 2012;22(12):2790– invasive evaluation of liver fibrosis: a com- 2014;86(2):241–247. 2797. parison of ultrasound-based transient elas- 125. Takaki S, Kawakami Y, Miyaki D, et al. Non- 136. Rouvière O, Yin M, Dresner MA, et al. MR tography and MR elastography in patients invasive liver fibrosis score calculated by elastography of the liver: preliminary re- with viral hepatitis B and C. Eur Radiol combination of virtual touch tissue quanti- sults. Radiology 2006;240(2):440–448. 2014;24(3):638–648. fication and serum liver functional tests in chronic hepatitis C patients. Hepatol Res 137. Rustogi R, Horowitz J, Harmath C, et al. 149. Vernon G, Baranova A, Younossi ZM. Sys- 2014;44(3):280–287. Accuracy of MR elastography and anatomic tematic review: the epidemiology and natural MR imaging features in the diagnosis of se- history of non-alcoholic fatty liver disease and 126. Chen SH, Li YF, Lai HC, et al. Effects of pa- vere hepatic fibrosis and cirrhosis. J Magn non-alcoholic steatohepatitis in adults. Ali- tient factors on noninvasive liver stiffness Reson Imaging 2012;35(6):1356–1364. ment Pharmacol Ther 2011;34(3):274–285. measurement using acoustic radiation force impulse elastography in patients with chronic 138. Yoon JH, Lee JM, Joo I, et al. Hepatic fibro- 150. Younossi ZM, Stepanova M, Afendy M, et hepatitis C. BMC Gastroenterol 2012;12:105. sis: prospective comparison of MR elastog- al. Changes in the prevalence of the most raphy and US shear-wave elastography for common causes of chronic liver diseases in 127. Lupsor M, Badea R, Stefanescu H, et al. evaluation. Radiology 2014;273(3):772–782. the United States from 1988 to 2008. Clin Performance of a new elastographic method Gastroenterol Hepatol 2011;9(6):524–530. (ARFI technology) compared to unidimen- 139. Takamura T, Motosugi U, Ichikawa S, et al. e1; quiz e60. sional transient elastography in the nonin- Usefulness of MR elastography for detect- vasive assessment of chronic hepatitis C: ing clinical progression of cirrhosis from 151. Ekstedt M, Hagström H, Nasr P, et al. Fi- preliminary results. J Gastrointestin Liver child-pugh class A to B in patients with type brosis stage is the strongest predictor for Dis 2009;18(3):303–310. C viral hepatitis. J Magn Reson Imaging disease-specific mortality in NAFLD after 2016;44(3):715–722. up to 33 years of follow-up. Hepatology 128. Hu X, Qiu L, Liu D, Qian L. Acoustic radi- 2015;61(5):1547–1554. ation force impulse (ARFI) elastography for 140. Hennedige TP, Wang G, Leung FP, et al. non‑invasive evaluation of hepatic fibrosis in Magnetic resonance elastography and diffu- 152. European Association for the Study of the chronic hepatitis B and C patients: a system- sion weighted imaging in the evaluation of Liver (EASL); European Association for the

758 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

Study of Diabetes (EASD); European Asso- noninvasive diagnosis of advanced fibrosis in agnosis of advanced fibrosis and cirrhosis in ciation for the Study of Obesity (EASO). EA- NAFLD: a prospective study. Am J Gastro- alcoholic liver disease. Eur J Gastroenterol SL-EASD-EASO clinical practice guidelines enterol 2016;111(7):986–994. Hepatol 2015;27(9):1074–1079. for the management of non-alcoholic fatty 164. Park CC, Nguyen P, Hernandez C, et al. 175. Nahon P, Kettaneh A, Tengher-Barna I, et al. liver disease. J Hepatol 2016;64(6):1388– Magnetic resonance elastography vs tran- Assessment of liver fibrosis using transient 1402. sient elastography in detection of fibrosis elastography in patients with alcoholic liver 153. Nobili V, Vizzutti F, Arena U, et al. Accuracy and noninvasive measurement of steatosis disease. J Hepatol 2008;49(6):1062–1068. and reproducibility of transient elastogra- in patients with biopsy-proven nonalco- 176. Thiele M, Detlefsen S, Sevelsted Møller L, et phy for the diagnosis of fibrosis in pediat- holic fatty liver disease. Gastroenterology al. Transient and 2-dimensional shear-wave ric nonalcoholic steatohepatitis. Hepatology 2017;152(3):598–607.e2. elastography provide comparable assess- 2008;48(2):442–448. 165. Singh S, Venkatesh SK, Loomba R, et al. ment of alcoholic liver fibrosis and cirrhosis. 154. Kwok R, Tse YK, Wong GL, et al. System- Magnetic resonance elastography for stag- Gastroenterology 2016;150(1):123–133. atic review with meta-analysis: non-inva- ing liver fibrosis in non-alcoholic fatty liver 177. Attia D, Pischke S, Negm AA, et al. Chang- sive assessment of non-alcoholic fatty liver disease: a diagnostic accuracy systematic re- es in liver stiffness using acoustic radiation disease—the role of transient elastography view and individual participant data pooled force impulse imaging in patients with ob- and plasma cytokeratin-18 fragments. Ali- analysis. Eur Radiol 2016;26(5):1431–1440. structive cholestasis and cholangitis. Dig ment Pharmacol Ther 2014;39(3):254–269. 166. Xu Q, Sheng L, Bao H, et al. Evalua- Liver Dis 2014;46(7):625–631. 155. Puigvehí M, Broquetas T, Coll S, et al. Impact tion of transient elastography in assess- of anthropometric features on the applicabil- ing liver fibrosis in patients with autoim- 178. Kiani A, Brun V, Lainé F, et al. Acoustic radi- ity and accuracy of FibroScan(®) (M and XL) mune hepatitis. J Gastroenterol Hepatol ation force impulse imaging for assessing liver in overweight/obese patients. J Gastroenterol 2017;32(3):639–644. fibrosis in alcoholic liver disease. World J Gas- Hepatol 2017;32(10):1746–1753. troenterol 2016;22(20):4926–4935. 167. Hartl J, Denzer U, Ehlken H, et al. Transient 156. Yoneda M, Suzuki K, Kato S, et al. Nonalco- elastography in autoimmune hepatitis: tim- 179. Mjelle AB, Mulabecirovic A, Hausken T, holic fatty liver disease: US-based acoustic ing determines the impact of inflammation Havre RF, Gilja OH, Vesterhus M. Ul- radiation force impulse elastography. Radi- and fibrosis. J Hepatol 2016;65(4):769–775. trasound and point shear wave elastog- ology 2010;256(2):640–647. raphy in livers of patients with primary 168. Yada N, Sakurai T, Minami T, et al. Influ- sclerosing cholangitis. Ultrasound Med Biol 157. Fierbinteanu Braticevici C, Sporea I, Panai- ence of liver inflammation on liver stiffness 2016;42(9):2146–2155. tescu E, Tribus L. Value of acoustic radiation measurement in patients with autoimmune force impulse imaging elastography for non- hepatitis evaluation by combinational elas- 180. Zhang D, Li P, Chen M, et al. Non-invasive invasive evaluation of patients with nonal- tography. Oncology 2017;92(Suppl 1):10–15. assessment of liver fibrosis in patients with coholic fatty liver disease. Ultrasound Med alcoholic liver disease using acoustic radia- 169. Corpechot C, Carrat F, Poujol-Robert A, et Biol 2013;39(11):1942–1950. tion force impulse elastography. Abdom Im- al. Noninvasive elastography-based assess- aging 2015;40(4):723–729. 158. Guzmán-Aroca F, Frutos-Bernal MD, Bas A, ment of liver fibrosis progression and prog- et al. Detection of non-alcoholic steatohepa- nosis in primary biliary cirrhosis. Hepatol- 181. Zhang DK, Chen M, Liu Y, Wang RF, Liu titis in patients with morbid obesity before ogy 2012;56(1):198–208. LP, Li M. Acoustic radiation force impulse bariatric surgery: preliminary evaluation elastography for non-invasive assessment of 170. Floreani A, Cazzagon N, Martines D, Ca- with acoustic radiation force impulse imag- disease stage in patients with primary bili- valletto L, Baldo V, Chemello L. Perfor- ing. Eur Radiol 2012;22(11):2525–2532. ary cirrhosis: a preliminary study. Clin Ra- mance and utility of transient elastography diol 2014;69(8):836–840. 159. Palmeri ML, Wang MH, Rouze NC, et al. and noninvasive markers of liver fibrosis Noninvasive evaluation of hepatic fibrosis in primary biliary cirrhosis. Dig Liver Dis 182. Wang J, Malik N, Yin M, et al. Magnetic res- using acoustic radiation force-based shear 2011;43(11):887–892. onance elastography is accurate in detecting stiffness in patients with nonalcoholic fatty advanced fibrosis in autoimmune hepatitis. 171. Gómez-Dominguez E, Mendoza J, García- liver disease. J Hepatol 2011;55(3):666–672. World J Gastroenterol 2017;23(5):859–868. Buey L, et al. Transient elastography to 160. Friedrich-Rust M, Romen D, Vermehren J, assess hepatic fibrosis in primary bili- 183. Eaton JE, Dzyubak B, Venkatesh SK, et al. et al. Acoustic radiation force impulse-imag- ary cirrhosis. Aliment Pharmacol Ther Performance of magnetic resonance elastog- ing and transient elastography for non-inva- 2008;27(5):441–447. raphy in primary sclerosing cholangitis. J Gas- sive assessment of liver fibrosis and steatosis troenterol Hepatol 2016;31(6):1184–1190. 172. Corpechot C, Gaouar F, El Naggar A, et in NAFLD. Eur J Radiol 2012;81(3):e325– al. Baseline values and changes in liver 184. Fleming KM, Aithal GP, Card TR, West J. e331. stiffness measured by transient elastogra- The rate of decompensation and clinical 161. Chen J, Talwalkar JA, Yin M, Glaser KJ, phy are associated with severity of fibro- progression of disease in people with cir- Sanderson SO, Ehman RL. Early detection of sis and outcomes of patients with primary rhosis: a cohort study. Aliment Pharmacol nonalcoholic steatohepatitis in patients with sclerosing cholangitis. Gastroenterology Ther 2010;32(11-12):1343–1350. nonalcoholic fatty liver disease by using MR 2014;146(4):970–979; quiz e15–e16. 185. Carrión JA, Navasa M, Bosch J, Bruguera elastography. Radiology 2011;259(3):749– 173. Ehlken H, Wroblewski R, Corpechot C, et M, Gilabert R, Forns X. Transient elastog- 756. al. Validation of transient elastography and raphy for diagnosis of advanced fibrosis and 162. Kim D, Kim WR, Talwalkar JA, Kim HJ, comparison with spleen length measure- portal hypertension in patients with hepati- Ehman RL. Advanced fibrosis in nonalco- ment for staging of fibrosis and clinical prog- tis C recurrence after liver transplantation. holic fatty liver disease: noninvasive as- nosis in primary sclerosing cholangitis. PLoS Liver Transpl 2006;12(12):1791–1798. sessment with MR elastography. Radiology One 2016;11(10):e0164224. 186. Foucher J, Chanteloup E, Vergniol J, et al. 2013;268(2):411–419. 174. Fernandez M, Trépo E, Degré D, et al. Tran- Diagnosis of cirrhosis by transient elastog- 163. Loomba R, Cui J, Wolfson T, et al. Novel sient elastography using Fibroscan is the raphy (FibroScan): a prospective study. Gut 3D magnetic resonance elastography for the most reliable noninvasive method for the di- 2006;55(3):403–408.

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 759 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

187. Kim SU, Lee JH, Kim DY, et al. Prediction of with HCV-related cirrhosis. Hepatology development in viral hepatitis. Hepatol Res liver-related events using fibroscan in chronic 2007;45(5):1290–1297. 2011;41(10):965–970. hepatitis B patients showing advanced liver 199. Wong GL, Chan HL, Yu Z, et al. Nonin- 210. Chon YE, Jung ES, Park JY, et al. The ac- fibrosis. PLoS One 2012;7(5):e36676. vasive assessments of liver fibrosis with curacy of noninvasive methods in predicting 188. de Lédinghen V, Vergniol J, Barthe C, et transient elastography and Hui index the development of hepatocellular carci- al. Non-invasive tests for fibrosis and liver predict survival in patients with chronic noma and hepatic decompensation in pa- stiffness predict 5-year survival of patients hepatitis B. J Gastroenterol Hepatol tients with chronic hepatitis B. J Clin Gas- chronically infected with hepatitis B virus. 2015;30(3):582–590. troenterol 2012;46(6):518–525. Aliment Pharmacol Ther 2013;37(10):979– 200. Singh S, Fujii LL, Murad MH, et al. Liver 211. Feier D, Lupsor Platon M, Stefanescu H, 988. stiffness is associated with risk of decom- Badea R. Transient elastography for the de- 189. Llop E, Berzigotti A, Reig M, et al. Assess- pensation, liver cancer, and death in patients tection of hepatocellular carcinoma in viral C ment of portal hypertension by transient with chronic liver diseases: a systematic re- liver cirrhosis. is there something else than elastography in patients with compensated view and meta-analysis. Clin Gastroenterol increased liver stiffness? J Gastrointestin cirrhosis and potentially resectable liver tu- Hepatol 2013;11(12):1573–1584.e1–e2; quiz Liver Dis 2013;22(3):283–289. mors. J Hepatol 2012;56(1):103–108. e88–e89. 212. Jung KS, Kim SU, Ahn SH, et al. Risk as- 190. Macías J, Camacho A, Von Wichmann 201. Shi KQ, Fan YC, Pan ZZ, et al. Transient sessment of hepatitis B virus-related hepa- MA, et al. Liver stiffness measurement elastography: a meta-analysis of diagnos- tocellular carcinoma development using liver versus liver biopsy to predict survival and tic accuracy in evaluation of portal hyper- stiffness measurement (FibroScan). Hepa- decompensations of cirrhosis among HIV/ tension in chronic liver disease. Liver Int tology 2011;53(3):885–894. hepatitis C virus-coinfected patients. AIDS 2013;33(1):62–71. 213. Klibansky DA, Mehta SH, Curry M, Nasser 2013;27(16):2541–2549. 202. Kim BK, Han KH, Park JY, et al. A liver I, Challies T, Afdhal NH. Transient elastog- 191. Merchante N, Rivero-Juárez A, Téllez F, stiffness measurement-based, noninvasive raphy for predicting clinical outcomes in et al. Liver stiffness predicts clinical out- prediction model for high-risk esophageal patients with chronic liver disease. J Viral come in human immunodeficiency virus/ varices in B-viral liver cirrhosis. Am J Gas- Hepat 2012;19(2):e184–e193. hepatitis C virus-coinfected patients with troenterol 2010;105(6):1382–1390. 214. Masuzaki R, Tateishi R, Yoshida H, et al. compensated liver cirrhosis. Hepatology 203. Berzigotti A, Seijo S, Arena U, et al. Elas- Prospective risk assessment for hepatocellu- 2012;56(1):228–238. tography, spleen size, and platelet count lar carcinoma development in patients with 192. Pang JX, Zimmer S, Niu S, et al. Liver stiff- identify portal hypertension in patients with chronic hepatitis C by transient elastography. ness by transient elastography predicts liver- compensated cirrhosis. Gastroenterology Hepatology 2009;49(6):1954–1961. related complications and mortality in pa- 2013;144(1):102–111.e1. 215. Narita Y, Genda T, Tsuzura H, et al. Pre- tients with chronic liver disease. PLoS One 204. de Franchis R; Baveno VI Faculty. Expanding diction of liver stiffness hepatocellular car- 2014;9(4):e95776. consensus in portal hypertension: report of cinoma in chronic hepatitis C patients on 193. Park MS, Kim SU, Kim BK, et al. Prog- the Baveno VI Consensus Workshop: strati- interferon-based anti-viral therapy. J Gas- nostic value of the combined use of tran- fying risk and individualizing care for portal troenterol Hepatol 2014;29(1):137–143. sient elastography and fibrotest in pa- hypertension. J Hepatol 2015;63(3):743– 216. Poynard T, Vergniol J, Ngo Y, et al. Staging tients with chronic hepatitis B. Liver Int 752. chronic hepatitis C in seven categories using 2015;35(2):455–462. 205. Colecchia A, Montrone L, Scaioli E, et al. fibrosis biomarker (FibroTest™) and tran- 194. Pérez-Latorre L, Sánchez-Conde M, Rincón Measurement of spleen stiffness to evalu- sient elastography (FibroScan®). J Hepatol D, et al. Prediction of liver complications in ate portal hypertension and the presence 2014;60(4):706–714. patients with hepatitis C virus-related cir- of esophageal varices in patients with 217. Wang HM, Hung CH, Lu SN, et al. Liver rhosis with and without HIV coinfection: HCV-related cirrhosis. Gastroenterology stiffness measurement as an alterna- comparison of hepatic venous pressure gra- 2012;143(3):646–654. tive to fibrotic stage in risk assessment dient and transient elastography. Clin Infect 206. Calvaruso V, Bronte F, Conte E, Simone F, of hepatocellular carcinoma incidence Dis 2014;58(5):713–718. Craxì A, Di Marco V. Modified spleen stiff- for chronic hepatitis C patients. Liver Int 195. Robic MA, Procopet B, Métivier S, et al. ness measurement by transient elastography 2013;33(5):756–761. Liver stiffness accurately predicts portal hy- is associated with presence of large oesoph- 218. Wong GL, Chan HL, Wong CK, et al. pertension related complications in patients ageal varices in patients with compensated Liver stiffness-based optimization of he- with chronic liver disease: a prospective hepatitis C virus cirrhosis. J Viral Hepat patocellular carcinoma risk score in pa- study. J Hepatol 2011;55(5):1017–1024. 2013;20(12):867–874. tients with chronic hepatitis B. J Hepatol 196. Vergniol J, Foucher J, Terrebonne E, et al. 207. Sharma P, Kirnake V, Tyagi P, et al. Spleen 2014;60(2):339–345. Noninvasive tests for fibrosis and liver stiff- stiffness in patients with cirrhosis in predict- 219. Kim DY, Song KJ, Kim SU, et al. Transient ness predict 5-year outcomes of patients ing esophageal varices. Am J Gastroenterol elastography-based risk estimation of hepa- with chronic hepatitis C. Gastroenterology 2013;108(7):1101–1107. titis B virus-related occurrence of hepatocel- 2011;140(7):1970–1979, 1979.e1–e3. 208. Stefanescu H, Grigorescu M, Lupsor M, lular carcinoma: development and validation 197. Vergniol J, Boursier J, Coutzac C, et al. Procopet B, Maniu A, Badea R. Spleen stiff- of a predictive model. Onco Targets Ther Evolution of noninvasive tests of liver fi- ness measurement using Fibroscan for the 2013;6:1463–1469. brosis is associated with prognosis in pa- noninvasive assessment of esophageal vari- 220. Vermehren J, Polta A, Zimmermann O, et tients with chronic hepatitis C. Hepatology ces in liver cirrhosis patients. J Gastroen- al. Comparison of acoustic radiation force 2014;60(1):65–76. terol Hepatol 2011;26(1):164–170. impulse imaging with transient elastography 198. Vizzutti F, Arena U, Romanelli RG, et 209. Akima T, Tamano M, Hiraishi H. Liver stiff- for the detection of complications in patients al. Liver stiffness measurement predicts ness measured by transient elastography with cirrhosis. Liver Int 2012;32(5):852– severe portal hypertension in patients is a predictor of hepatocellular carcinoma 858.

760 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

221. Takuma Y, Morimoto Y, Takabatake H, et al. 232. Guo J, Büning C, Schott E, et al. In vivo transient elastography after hepatitis C virus Measurement of spleen stiffness with acous- abdominal magnetic resonance elastography protease inhibitor-based triple therapy. Eur tic radiation force impulse imaging predicts for the assessment of portal hypertension J Gastroenterol Hepatol 2016;28(3):305– mortality and hepatic decompensation in pa- before and after transjugular intrahepatic 312. tients with liver cirrhosis. Clin Gastroenterol portosystemic shunt implantation. Invest 244. Elsharkawy A, Alem SA, Fouad R, et al. Hepatol 2017;15(11):1782–1790.e4. Radiol 2015;50(5):347–351. Changes in liver stiffness measurements and 222. Morishita N, Hiramatsu N, Oze T, et al. 233. Morisaka H, Motosugi U, Ichikawa S, Sano fibrosis scores following sofosbuvir based Liver stiffness measurement by acoustic K, Ichikawa T, Enomoto N. Association treatment regimens without interferon. J radiation force impulse is useful in pre- of splenic MR elastographic findings with Gastroenterol Hepatol 2017;32(9):1624– dicting the presence of esophageal varices gastroesophageal varices in patients with 1630. or high-risk esophageal varices among pa- chronic liver disease. J Magn Reson Imaging 245. Bachofner JA, Valli PV, Kröger A, et al. tients with HCV-related cirrhosis. J Gastro- 2015;41(1):117–124. Direct antiviral agent treatment of chronic enterol 2014;49(7):1175–1182 [Published 234. Ronot M, Lambert S, Elkrief L, et al. As- hepatitis C results in rapid regression of correction appears in J Gastroenterol sessment of portal hypertension and high- transient elastography and fibrosis markers 2015;50(6):705.]. risk oesophageal varices with liver and fibrosis-4 score and aspartate amino- 223. Aoki T, Iijima H, Tada T, et al. Prediction spleen three-dimensional multifrequency transferase-platelet ratio index. Liver Int of development of hepatocellular carci- MR elastography in liver cirrhosis. Eur Ra- 2017;37(3):369–376. diol 2014;24(6):1394–1402. noma using a new scoring system involv- 246. Yada N, Sakurai T, Minami T, et al. Ultra- ing virtual touch quantification in patients 235. Shin SU, Lee JM, Yu MH, et al. Prediction of sound elastography correlates treatment re- with chronic liver diseases. J Gastroenterol esophageal varices in patients with cirrhosis: sponse by antiviral therapy in patients with 2017;52(1):104–112. usefulness of three-dimensional MR elastog- chronic hepatitis C. Oncology 2014;87(Sup- 224. Takamura M, Kanefuji T, Suda T, et al. Value raphy with echo-planar imaging technique. pl 1):118–123. Radiology 2014;272(1):143–153. of shear wave velocity measurements for the 247. Casado JL, Quereda C, Moreno A, Pérez- risk assessment of hepatocellular carcinoma 236. Talwalkar JA, Yin M, Venkatesh S, et al. Fea- Elías MJ, Martí-Belda P, Moreno S. Regres- development in patients with nonalcoholic sibility of in vivo MR elastographic splenic sion of liver fibrosis is progressive after sus- fatty liver disease : HCC risk assessment by stiffness measurements in the assessment of tained virological response to HCV therapy VTTQ. Hepatol Int 2014;8(2):240–249. portal hypertension. AJR Am J Roentgenol in patients with hepatitis C and HIV coinfec- 225. Choi SY, Jeong WK, Kim Y, Kim J, Kim TY, 2009;193(1):122–127. tion. J Viral Hepat 2013;20(12):829–837. Sohn JH. Shear-wave elastography: a nonin- 237. Dyvorne HA, Jajamovich GH, Besa C, Coo- 248. Martinez SM, Foucher J, Combis JM, vasive tool for monitoring changing hepatic per N, Taouli B. Simultaneous measurement et al. Longitudinal liver stiffness assess- venous pressure gradients in patients with of hepatic and splenic stiffness using MR ment in patients with chronic hepatitis C cirrhosis. Radiology 2014;273(3):917–926. elastography: preliminary experience. Ab- undergoing antiviral therapy. PLoS One 226. Procopet B, Berzigotti A, Abraldes JG, et al. dom Imaging 2015;40(4):803–809. 2012;7(10):e47715. Real-time shear-wave elastography: applica- 238. Stefanescu H, Radu C, Procopet B, et al. 249. Suda T, Okawa O, Masaoka R, et al. Shear bility, reliability and accuracy for clinically Non-invasive ménage à trois for the predic- wave elastography in hepatitis C patients significant portal hypertension. J Hepatol tion of high-risk varices: stepwise algorithm before and after antiviral therapy. World J 2015;62(5):1068–1075. using lok score, liver and spleen stiffness. Hepatol 2017;9(1):64–68. Liver Int 2015;35(2):317–325. 227. Elkrief L, Rautou PE, Ronot M, et al. Pro- 250. Tachi Y, Hirai T, Kojima Y, et al. Liver stiff- spective comparison of spleen and liver 239. Anaparthy R, Talwalkar JA, Yin M, Roberts ness measurement using acoustic radiation stiffness by using shear-wave and tran- LR, Fidler JL, Ehman RL. Liver stiffness force impulse elastography in hepatitis C sient elastography for detection of por- measurement by magnetic resonance elas- virus-infected patients with a sustained vi- tal hypertension in cirrhosis. Radiology tography is not associated with developing rological response. Aliment Pharmacol Ther 2015;275(2):589–598. hepatocellular carcinoma in subjects with 2016;44(4):346–355. compensated cirrhosis. Aliment Pharmacol 228. Cassinotto C, Charrie A, Mouries A, et al. 251. Schmid P, Bregenzer A, Huber M, et al. Pro- Ther 2011;34(1):83–91. Liver and spleen elastography using super- gression of liver fibrosis in HIV/HCV co-in- sonic shear imaging for the non-invasive di- 240. Motosugi U, Ichikawa T, Koshiishi T, et fection: a comparison between non-invasive agnosis of cirrhosis severity and oesophageal al. Liver stiffness measured by magnetic assessment methods and liver biopsy. PLoS varices. Dig Liver Dis 2015;47(8):695–701. resonance elastography as a risk fac- One 2015;10(9):e0138838. tor for hepatocellular carcinoma: a pre- 229. Kasai Y, Moriyasu F, Saito K, et al. Value of 252. Tian WS, Lin MX, Zhou LY, et al. Maximum liminary case-control study. Eur Radiol shear wave elastography for predicting he- value measured by 2-D shear wave elastog- 2013;23(1):156–162. patocellular carcinoma and esophagogastric raphy helps in differentiating malignancy varices in patients with chronic liver disease. 241. Foster GR, Irving WL, Cheung MC, et from benign focal liver lesions. Ultrasound J Med Ultrason (2001) 2015;42(3):349–355. al. Impact of direct acting antiviral ther- Med Biol 2016;42(9):2156–2166. apy in patients with chronic hepatitis C 230. Asrani SK, Talwalkar JA, Kamath PS, et al. 253. Guo LH, Wang SJ, Xu HX, et al. Differ- and decompensated cirrhosis. J Hepatol Role of magnetic resonance elastography entiation of benign and malignant focal 2016;64(6):1224–1231. in compensated and decompensated liver liver lesions: value of virtual touch tis- disease. J Hepatol 2014;60(5):934–939. 242. Flemming JA, Kim WR, Brosgart CL, Ter- sue quantification of acoustic radiation rault NA. Reduction in liver transplant wait- force impulse elastography. Med Oncol 231. Lee DH, Lee JM, Yi NJ, et al. Hepatic listing in the era of direct-acting antiviral 2015;32(3):68. stiffness measurement by using MR elas- therapy. Hepatology 2017;65(3):804–812. tography: prognostic values after hepatic 254. Ronot M, Di Renzo S, Gregoli B, et al. Char- resection for hepatocellular carcinoma. Eur 243. Sáez-Royuela F, Linares P, Cervera LA, et al. acterization of fortuitously discovered focal Radiol 2017;27(4):1713–1721. Evaluation of advanced fibrosis measured by liver lesions: additional information pro-

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 761 STATE OF THE ART : Quantitative Elastography Methods in Liver Diseases Kennedy et al

vided by shearwave elastography. Eur Radiol histopathology features—preliminary find- 279. Gerber L, Kasper D, Fitting D, et al. As- 2015;25(2):346–358. ings. Magn Reson Imaging 2017;37:41–45. sessment of liver fibrosis with 2-D shear wave elastography in comparison to tran- 255. Park HS, Kim YJ, Yu MH, Jung SI, Jeon 268. Gordic S, Ayache JB, Kennedy P, et al. sient elastography and acoustic radiation HJ. Shear wave elastography of focal liver Value of tumor stiffness measured with force impulse imaging in patients with lesion: intraobserver reproducibility and MR elastography for assessment of re- chronic liver disease. Ultrasound Med Biol elasticity characterization. Ultrasound Q sponse of hepatocellular carcinoma to lo- 2015;41(9):2350–2359. 2015;31(4):262–271. coregional therapy. Abdom Radiol (NY) 2017;42(6):1685–1694. 256. Özmen E, Adaletli I, Kayadibi Y, et al. The 280. Sporea I, Bota S, Jurchis A, et al. Acous- impact of share wave elastography in dif- 269. Mariappan YK, Glaser KJ, Ehman RL. Mag- tic radiation force impulse and supersonic ferentiation of hepatic hemangioma from netic resonance elastography: a review. Clin shear imaging versus transient elastography malignant liver tumors in pediatric popula- Anat 2010;23(5):497–511. for liver fibrosis assessment. Ultrasound tion. Eur J Radiol 2014;83(9):1691–1697. Med Biol 2013;39(11):1933–1941. 270. Johnson CL, Schwarb H, D J McGarry 257. Zhang P, Zhou P, Tian SM, Qian Y, Deng J, M, et al. Viscoelasticity of subcortical 281. Arena U, Lupsor Platon M, Stasi C, et al. Zhang L. Application of acoustic radiation gray matter structures. Hum Brain Mapp Liver stiffness is influenced by a standard- force impulse imaging for the evaluation of 2016;37(12):4221–4233. ized meal in patients with chronic hepatitis focal liver lesion elasticity. Hepatobiliary C virus at different stages of fibrotic evolu- 271. Ichikawa S, Motosugi U, Morisaka H, et al. Pancreat Dis Int 2013;12(2):165–170. tion. Hepatology 2013;58(1):65–72. Comparison of the diagnostic accuracies of 258. Park H, Park JY, Kim DY, et al. Character- magnetic resonance elastography and tran- 282. Berzigotti A, De Gottardi A, Vukotic R, et ization of focal liver masses using acoustic sient elastography for hepatic fibrosis. Magn al. Effect of meal ingestion on liver stiffness radiation force impulse elastography. World Reson Imaging 2015;33(1):26–30. in patients with cirrhosis and portal hyper- J Gastroenterol 2013;19(2):219–226. tension. PLoS One 2013;8(3):e58742. 272. Guo Y, Parthasarathy S, Goyal P, McCar- 259. Guibal A, Boularan C, Bruce M, et al. Evalua- thy RJ, Larson AC, Miller FH. Magnetic 283. Hines CD, Lindstrom MJ, Varma AK, tion of shearwave elastography for the charac- resonance elastography and acoustic ra- Reeder SB. Effects of postprandial state and terisation of focal liver lesions on ultrasound. diation force impulse for staging hepatic mesenteric blood flow on the repeatability of Eur Radiol 2013;23(4):1138–1149. fibrosis: a meta-analysis. Abdom Imaging MR elastography in asymptomatic subjects. 2015;40(4):818–834. J Magn Reson Imaging 2011;33(1):239–244. 260. Kapoor A, Kapoor A, Mahajan G, Sidhu BS, Lakhanpal VP. Real-time elastography 273. Song P, Mellema DC, Sheedy SP, et al. Per- 284. Mederacke I, Wursthorn K, Kirschner J, et in differentiating metastatic from nonmeta- formance of 2-dimensional ultrasound shear al. Food intake increases liver stiffness in static liver nodules. Ultrasound Med Biol wave elastography in liver fibrosis detection patients with chronic or resolved hepatitis C 2011;37(2):207–213. using magnetic resonance elastography as virus infection. Liver Int 2009;29(10):1500– the reference standard: a pilot study. J Ul- 1506. 261. Davies G, Koenen M. Acoustic radiation trasound Med 2016;35(2):401–412. force impulse elastography in distinguish- 285. Popescu A, Bota S, Sporea I, et al. The influ- ing hepatic haemangiomata from metasta- 274. Ichikawa S, Motosugi U, Morisaka H, et ence of food intake on liver stiffness values ses: preliminary observations. Br J Radiol al. MRI-based staging of hepatic fibrosis: assessed by acoustic radiation force impulse 2011;84(1006):939–943. comparison of intravoxel incoherent motion elastography-preliminary results. Ultra- diffusion-weighted imaging with magnetic sound Med Biol 2013;39(4):579–584. 262. Shuang-Ming T, Ping Z, Ying Q, Li-Rong C, resonance elastography. J Magn Reson Im- Ping Z, Rui-Zhen L. Usefulness of acoustic aging 2015;42(1):204–210. 286. Yin M, Talwalkar JA, Glaser KJ, et al. Dy- radiation force impulse imaging in the differ- namic postprandial hepatic stiffness aug- ential diagnosis of benign and malignant liver 275. Park HS, Kim YJ, Yu MH, Choe WH, Jung mentation assessed with MR elastography in lesions. Acad Radiol 2011;18(7):810–815. SI, Jeon HJ. Three-Tesla magnetic reso- patients with chronic liver disease. AJR Am nance elastography for hepatic fibrosis: 263. Cho SH, Lee JY, Han JK, Choi BI. Acous- J Roentgenol 2011;197(1):64–70. comparison with diffusion-weighted imag- tic radiation force impulse elastography for ing and gadoxetic acid-enhanced magnetic 287. Millonig G, Reimann FM, Friedrich S, et the evaluation of focal solid hepatic lesions: resonance imaging. World J Gastroenterol al. Extrahepatic cholestasis increases liver preliminary findings. Ultrasound Med Biol 2014;20(46):17558–17567. stiffness (FibroScan) irrespective of fibrosis. 2010;36(2):202–208. Hepatology 2008;48(5):1718–1723. 276. Wang QB, Zhu H, Liu HL, Zhang B. Per- 264. Venkatesh SK, Yin M, Glockner JF, et al. formance of magnetic resonance elastogra- 288. Pfeifer L, Strobel D, Neurath MF, Wildner MR elastography of liver tumors: pre- phy and diffusion-weighted imaging for the D. Liver stiffness assessed by acoustic radia- liminary results. AJR Am J Roentgenol staging of hepatic fibrosis: a meta-analysis. tion force impulse (ARFI) technology is con- 2008;190(6):1534–1540. Hepatology 2012;56(1):239–247. siderably increased in patients with chole- 265. Garteiser P, Doblas S, Daire JL, et al. MR stasis. Ultraschall Med 2014;35(4):364–367. 277. Bota S, Herkner H, Sporea I, et al. Meta- elastography of liver tumours: value of visco- analysis: ARFI elastography versus transient 289. Sporea I, Raţiu I, Bota S, Şirli R, Jurchiş A. elastic properties for tumour characterisa- elastography for the evaluation of liver fibro- Are different cut-off values of liver stiffness tion. Eur Radiol 2012;22(10):2169–2177. sis. Liver Int 2013;33(8):1138–1147. assessed by transient elastography accord- 266. Hennedige TP, Hallinan JT, Leung FP, et al. ing to the etiology of liver cirrhosis for pre- 278. Sporea I, Sirli R, Bota S, Popescu A, Send- Comparison of magnetic resonance elastog- dicting significant esophageal varices? Med roiu M, Jurchis A. Comparative study con- raphy and diffusion-weighted imaging for Ultrason 2013;15(2):111–115. cerning the value of acoustic radiation force differentiating benign and malignant liver le- impulse elastography (ARFI) in comparison 290. Lombardi R, Buzzetti E, Roccarina D, sions. Eur Radiol 2016;26(2):398–406. with transient elastography (TE) for the as- Tsochatzis EA. Non-invasive assessment 267. Thompson SM, Wang J, Chandan VS, et sessment of liver fibrosis in patients with of liver fibrosis in patients with alco- al. MR elastography of hepatocellular car- chronic hepatitis B and C. Ultrasound Med holic liver disease. World J Gastroenterol cinoma: correlation of tumor stiffness with Biol 2012;38(8):1310–1316. 2015;21(39):11044–11052.

762 radiology.rsna.org n Radiology: Volume 286: Number 3—March 2018 STATE OF THE ART: Quantitative Elastography Methods in Liver Diseases Kennedy et al

291. Xu J, Liu X, Koyama Y, et al. The types of withdrawal in heavy drinkers. Alcohol Clin 315. Sasso M, Audière S, Kemgang A, et al. Liver hepatic myofibroblasts contributing to liver Exp Res 2012;36(8):1407–1411. steatosis assessed by controlled attenuation fibrosis of different etiologies. Front Phar- parameter (CAP) measured with the XL 305. Wong GL, Wong VW, Choi PC, et al. In- macol 2014;5:167. probe of the FibroScan: a pilot study assess- creased liver stiffness measurement by tran- ing diagnostic accuracy. Ultrasound Med 292. Pinzani M. Pathophysiology of liver fibrosis. sient elastography in severe acute exacerba- Biol 2016;42(1):92–103. Dig Dis 2015;33(4):492–497. tion of chronic hepatitis B. J Gastroenterol Hepatol 2009;24(6):1002–1007. 293. Friedrich-Rust M, Nierhoff J, Lupsor M, et 316. Reeder SB, Hu HH, Sirlin CB. Proton density fat-fraction: a standardized MR-based al. Performance of acoustic radiation force 306. Mueller S, Englert S, Seitz HK, et al. In- biomarker of tissue fat concentration. J impulse imaging for the staging of liver fi- flammation-adapted liver stiffness values for Magn Reson Imaging 2012;36(5):1011–1014. brosis: a pooled meta-analysis. J Viral Hepat improved fibrosis staging in patients with 2012;19(2):e212–e219. hepatitis C virus and alcoholic liver disease. 317. Weaver JB, Van Houten EE, Miga MI, Ken- 294. Yun MH, Seo YS, Kang HS, et al. The ef- Liver Int 2015;35(12):2514–2521. nedy FE, Paulsen KD. Magnetic resonance fect of the respiratory cycle on liver stiffness 307. Boursier J, de Ledinghen V, Sturm N, et elastography using 3D gradient echo mea- values as measured by transient elastogra- al. Precise evaluation of liver histology surements of steady-state motion. Med Phys phy. J Viral Hepat 2011;18(9):631–636. by computerized morphometry shows 2001;28(8):1620–1628. 295. Chapman T, Dubinsky T, Barr RG. Ultra- that steatosis influences liver stiffness 318. Hirsch S, Guo J, Reiter R, et al. Towards sound elastography of the liver: what the measured by transient elastography in compression-sensitive magnetic reso- clinician needs to know. J Ultrasound Med chronic hepatitis C. J Gastroenterol nance elastography of the liver: sensitiv- 2017;36(7):1293–1304. 2014;49(3):527–537. ity of harmonic volumetric strain to portal hypertension. J Magn Reson Imaging 296. Wang K, Manning P, Szeverenyi N, et al. 308. Gaia S, Carenzi S, Barilli AL, et al. Reliabil- 2014;39(2):298–306. Repeatability and reproducibility of 2D and ity of transient elastography for the detec- 3D hepatic MR elastography with rigid and tion of fibrosis in non-alcoholic fatty liver 319. Glaser KJ, Manduca A, Ehman RL. Review flexible drivers at end-expiration and end- disease and chronic viral hepatitis. J Hepa- of MR elastography applications and re- inspiration in healthy volunteers. Abdom tol 2011;54(1):64–71. cent developments. J Magn Reson Imaging Radiol (NY) 2017 Jun 13 [Epub ahead of 309. Arena U, Vizzutti F, Abraldes JG, et al. 2012;36(4):757–774. print]. Reliability of transient elastography for the 320. Morisaka H, Motosugi U, Glaser KJ, et al. 297. Echosens. FibroScan reimbursement. diagnosis of advanced fibrosis in chronic Comparison of diagnostic accuracies of http://www.echosens.us/reimbursement-0. hepatitis C. Gut 2008;57(9):1288–1293. two- and three-dimensional MR elastog- Accessed March 4, 2017. 310. Wong VW, Vergniol J, Wong GL, et al. Di- raphy of the liver. J Magn Reson Imaging 2017;45(4):1163–1170. 298. Arena U, Vizzutti F, Corti G, et al. Acute agnosis of fibrosis and cirrhosis using liver viral hepatitis increases liver stiffness values stiffness measurement in nonalcoholic fatty 321. Asbach P, Klatt D, Hamhaber U, et al. As- measured by transient elastography. Hepa- liver disease. Hepatology 2010;51(2):454– sessment of liver viscoelasticity using mul- tology 2008;47(2):380–384. 462. tifrequency MR elastography. Magn Reson 299. Coco B, Oliveri F, Maina AM, et al. Tran- 311. Wagner M, Besa C, Bou Ayache J, et al. Med 2008;60(2):373–379. sient elastography: a new surrogate Magnetic resonance elastography of the 322. Klatt D, Hamhaber U, Asbach P, Braun J, marker of liver fibrosis influenced by major liver: qualitative and quantitative compari- Sack I. Noninvasive assessment of the rhe- changes of transaminases. J Viral Hepat son of gradient echo and spin echo echo- ological behavior of human organs using 2007;14(5):360–369. planar imaging sequences. Invest Radiol multifrequency MR elastography: a study of 2016;51(9):575–581. 300. Sagir A, Erhardt A, Schmitt M, Häussinger brain and liver viscoelasticity. Phys Med Biol D. Transient elastography is unreliable for 312. Mariappan YK, Dzyubak B, Glaser KJ, 2007;52(24):7281–7294. et al. Application of modified spin-echo- detection of cirrhosis in patients with acute 323. Etchell E, Jugé L, Hatt A, Sinkus R, Bilston based sequences for hepatic MR elastog- liver damage. Hepatology 2008;47(2):592– LE. Liver stiffness values are lower in pedi- raphy: evaluation, comparison with the 595. atric subjects than in adults and increase conventional gradient-echo sequence, and 301. Millonig G, Friedrich S, Adolf S, et al. Liver with age: a multifrequency MR elastogra- preliminary clinical experience. Radiology stiffness is directly influenced by central ve- phy study. Radiology 2017;283(1):222–230. 2017;282(2):390–398. nous pressure. J Hepatol 2010;52(2):206– 324. Barnhill E, Hollis L, Sack I, et al. Nonlin- 313. Sasso M, Beaugrand M, de Ledinghen V, et 210. ear multiscale regularisation in MR elastog- al. Controlled attenuation parameter (CAP): 302. Bardou-Jacquet E, Legros L, Soro D, et al. Ef- raphy: towards fine feature mapping. Med a novel VCTE™ guided ultrasonic attenuation fect of alcohol consumption on liver stiffness Image Anal 2017;35:133–145. measurement for the evaluation of hepatic measured by transient elastography. World J steatosis: preliminary study and validation in 325. Papazoglou S, Hirsch S, Braun J, Sack Gastroenterol 2013;19(4):516–522. a cohort of patients with chronic liver disease I. Multifrequency inversion in magnetic 303. Mueller S, Millonig G, Sarovska L, et from various causes. Ultrasound Med Biol resonance elastography. Phys Med Biol al. Increased liver stiffness in alcoholic 2010;36(11):1825–1835. 2012;57(8):2329–2346. liver disease: differentiating fibrosis from 314. Karlas T, Petroff D, Sasso M, et al. Indi- 326. Tzschätzsch H, Guo J, Dittmann F, et al. steatohepatitis. World J Gastroenterol vidual patient data meta-analysis of con- Tomoelastography by multifrequency wave 2010;16(8):966–972. trolled attenuation parameter (CAP) tech- number recovery from time-harmonic 304. Trabut JB, Thépot V, Nalpas B, et al. Rapid nology for assessing steatosis. J Hepatol propagating shear waves. Med Image Anal decline of liver stiffness following alcohol 2017;66(5):1022–1030. 2016;30:1–10.

Radiology: Volume 286: Number 3—March 2018 n radiology.rsna.org 763