Original Research

Portal Hypertension Correlates With SplenicStiffness as Measured With MR Elastography

Geir I. Nedredal, MD, PhD,1 Meng Yin, PhD,2 Travis McKenzie, MD,1

Joseph Lillegard, MD, PhD,1 Jennifer Luebke-Wheeler, PhD,1 Jayant Talwalkar, MD,2

Richard Ehman, MD,2 and Scott L. Nyberg, MD, PhD1*

Purpose: To investigate the correlation between MR elas-tography (MRE) assessed spleen stiffness and direct por-tal vein pressure gradient (D-HVPG) measurements in alarge animal model of portal hypertension.

Materials and Methods: Cholestatic liver disease wasestablished in adult canines by common bile duct liga-tion. A spin echo based echo planar imaging (EPI) MREsequence was used to acquire three-dimensional/threeaxis (3D/3-axis) abdominal MRE data at baseline, 4weeks, and 8 weeks. Liver biopsies, blood samples, andD-HVPG measurements were obtained simultaneously.

Results: Animals developed portal hypertension (D-HVPG: 11.0 6 5.1 mmHg) with only F1 fibrosis after 4weeks. F3 fibrosis was confirmed after 8 weeks despite nofurther rise in portal hypertension (D-HVPG: 11.3 6 3.2mmHg). Mean stiffnesses of the spleen increased overtwo-fold from baseline (1.72 6 0.33 kPa) to 4 weeks (3.546 0.31 kPa), and stabilized at 8 weeks (3.38 6 0.06 kPa)in a pattern consistent with changes in portal pressure. Apositive correlation was observed between spleen stiffnessand D-HVPG (r2 ¼ 0.86; P < 0.01).

Conclusion: These findings indicate a temporal relation-ship between portal hypertension and the development ofliver fibrosis in a large animal model of cholestatic liverdisease. The observed direct correlation between spleenstiffness and D-HVPG suggest a noninvasive MREapproach to diagnose and screen for portal hypertension.

Key Words: magnetic resonance elastography; portalhypertension; spleen stiffness; HVPG; cirrhosisJ. Magn. Reson. Imaging 2011; 34:79–87.VC 2011 Wiley-Liss, Inc.

INCIDENCE OF LIVER fibrosis is increasing world-wide due to nonalcoholic steatohepatitis and hepatitisC (1). As a result, the burden of liver disease is alsoexpanding (2) and includes the clinical sequelae ofliver fibrosis such as portal hypertension. The goldstandard for assessing liver fibrosis is percutaneousliver biopsy. Percutaneous liver biopsy has risks suchas bleeding and other issues such as poor acceptanceby patients. The associated morbidity and mortalityrates of liver biopsy are estimated at 3% and 0.03%,respectively (3). Thus, there is significant demand fornew noninvasive methods to assess liver fibrosis andportal hypertension.

MR elastography (MRE) is a MRI technique basedon quantitatively assessing the mechanical propertiesof tissues based on the propagation of shear waves.MRE is a novel noninvasive technology for diagnosisand monitoring liver stiffness (4). The estimation ofliver stiffness by MRE involves a mechanical tech-nique to generate low frequency (40–120 Hz) shearwaves through the abdomen from which liver fibrosiscan be estimated. A correlation between liver stiffnessand human liver fibrosis has been established (5).Ultrasound elastography has also been proposed as anoninvasive approach for assessing liver fibrosis.However, MRE has a higher technical success rateand a better diagnostic accuracy than ultrasoundelastography for staging liver fibrosis (4,6).

Splenomegaly, occurring in the setting of untreatedportal hypertension, is a common feature of liver cir-rhosis (7). Based on this observation, attempts havebeen made to correlate portal venous pressure andthe hepatic venous pressure gradient (HVPG) withspleen size using ultrasonography and radionuclideimaging (8). Both platelet count and the ratio ofaspartate transaminase to platelet count also havebeen used to predict hepatic venous pressure gradi-ent. Unfortunately, neither spleen size nor platelet-based indices have shown a good correlation with por-tal venous pressure. More encouraging was a recentreport by Talwalkar et al suggesting feasibility of MREin assessing spleen stiffness as a noninvasive mea-sure of portal pressure (9). Portal hypertension wasassessed indirectly in this study by the presence ofesophageal varices on endoscopy or splenic

1Division of Transplantation Surgery, Department of Surgery, MayoClinic, 200 First Street SW, Rochester, Minnesota, USA.2Department of Radiology, Mayo Clinic, 200 First Street SW, Rochester,Minnesota, USA.

Contract grant sponsor: Norwegian Research Council; NIH; Contractgrant number: RO1-DK56733; Contract grant number: EB001981.

*Address reprint requests to: S.L.N., Division of Transplant Surgery,Department of Surgery, 200 First Street SW, Mayo Clinic, Rochester,MN 55905. E-mail: nyberg.scott@mayo.edu

Received July 30, 2010; Accepted March 14, 2011.

DOI 10.1002/jmri.22610View this article online at wileyonlinelibrary.com.

JOURNAL OF MAGNETIC RESONANCE IMAGING 34:79–87 (2011)

CME

VC 2011 Wiley-Liss, Inc. 79

enlargement on ultrasound. Therefore, a correlationbetween splenic stiffness and portal hypertension wasproposed but could not be confirmed because the Tal-walkar study lacked a direct measure of portal pres-sure or HVPG.

Therefore, the current study was designed to con-firm the relationship between splenic stiffness andportal hypertension using a direct pressure measure-ment in a large animal model of cholestatic liver dis-ease. Furthermore, we hypothesized that portalhypertension would precede the onset of advancedliver fibrosis. Our findings suggested that portalhypertension preceded liver fibrosis and portal pres-sure correlated well with spleen stiffness. And lastly,our findings suggest that MRE may serve as a nonin-vasive measure of portal hypertension.

METHODS

Animal Preparation

All animals received care according to the criteria out-lined in the Guide for the Care and Use of LaboratoryAnimals prepared by the National Academy of Scien-ces and published by the National Institutes of Health(publication 86-23, revised 1985). Twenty kilogramadult mongrel canines were observed for at least 7days before the study start, and fasted overnight withfree access to water before general anesthesia. Seestudy outline in Figure 1. An 18 G intravenous (IV)cannula (Surflo, Terumo) was introduced in thebasilic vein of the front leg. Ketamine (10 mg/kg) anddiazepam (0.5 mg/kg) were given intravenously asinduction. The animals were then intubated, andthereafter connected to a Servo Ventilator 900C (Sie-mens). Continuous anesthesia was maintained withisoflurane. Anesthesia depth was assessed by cornealreflex and heart rate. All animals received 1000 mL0.9% NaCl IV. Tidal volume was adjusted by means ofrepeated arterial blood gas analyses to achieve apCO2 within the range 4.0–4.5 kPa during surgery.The bladder was drained by means of a transurethralcatheter and time diuresis was measured. Arterialand venous blood samples were drawn from two 22-Gcatheters (Arrow, Reading, GA) in the femoral artery

and vein. Core body temperature was maintained at37�C with a heating pad and blankets. Animalsreceived IV cefazolin (1 g) perioperatively.

Induction of cholestatic liver disease was performedas previously described (10,11). Laparotomy was per-formed by means of an upper midline incision. Theliver was retracted cephalad to provide exposure ofthe gallbladder. Cholecystectomy was performed. Thecommon bile duct was identified and exposed. Theduct was double ligated with nonabsorbable sutureand divided. The cystic duct was then double ligatedwith nonabsorbable suture. Two catheters wereplaced into the hepatic vein and the portal vein, tun-neled, and thereafter connected to vascular accessports (VAP) positioned subcutaneously paravertebrallywith skin closure over the ports (M.R.I. Hard BaseImplanted Port 9.6 Fr single-lumen catheter, BardAccess Systems, Salt Lake City, UT). Seldinger tech-nique was used for the introduction of the catheters.The point of entry into the portal vein was 2 cm distalto the splenic vein; 4 cm of the catheter was advancedproximally into the vessel. The left hepatic vein wascannulated at a superficial site between the left laterallobe and the median lobe; 3 cm of the catheter wasadvanced into the left hepatic vein. Both catheterswere secured to their respective veins by a purse-string of 5-0 Prolene. Catheter positions were verifiedwith fluoroscopy (Fig. 2).

A low-profile gastrostomy feeding tube (14 Fr and1.5 cm length; MIC-KEY, Kimberly Clark, Roswell, GA)was brought out through the abdominal wall andsecured to the skin with a nylon suture. Additionalsilk sutures were used to pexy the stomach to the ab-dominal wall and to prevent leakage around the g-tube. The abdominal wall was closed in three layers.

Postoperative analgesia was provided for 3 days.Suppressive antibiotics (enrofloxacin 5 mg/kg IV orPO) were administered daily to prevent spontaneousbacterial peritonitis. Pantoprazole IV (1 mg/kg) oromeprazole PO (1 mg/kg) was given daily to preventulcer formation. The vascular access ports wereflushed 3 times per week with 1000 U/mL Heparinand clindamycin (1.5 mg/dL).

The animals were weighed daily. A 500-mL bolus of5% dextrose in normal saline was given by means ofthe hepatic venous port to prevent dehydration duringthe first 10 days after surgery. The animals were fedcooked ground beef (83/17) in addition to cannedfood and dry food. The feeding tube was used toadminister a supplemental enteric diet (Ensure,Abbott Park, IL) when the weight loss exceeded 5%.Paracentesis was performed up to twice weekly whensignificant ascites developed and the animals becamesymptomatic as evidenced by discomfort, or increasedrespiratory effort. Animals loosing in excess of 20%body weight were euthanized and removed from fur-ther study.

Transjugular Liver Biopsies, Hepatic and Portal VeinPressure Measurements

Pressure measurements were obtained through a 19GNon-coring needle (Huber Plus, Bard Access Systems,

Figure 1. Study outline. MR elastography (MRE) wasobtained at baseline before common bile duct ligation(CBDL). Other baseline measurements included a wedge bi-opsy of the liver, liver function blood tests, and pressuremeasurements from the portal vein and hepatic vein. Pres-sure measurements, along with MRE measurements, blooddraws, and transjugular liver biopsies were repeated at 4and 8 weeks.

80 Nedredal et al.

Inc, Salt Lake City, UT) connected to pressure tubingand a pressure transducer (Edwards Lifesciences,Irvine, CA). The transducers were connected to anamplifier (Press Module, M1006B, Hewlett-Packard).All measurements were performed in triplicate. Portalhypertension was defined as an hepatic venous pres-sure gradient >10 mmHg in accordance to a consen-sus definition (12). This definition of portal hyperten-sion (hepatic venous pressure gradient >10 mmHg)has been used previously in canine studies (11).Under general anesthesia, transjugular liver biopsieswere obtained with Seldinger technique by means ofthe external jugular vein using a liver access and bi-opsy set (LABS-200, Cook Medical, Bloomington, IN).Catheter position in the left hepatic vein was verifiedby fluoroscopy and injection of contrast (NovaplusOmnipaque, GE Healthcare, Princeton, NJ) beforeperforming the biopsies. Liver tissue was fixed in buf-fered formalin, embedded in paraffin, and stainedwith hematoxylin and eosin (H&E) or Masson’s tri-chrome. Liver biopsy specimens were considered suit-able for fibrosis staging when they contained at least10 portal tracts or obvious regenerating nodules. Thestage of fibrosis was evaluated semiquantitativelyusing Masson’s trichrome staining according to theMETAVIR scoring system (13); with this score, F0 rep-resents no fibrosis, F1 portal fibrosis without septa;F2 portal fibrosis and few septa, F3 numerous septawithout cirrhosis, and F4 cirrhosis.

MRE

MRE is an MR technique for quantitative assessment ofthe mechanical properties of soft tissues. These me-chanical properties are quantified by their propagationof shear waves using MRE. We applied a three-dimen-sional/three axis (3D/3-axis) spin echo based EPI MREsequence to assess the mechanical properties of thespleen and liver in vivo. Animals were kept under gen-eral anesthesia with isoflurane, and connected to a ven-tilator. The imaging was done in the supine positionwith two acoustic pressure-activated drivers placedagainst the body wall adjacent to the liver and the spleenas illustrated in Figure 3. MRE examinations were per-formed with a 1.5 Tesla (T) whole-body GE imager(Signa, GE Medical System, Milwaukee, WI). A single-shot 3D/3-axis echo planar imaging (EPI) MREsequence was used to collect 40 axial wave imagesshown in Figure 4. Other imaging parameters are listedin Figure 4. All the trapezoidal motion-encoding gra-dients were applied with zeroth and first moment nullingalong the through-plane direction. Two spatial presatu-ration bands were applied on each side of the selectedslice to reduce motion artifacts from blood flow. Due tochemical shift artifacts, spatial spectral pulses are usedto generate the initial 90� radio frequency (RF) pulse.Increased signal and better wave depiction wereobtained with the EPI sequence. Susceptibility artifactsassociated with EPI based sequences were not a prob-lem in the liver and spleen applications.

Figure 2. Fluoroscopic confir-mation of portal vein and he-patic vein catheters. A: Asubcutaneous vascular accessport (VAP) positioned paraverte-brally. Contrast was injectedwith a non-coring needlethrough the skin and into thelumen of the port. B: A porto-gram during injection of con-trast into the portal veincatheter. C: The hepatic veincatheter and passage of con-trast into the suprahepatic venacava. The portal vein cathetercan be observed on the rightside of the image.

Spleen MRE Predicts Portal Hypertension 81

The acquired wave images were then processedusing an inversion algorithm to generate quantitativeimages called elastograms that depict tissue stiffness.Preprocessing algorithms for the wave data includedremoval of concomitant gradient field effects, phaseunwrapping with minimum discontinuity, removal oflongitudinal wave propagation with curl filtering and20 evenly spaced 3D directional filtering (14) to

enhance the accuracy of the elastograms. Tissue stiff-ness was estimated by a 3D local frequency estima-tion inversion algorithm (15) using the shear waveestimate obtained within each region. The elasto-grams were analyzed by measuring mean shearstiffness within a large, manually specified region ofinterest that included an entire cross-sectional imageof liver or spleen, while excluding major blood vesselsof width greater than six pixels (approximately 8 mm).Regions without adequate magnitude signal or waveamplitude were excluded as well because these lowamplitude regions result in erroneous stiffness values.

Statistics

We used the SPSS 11.0 statistical package to estab-lish significance between the groups (Chicago, IL). Allresults were expressed as mean 6 standard deviation(SD). The paired Student t-test was applied for com-parisons of normally distributed variables. P values<0.05 were considered significant. Pearson correla-tion coefficients were estimated between spleen/liverstiffness and D-HVPG. Repeat measures models withan equal correlation, within-subject structure wereused to test for associations between stiffness andD-HVPG. A multivariate repeat measures model wasfit to estimate the adjusted effects of each stiffnessmeasurement.

RESULTS

Study Outline

Eight animals were included in this study. Six animalsunderwent surgical intervention (catheter placement

Figure 3. MRE in a large animal model. The MRE apparatus is illustrated including a mechanical system for application ofshear waves through the abdomen. The animals were intubated and kept under general anesthesia. Two small separate driv-ers were strapped around the abdomen (as opposed to one large driver applied to the abdomen of humans) due to the V-shaped form of the canine torso. During data acquisition, acoustic pressure waves (60 Hz) were generated by an active pneu-matic driver located outside of the magnetic field of the whole body imager and conveyed by means of a flexible tube to thetwo passive pneumatic drivers strapped around each side of the upper abdomen. [Color figure can be viewed in the onlineissue, which is available at wileyonlinelibrary.com.]

Figure 4. Spin echo EPI MRE pulse sequence. Continuousvibrations were delivered throughout the abdomen at 60 Hz.Imaging parameters were listed as following: TR/TE ¼ 1634/48 ms, 3 � 3 � 3 mm3 imaging resolution, parallel imagingacceleration factor of two, one pair of motion encoding gradi-ent with the duration of 12.5 ms and sensitivity of 25.7 mm/p (radians), 4 phase offsets, and 118 s acquisition time splitinto eight intervals of suspended respiration. ACQ, acquisi-tion; MEQ, motion encoding gradient; G, gauss.

82 Nedredal et al.

and CBDL) and two control animals underwent cathe-ter placement and sham biliary surgery. Both controlanimals completed the 8 week study, while only two ofthe intervention animals completed the 8-week studyby maintaining a weight above 80% of baseline. Twointervention animals were killed at 5 and 6 weeks afterCBDL due to weight loss. One intervention animal waskilled 3 weeks after CBDL due to a bile duct leak. Oneintervention animal died of a bleeding complication 3days after CBDL. All animals were studied at baselineand served as their own internal control. Baselinemeasurements obtained before CBDL included mag-netic resonance elastography, blood draws, D-HVPGmeasurements, and wedge liver biopsies (Fig. 1). MRE,blood draws, transjugular liver biopsies, and D-HVPGmeasurements were repeated 4 and 8 weeks afterCBDL.

Liver Function

Liver function tests obtained at baseline and at 4weeks and 8 weeks after CBDL are summarized in Ta-ble 1. Animals developed progressive hyperammone-mia after CBDL. The levels of ammonia increasedmore than six-fold (P < 0.05). Total bilirubin rose 50-fold during the first 4 weeks (P < 0.05), and stabilizedthereafter. ALT levels exhibited a similar pattern risingover 35-fold in the first 4 weeks (P < 0.05). The ani-mals developed progressive hypoalbuminemia (P <0.05), suggesting decreased synthetic function andsuboptimal nutrition of chronic liver disease despitesupplemental feeding. International normalized ratio(INR), another measure of hepatic synthetic function,increased mildly from baseline. The Child-Pugh scor-ing system was used an overall assessment of liverdysfunction (16). As expected, Child-Pugh score rosewith all animals reaching stage C at 8 weeks.

Histological Findings

The liver histology was graded according to the META-VIR classification (13). According to this scoring sys-tem, F0 represented no fibrosis, F1 portal fibrosiswithout septa, F2 portal fibrosis and few septa, F3bridging fibrosis connecting the portal triads, and F4cirrhosis. Liver fibrosis was identified by both H&Eand Masson’s trichrome staining of liver tissue (Fig.5). After 4 weeks F1 fibrosis developed around theportal triads in all liver biopsies. Fibrosis progressedto bridging fibrosis (F3) in both livers examined at 8weeks. One animal was biopsied after six weeks withportal fibrosis and a few septae (F2).

Pressure Measurements

Portal vein and hepatic vein pressures were measuredby means of vascular access ports positioned subcu-taneously with their catheter tips positioned in the re-spective veins (Fig. 2). The baseline pressures for theportal and hepatic veins were 6.6 6 2.2 mmHg and2.5 6 2.6 mmHg, the 4 week pressures were 11.0 65.1 mmHg and �1.5 6 3.4 mmHg, and the 8 weekpressures were 11.3 6 3.2 mmHg and �2.0 6 3.6mmHg, respectively. All venous pressure measure-ments were obtained under general anesthesia inhydrated dogs. Hydration was standardized by intra-venous infusion of saline to a systolic blood pressuregreater than 70 mmHg.

Direct Hepatic Venous Pressure Gradient (D-HVPG)was determined by subtracting hepatic vein pressurefrom portal vein pressure (D-HVPG ¼ PVP – HVP). Astriking finding was the significant (P < 0.05) three-fold increase in D-HVPG from baseline to 4 weeks(Fig. 6). D-HVPG increased only slightly after 4weeks.

Table 1

Laboratory and Clinical Findings

Weeks 0 4 8

Ammonia (mg/dL)CBDL 35.3 6 16.5 133.7 6 77.4* 237.7 6 127.5*

Control 36.2 6 13.2 37.8 6 16.9 37.5 6 15.4

Total bilirubin (mg/dL)

CBDL 0.2 6 0.1 11.0 6 4.1* 10.2 6 8.9*

Control 0.3 6 0.1 0.2 6 0.1 0.3 6 0.2

Albumin (g/dL)

CBDL 3.1 6 0.2 2.3 6 0.5 1.6 6 0.3*

Control 3.0 6 0.2 3.2 6 0.3 3.3 6 0.3

ALT (U/L)

CBDL 46.3 6 23.2 1679.3 6 470.1* 891.0 6 304.0*

Control 48.1 6 22.9 50.1 6 24.3 53.4 6 22.3

INR

CBDL 0.8 6 0.1 1.1 6 0.1 1.1 6 0.1

Control 0.8 6 0.1 0.7 6 0.0 0.8 6 0.1

Creatinine (mg/dL)

CBDL 0.7 6 0.2 0.6 6 0.1 0.7 6 0.2

Control 0.8 6 0.1 0.8 6 0.2 0.8 6 0.1

Child-Pugh score

CBDL 0 B C

Control 0 0 0

*Significant difference between CBDL group and Control group (P < 0.05).

Spleen MRE Predicts Portal Hypertension 83

MRE

The mean spleen stiffness increased more than two-fold from baseline to 4 weeks (1.72 6 0.33 kPa and3.54 6 0.31 kPa; P < 0.05) (Fig. 7). The mean liverstiffness also increased from baseline to 4 weeks(1.50 6 0.15 kPa and 3.49 6 0.75, respectively).Mean stiffnessess of the spleen and the liver did not

increase from 4 weeks to 8 weeks (3.38 6 0.06 kPaand 3.37 6 0.05, respectively).

Regression Analysis

Positive correlations were observed between spleenstiffness measurement and D-HVPG (r2 ¼ 0.86; P <0.01; y ¼ 3.8225x–1.2795) and between liver stiffnessmeasurement and D-HVPG (r2 ¼ 0.90; P < 0.01; y ¼3.9058x–0.7208) as illustrated in Figure 8.

DISCUSSION

This study established a 3D/3-axis MRE imagingmethod in a large animal model for estimating mechan-ical properties of abdominal organs. We observed atemporal relationship between the rise in portal pres-sure and the development of liver fibrosis using a largeanimal model of portal hypertension. In addition, ourresults suggest important correlations between spleenstiffness and liver stiffness and D-HVPG (i.e., portalhypertension), suggesting a noninvasive approach tothe diagnosis and serial examinations of patients withportal hypertension using MRE. Furthermore, we dem-onstrated that portal hypertension (determined bydirect measure of HVPG) preceded liver fibrosis. Thesefindings may contribute to a better understanding ofthe pathophysiology of portal hypertension.

Figure 5. Liver histology. A–C:Low power magnification of ca-nine liver stained with Masson’strichrome at baseline (A), 4weeks (B), and 8 weeks (C). Notethe mild patches of blue stainingaround the portal triads indicat-ing an F1 fibrosis after 4 weeks(B), and progression to bridgingfibrosis (F3) after 8 weeks (C). D:A nutmeg appearance of thesliced liver characteristic ofcholestatic liver disease wasobserved after 8 weeks (D).[Color figure can be viewed inthe online issue, which is avail-able at wileyonlinelibrary.com.]

Figure 6. Direct hepatic venous pressure gradient. D-HVPGwas obtained using subcutaneous vascular ports with theircatheter tips positioned in the portal and hepatic veins. TheD-HVPG increased significantly from baseline to 4 weeks (P< 0.05). Only a slight increase in D-HVPG was detected from4 to 8 weeks. There was no increase in D-HVPG in the con-trols. D-HVPG, direct hepatic venous pressure gradient;CBDL, common bile duct ligation.

84 Nedredal et al.

Figure 7. MRE. Anatomical images (left column) and wave images (middle column) were obtained at baseline, 4 weeks, and 8week. Maximum intensity projection elastograms (right column) were determined from the wave image data at these timepoints. After 8 weeks, liver and spleen were separated due to displacement of the spleen in the lower abdomen by ascites.Also, note the indentations made by the passive drivers on the anterior surfaces of the abdomen during data collection atweek 8.

Figure 8. Regression analysis. Regression analysis between stiffness measurements and D-HVPG (direct hepatic venouspressure gradient). Regression values correlating spleen stiffness with D-HVPG and liver stiffness with D-HVPG were r2 ¼0.86 (P < 0.01) and r2 ¼ 0.90 (P < 0.01), respectively.

Spleen MRE Predicts Portal Hypertension 85

Spleen Stiffness Correlates With PortalHypertension

Our study showed that splenic stiffness increasedbefore significant (F3) liver fibrosis was observed. So,which morphological changes are observed in thespleen among patients progressing to chronic liverdisease? Observed changes include pulp hyperplasia,congestion due to increased blood flow, and eventu-ally fibrosis (17–19). Our results suggest that splenicstiffness as determined by MRE can reflect thesechanges and may serve as an indirect and noninva-sive measure of portal hypertension. However, a limi-tation of the results are the lack of a broad range ofd-HVPG sampling and MRE imaging in the intervalsbetween sampling.

A relationship between spleen stiffness and portalhypertension was hypothesized, but remained uncon-firmed before our current report. Direct confirmationof this relationship has been difficult because meas-urements of portal pressures required invasive cathe-terization, a procedure considered too risky inhumans. Therefore, Talwalker et al diagnosed portalhypertension indirectly by physical findings such asesophageal varices and splenic enlargement on endos-copy and ultrasound imaging (9). A direct measure-ment of portal pressure synchronously with spleenstiffness was required to validate this relationship.Accordingly, with verification of this relationship ourstudy suggests a noninvasive approach to the diagno-sis and serial examination of patients with portalhypertension using MRE. A potential limitation of thisstudy is the necessity of general anesthesia and me-chanical ventilation. The animals were disconnectedfrom the ventilator for 10–20 s to simulate breathholdand to obtain high-quality MRE images.

Taking into account the high prevalence of hepatitisC worldwide, the emergence of nonalcoholic fatty liverdisease, and excessive alcohol consumption, the inci-dence of cirrhosis and portal hypertension is expectedto increase significantly in the next decade. Untreatedportal hypertension (HVPG > 10–12 mmHg) increasesthe risk of serious complications such as ascites, he-patic encephalopathy, and bleeding complications (7).With proper pharmacological treatment (HVPG < 12mmHg) the risk of fatal complications such as varicealbleeding can be reduced. Therefore, a noninvasivemodality could be useful in mass screening and diag-nosis of patients with portal hypertension and tomonitor their response to therapy.

Portal Hypertension

We also observed the development of sustained portalhypertension and a significant hepatoportal venouspressure gradient within 4 weeks of CBDL, well beforethe development of a significant fibrotic response inthe liver. Moreover, but not until 8 weeks after CBDLwas bridging fibrosis (F3) present on histological ex-amination. In our study, portal pressure correlatedwith stiffness of both the liver and spleen. There areproposed simultaneous architectural and pathophys-iological mechanisms in the development of portal

hypertension (20). Some studies have suggested thatportal hypertension precedes the development offibrosis (21,22). Pressure is a function of vessel resist-ance multiplied by vessel flow. Therefore, portalhypertension is not only related to the degree of fibro-sis (increased resistance) but also to the flow. Accord-ingly, both resistance and flow contribute to the stateof portal hypertension (23). Cirrhotic liver disease ischaracterized by a hyperdynamic state, i.e. increasedflow, as well as variable flow and variable resistance(24). Cirrhotic liver disease is also associated withprogressive vasodilatation in the splanchnic circula-tion and extrahepatic organs. These changes may beheightened if combined with infection, sepsis andmulti-organ failure syndrome.

Liver fibrosis did not correlate with increased liverstiffness suggesting that other factors were involvedsuch as cholestasis and an inflammatory response.A study indicated that extrahepatic cholestasisincreased liver stiffness regardless of fibrosis (25).Activation of stellate cells has been proposed as anearly event in portal hypertension leading to anincreased secretion of a powerful vasoconstrictor;endothelin-1 (26). Increased nitric oxide release insplanchnic arteries has also been proposed as amechanism of portal hypertension in rats (27). More-over, conditions of prolonged liver injury, such aschronic cholestasis and repeated toxic exposures,lead to liver fibrosis development and are associatedwith increased production of major matrix proteins(collagen, fibronectin, proteoglycans, and laminin)(26).

Justification of a Large Animal Model of PortalHypertension

The animals developed cirrhotic cholestatic disease by8 weeks (Fig. 5). A large animal model of liver diseasehad several benefits in our study. These benefitsincluded the use of a full-size clinical MRE systemand placement of chronic indwelling catheters forrepeated measures of portal pressure. The same ani-mal could also be used for multiple liver biopsies andrepeat blood draws. A large animal model also pro-vided a dynamic living system more consistent withhuman hepato-portal physiology in terms of both por-tal pressure and splenic and hepatic stiffnesses. Asimilar study performed in canines showed a portalhypertension of 13.3 mmHg and HVPG of 11.1 mmHgafter 8 weeks. Unlike our model with continuousdirect access to the portal and hepatic veins, the por-tal pressure measurements after 8 weeks required are-laparotomy to gain direct access to the portal vein(11). Clinical MRE systems cannot be validated inrodents. The magnitude and variability of portal pres-sure profiles of cirrhotic humans are more similar tocirrhotic dogs than cirrhotic rodents. For example,mean portal pressures rarely exceed 10 mmHg inmice during the development of cirrhosis (28). In con-trast, all four study dogs had portal pressures rangingfrom 15–30 mmHg 4 weeks after CBDL. The magni-tude of portal pressures in canine cirrhosis were com-parable to human portal hypertension. Large animals

86 Nedredal et al.

also allowed us to maintain patency of chronic in-dwelling catheters placed in the hepatic and portalveins. Portal pressures and MRE images wereobtained serially from the same large animal. In con-trast, chronic indwelling portal vein and hepatic veincatheters were not possible in rodents (28,29).

In conclusion, the application-specific 3D/3-axisabdominal MRE in a large animal model has beenestablished and validated in this investigation. Goodcorrelations were observed between portal pressure,D-HVPG and both spleen stiffness and liver stiffnessin our large animal study of cholestatic liver diseaseand portal hypertension. These findings suggest atemporal relationship between portal hypertensionand the development of liver fibrosis. Moreover, thesefindings suggest a noninvasive approach to the diag-nosis and serial examinations of patients with portalhypertension using MRE. Further study of this noveltechnology is warranted.

ACKNOWLEDGMENTS

Funding for the study was provided by the NorwegianResearch Council, the National Institutes of Health,and a Career Development Award from the MayoFoundation, Rochester, MN. MRI compatible vascularaccess ports were generously provided by Bard AccessSystems, Salt Lake City, UT. Liver Access Biopsy Setswere generously provided from Cook Medical, Bloo-mington, IN.

REFERENCES

1. Lauer GM, Walker BD. Hepatitis C virus infection. N Engl J Med2001;345:41–52.

2. Kim WR, Brown RS Jr, Terrault NA, El-Serag H. Burden of liverdisease in the United States: summary of a workshop. Hepatology2002;36:227–242.

3. Piccinino F, Sagnelli E, Pasquale G, Giusti G. Complications fol-lowing percutaneous liver biopsy. A multicentre retrospectivestudy on 68,276 biopsies. J Hepatol 1986;2:165–173.

4. Huwart L, Sempoux C, Vicaut E, et al. Magnetic resonance elas-tography for the noninvasive staging of liver fibrosis. Gastroenter-ology 2008;135:32–40.

5. Talwalkar JA, Yin M, Fidler JL, Sanderson SO, Kamath PS,Ehman RL. Magnetic resonance imaging of hepatic fibrosis:emerging clinical applications. Hepatology 2008;47:332–342.

6. Talwalkar JA. Elastography for detecting hepatic fibrosis: optionsand considerations. Gastroenterology 2008;135:299–302.

7. Turnes J, Garcia-Pagan JC, Abraldes JG, Hernandez-Guerra M,Dell’Era A, Bosch J. Pharmacological reduction of portal pressureand long-term risk of first variceal bleeding in patients with cir-rhosis. Am J Gastroenterol 2006;101:506–512.

8. Shah SH, Hayes PC, Allan PL, Nicoll J, Finlayson ND. Measure-ment of spleen size and its relation to hypersplenism and portalhemodynamics in portal hypertension due to hepatic cirrhosis.Am J Gastroenterol 1996;91:2580–2583.

9. Talwalkar JA, Yin M, Venkatesh S, et al. Feasibility of in vivo MRelastographic splenic stiffness measurements in the assessmentof portal hypertension. AJR Am J Roentgenol 2009;193:122–127.

10. Rutgers HC, Stradley RP, Johnson SE. Serum bile acid analysisin dogs with experimentally induced cholestatic jaundice. Am JVet Res 1988;49:317–320.

11. Bosch J, Enriquez R, Groszmann RJ, Storer EH. Chronic bileduct ligation in the dog: hemodynamic characterization of a por-tal hypertensive model. Hepatology 1983;3:1002–1007.

12. de Franchis R. Updating consensus in portal hypertension: reportof the Baveno III Consensus Workshop on definitions, methodol-ogy and therapeutic strategies in portal hypertension. J Hepatol2000;33:846–852.

13. METAVIR. Intraobserver and interobserver variations in liver bi-opsy interpretation in patients with chronic hepatitis C. TheFrench METAVIR Cooperative Study Group. Hepatology 1994;20(Pt 1):15–20.

14. Manduca A, Lake DS, Kruse SA, Ehman RL. Spatio-temporaldirectional filtering for improved inversion of MR elastographyimages. Med Image Anal 2003;7:465–473.

15. Manduca A, Oliphant TE, Dresner MA, et al. Magnetic resonanceelastography: non-invasive mapping of tissue elasticity. MedImage Anal 2001;5:237–254.

16. Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R.Transection of the oesophagus for bleeding oesophageal varices.Br J Surg 1973;60:646–649.

17. Manenti A, Botticelli A, Gibertini G, Botticelli L. Experimentalcongestive splenomegaly: histological observations in the rat.Pathologica 1993;85:721–724.

18. Cavalli G, Re G, Casali AM. Red pulp arterial terminals in conges-tive splenomegaly. A morphometric study. Pathol Res Pract 1984;178:590–594.

19. Terayama N, Makimoto KP, Kobayashi S, et al. Pathology of thespleen in primary biliary cirrhosis: an autopsy study. Pathol Int1994;44:753–758.

20. Bosch J, Garcia-Pagan JC. Complications of cirrhosis. I. Portalhypertension. J Hepatol 2000;32(Suppl):141–156.

21. Rockey D. The cellular pathogenesis of portal hypertension: stel-late cell contractility, endothelin, and nitric oxide. Hepatology1997;25:2–5.

22. Rockey DC. Noninvasive assessment of liver fibrosis and portalhypertension with transient elastography. Gastroenterology 2008;134:8–14.

23. Iwakiri Y, Groszmann RJ. The hyperdynamic circulation ofchronic liver diseases: from the patient to the molecule. Hepato-logy 2006;43(Suppl 1):S121–S131.

24. Kowalski HJ, Abelmann WH. The cardiac output at rest in Laen-nec’s cirrhosis. J Clin Invest 1953;32:1025–1033.

25. Millonig G, Reimann FM, Friedrich S, et al. Extrahepatic choles-tasis increases liver stiffness (FibroScan) irrespective of fibrosis.Hepatology 2008;48:1718–1723.

26. Rockey DC. Hepatic fibrosis, stellate cells, and portal hyperten-sion. Clin Liver Dis 2006;10:459–479, vii–viii.

27. Hori N, Wiest R, Groszmann RJ. Enhanced release of nitric oxidein response to changes in flow and shear stress in the superiormesenteric arteries of portal hypertensive rats. Hepatology 1998;28:1467–1473.

28. Geerts AM, Vanheule E, Praet M, Van Vlierberghe H, De Vos M,Colle I. Comparison of three research models of portal hyperten-sion in mice: macroscopic, histological and portal pressure evalu-ation. Int J Exp Pathol 2008;89:251–263.

29. Cardenas A, Lowe R, Oh S, et al. Hemodynamic effects of sub-stance P and its receptor antagonist RP67580 in anesthetizedrats with carbon tetrachloride-induced cirrhosis. Scand J Gastro-enterol 2008;43:328–333.

Spleen MRE Predicts Portal Hypertension 87