Improved hepatocyte function offuture liver remnant of cirrhotic ratsafter portal vein ligation: A bonusother than volume shiftingKun-Ju Lin, MD, PhD,a Chien-Hung Liao, MD,b Ing-Tsung Hsiao, PhD,d Tzu-Chen Yen, MD, PhD,a

Tse-Ching Chen, MD, PhD,c Yi-Yin Jan, MD,b Miin-Fu Chen, MD,b and Ta-Sen Yeh, MD, PhD,b

Taoyuan, Taiwan

Background. Preoperative portal vein embolization is increasingly employed for those with hepatocellularcarcinoma and cirrhosis to gain a volume-shifting effect. However, the alterations of histologicarchitecture and hepatocyte function of future liver remnant (FLR) remain unexplored.Methods. Portal vein ligation (PVL) was performed in cirrhotic and noncirrhotic rats. Regenerationindices that include the DNA synthesis index, restituted liver mass, and the redistributed volume ratiowere measured. The indocyanine green 159 retention test (ICG-R15), histologic changes, total Knodellscore, and activated hepatic stellate cells (HSCs) were measured before and after PVL. Tc-99m sulfur-colloid liver single photon emission computed tomography (SPECT) and diisopropyl iminoacetic acid(DISIDA) SPECT were conducted.Results. The redistributed volume ratio of cirrhotic rats was less than noncirrhotic rats (63% vs 80%,P < .01). The ICG-R15 of cirrhotic rats at day 7 after PVL was improved compared with baseline (6.0 ±4.1% vs 15.8 ± 4.6%, P < .01). The total Knodell score and activated HSCs of FLR in cirrhotic ratsboth were decreased compared with those of baseline. The redistributed volume ratio of noncirrhotic andcirrhotic rats based on 99mTc sulfur-colloid SPECT were 79% and 64%, respectively. The clearanceT1/2 of FLR in cirrhotic rats based on DISIDA SPECT was decreased compared with baseline (5.2 ± 1.9min vs 8.6 ± 3.1 min).Conclusion. The regenerated functional liver mass of cirrhotic rats after PVL is less than noncirrhoticrats, whereas the hepatocyte function of FLR in cirrhotic rats is improved relevant to tissue remodeling.(Surgery 2009;145:202-11.)

From the Departments of Nuclear Medicine and Molecular Imaging Center,a Surgery,b and Pathology,c

Chang Gung Memorial Hospital, Taoyuan, Taiwan; and the Department of Medical Imaging andRadiological Sciences, Chang Gung University,d Taoyuan, Taiwan

PORTAL VEIN EMBOLIZATION has been increasinglyapplied prior to major hepatic resection for he-patocellular carcinoma, bile duct malignancies,and metastatic liver tumors.1-5 The procedureacts to embolize the portal branches of the lobetargeted for resection and induces compensatory

K-J.L. and C-H.L. both contributed equally to this article.

Supported by Grants NMRPG366031, 96-2314-B-182A-008-MY2from the National Medical Research Program, Taiwan.

Accepted for publication October 21, 2008.

Reprint requests: Ta-Sen Yeh, MD, PhD, Department of Surgery,Chang Gung Memorial Hospital, 5 Fu-Hsing Street, Kwei-ShanShiang, Taoyuan, Taiwan. E-mail: Surglab@adm.cgmh.org.tw;tsy471027@adm.cgmh.org.tw.

0039-6060/$ - see front matter

� 2009 Mosby, Inc. All rights reserved.

doi:10.1016/j.surg.2008.10.006

202 SURGERY

hypertrophy of the nonembolized lobe and futureliver remnant (FLR) with the aim of minimizingpostoperative liver dysfunction and of increasingthe resectability of the disease. It is important topredict hypertrophy of FLR after portal vein embo-lization to decide a subsequent safe hepatectomy.The amount of liver regeneration concerningFLR after portal vein embolization varies enor-mously, with an increased volume of 7--79%.1-6

Given that chronically damaged livers are believedto be less able to regenerate, evaluating the effectsof portal vein embolization toward normal and ab-normal liver parenchyma separately is warranted.Cirrhosis has been considered one of the most im-portant prognostic parameter for liver surgery be-cause of inherent bleeding tendency, impairedliver regeneration, and poor liver reserve functionafter hepatectomy.6 Many centers have reported a

SurgeryVolume 145, Number 2

Lin et al 203

promising surgical result of preoperative portalvein embolization on patients with cirrhosis/liverfibrosis, with an emphasis on ultimately increasingthe volume of FLR.6-9 Although a change of livervolume can be determined with serial computedtomography (CT),10 the alterations of histologicarchitecture and lobar function in parallel withvolume change of FLR remain unexplored. Theaim of this study was to fulfill these intriguingissues.

MATERIALS AND METHODS

Cirrhosis model induced by thioacetamide. Theexperimental protocol was approved by the EthicsReview Committee for Animal Experimentation ofChang Gung University. The study used maleSprague-Dawley rats that weighed between 200 gand 250 g. The rats were housed under a 12-hlight/dark cycle with free access to a standard ratchow and water. Cirrhosis was induced by oraladministration of thioacetamide (0.03% in tapwater) for 16 weeks, as previously described.11

This long-term administration of thioacetamidecauses characteristic micronodular cirrhosis inthe animal livers.

Portal vein ligation (PVL). Under general anes-thesia with an intraperitoneal injection of keta-mine, the portal branches of cirrhotic ornoncirrhotic rats that perfuse the median andleft lobes of the liver were ligated with fine silk.12

Animals of the sham-operated group were sub-jected to laparotomy, and their livers were manipu-lated. All cirrhotic and sham-operated ratscontinued to be fed after PVL with a thioaceta-mide-containing diet until they were killed to pre-vent spontaneous recovery from cirrhosis.13 Atindicated times (days 0, 1, 3, 5, and 7) after PVL,the animals with and without cirrhosis as well asthe sham-operated cirrhotic rats were killed, andthe blood and livers were collected for analysis(n = 6 at least for each group and each time point).The ligated lobes (left and median lobes) and un-ligated lobes (right and caudate lobes) were re-trieved and weighed, respectively. A portion ofthe unligated lobes of the liver was immediatelyfrozen in liquid nitrogen and stored at --80�C forsubsequent investigation. For histologic and immu-nohistochemical examination, the tissue sampleswere fixed in 10% neutral-buffered formalin, pro-cessed with graded ethanol solutions, and embed-ded in paraffin.

Conventional liver function tests. The serumalbumin, alanine aminotransferase (ALT), andtotal bilirubin of the experimental animals weredetermined prior and at day 1 and day 7 after PVL.

Indocyanine green 15-min retention test (ICG-R15). ICG (0.5 mg/kg of body weight) was givenvia the femoral vein. The samples of venous bloodwere taken before and 15 min after the injectionvia the contralateral femoral vein. Each sample wasdiluted to one fourth of the concentration withsaline, and the specimens were analyzed for ICGconcentration in a spectrophotometer at 805 nm.The ICG-R15 was performed prior and at day 7after PVL in experimental rats.

BrdU (bromodeoxyuridine) assay for evaluationof DNA synthesis. The experimental animals re-ceived BrdU (Sigma, St. Louis, Mo) at a dose of30 mg/kg body weight by intraperitoneal injection2 h before they were killed. The livers wereremoved, fixed in 95% ethanol that contained1% acetic acid at 4�C for 2 h, set overnight inethanol 4�C, and processed to paraffin sections.Incorporated BrdU was immunohistochemicallystained using a monoclonal anti-BrdU antibody(Becton Dickson, Mountain View, Calif). Ten ran-domly chosen visual fields per section were exam-ined at a magnification of 3 200. The labelingindex was defined as the mean number of BrdU-positive cells per HPF (3 200).

Evaluation of liver regeneration. The normal-ized restituted liver mass (%) after PVL was calcu-lated using the following formula: (weight of [rightlobe ± caudate lobe]/body weight at day 7)/(weight of [right lobe ± caudate lobe]/body weightat day 0) 3 100%, which is based on the results ofautopsy data. The redistributed volume ratio ofunligated lobes to the whole liver after PVL at eachtime point was calculated using the formula:weight of right lobe + caudate lobe/weight of thewhole liver 3 100%, which is based on the resultsof autopsy data.

Histologic evaluation. The hepatitis activity in-dex (HAI), which was designed by Knodell andDesmet,14,15 was used to grade the severity of thenecroinflammation and fibrosis. HAI comprises 4separate scores, which include periportal necrosiswith and without bridging necrosis (0--10), intra-lobular degeneration and focal necrosis (0--4), por-tal inflammation (0--4), and fibrosis. Additionally,Masson trichrome staining was performed to facil-itate fibrosis scoring. The Knodell HAI of each an-imal was assessed by examining 15 consecutiveslides per paraffin-embedded block by 2 seniorsurgical pathologists (T.C.C. and T.S.Y.). The Mas-son-trichrome stained area was measured by a com-puter-assisted image-analysis system with MacScope software (Mitani Co., Fukui, Japan), andthe relative area of fibrosis to the whole liver wascalculated.11

SurgeryFebruary 2009

204 Lin et al

Immunohistochemical staining for alfa-smoothmuscle actin (a-SMA). The paraffin-embeddedliver sections (4 mm thick) were deparaffinized inxylene, rehydrated in decreasing concentrations ofethanol in water, and treated with 3% hydrogenperoxidase in methanol for 30 min. The tissuesections were processed in 10 mmol/L citratebuffer (pH = 6.0) and heated to 120�C in anautoclave for 10 min. The sections were thenincubated with primary monoclonal antibodiesagainst a-SMA (1/500; Sigma-Aldrich Corp.,St. Louis, Mo). The sections were subsequentlyincubated with biotinylated anti-mouse immuno-globulins for 30 min and treated with an avidin-biotin complex for 30 min at room temperature.3-Amino-9-ethyl carbazole was used as the colorreagent, and hematoxylin was used as a counter-stain. Quantitative expression of a-SMA was mea-sured by a computed assisted image analysis systemwith Mac Scope software (Mitani Co.).11

In vivo liver-function image study. A high-reso-lution dual-modality system NanoSPECT/CT (Bio-scan, Inc., Washington, DC) was used for theanimal image study. This system features helicalscanning for both single-photon emission com-puted tomography (SPECT) and CT acquisitionsusing a translation stage with variable axialscanning range from 30 to 300 mm. The multi-ple-pinhole SPECT system is equipped with 4 large-area detectors of 230 3 215 mm each and withmultiple-pinhole (nine 2.5-mm pinholes) collima-tors. Overlapping projections are produced toincrease the sensitivity in the multiple-pinholeapertures. Three sessions of image studies wereperformed 2 days before operation as well as atdays 1 and 7 after operation. Each image sessionincluded 1 Tc-99m diisopropyl iminodiacetic acid(DISIDA) dynamic SPECT image for hepatobiliaryfunction assay and 1 Tc-99m sulfur-colloid image todetermine the liver function volume. The func-tional liver volume studies were performed 15 minafter tail vein injection of 2 mCi/0.2 mL Tc-99msulfur colloid for 30 min. To evaluate the hepato-biliary function, the Tc-99m DISIDA dynamicSPECT images (at 2 min per frame) were acquiredimmediately after tail vein injection of 2 mCi/0.2mL radiotracer for 16 min. All SPECT images werereconstructed using the iterative ordered subsetexpectation maximization method from overlap-ping projects.

Functional liver volume assay. All SPECT imagedata were processed on a PMOD image analysisworkstation (PMOD Technologies Ltd., Zurich,Switzerland). To determine the total liver functionvolume, the whole liver volume of interest (VOI)

was drawn automatically on a threshold-basedalgorithm using 20% of the maximum liver valueas cutoff on a 3-dimensional (3D) Tc-99m sulfur-colloid image. The redistributed liver functionvolumes were determined using the same thresh-old-based algorithm within a predefined liver maskthat contained the right and caudate lobes. Theliver masks were drawn by a single trained surgeon(T.S.Y.) with familiarity of the rodent liver anatomy.The redistributed volume ratios were calculatedusing the following formula: the redistributedvolume ratio = (right plus caudate lobe functionvolume)/total liver function volume.

Hepatobiliary function assay. The hepatobiliaryfunction assay was measured from the dynamic Tc-99m DISIDA study. To define the VOI for the time-activity curve assay, the first 2 frames of thedynamic study were summed together for bettervisualization. Two sets of 3D masks, in which 1contains right and caudate lobes and the othercontains left and median lobes, were predefined bythe same surgeon. Later, 2 liver VOIs were thandrawn automatically on a threshold-based algo-rithm using 20% of the maximum liver value ascutoff within the above-mentioned liver masks onthe Tc-99m DISIDA SPECT images. The time-activity curves from the 2 VOIs were generatedand corrected for radioactivity decay. The meanradioactivity values were normalized to their max-imum for comparison. A quantitative analysis ofthe time-activity curve was performed to measuretime-point-to-peak counts and to calculate excre-tion of T1/2 by nonlinear regression using 1-phaseexponential decay. For curve fitting, the initialvalues were automatically calculated, and the pla-teau constant was fixed to 0.

Statistical analysis. All continuous variables wereexpressed as the mean ± standard deviation. Astatistical analysis for continuous variables wasperformed using a Student t test. A P value lessthan .05 was considered to represent statisticallysignificant difference between tested data sets.

RESULTS

Evaluation of liver regeneration based on au-topsy data. After PVL, the ligated lobes (medianand left lobe) of cirrhotic rats became atrophic,whereas the unligated lobes (right and caudatelobes) and FLR became hypertrophic and ap-proached their plateau at day 7. The BrdU labelingindex was peaked at day 1 after PVL in cirrhoticrats, which was lower than noncirrhotic rats; thebasal level of the BrdU labeling index in cirrhoticrats was modestly greater than noncirrhotic rats(Fig 1, A). The restituted liver mass of cirrhotic rats

SurgeryVolume 145, Number 2

Lin et al 205

was less than noncirrhotic rats at day 7 after PVL(1.8 ± 0.2-fold vs 3.0 ± 0.7-fold, P < .01) (Fig 1, B).The redistributed volume ratio of cirrhotic ratswas accordingly less than noncirrhotic rats at day7 after PVL (63.2 ± 4.2% vs 80.5 ± 3.9%, P < .01)(Fig 1, C).

Conventional parameters of liver function afterPVL. The serum levels of albumin, ALT, and totalbilirubin in cirrhotic and noncirrhotic rats beforeand 7 days after PVL were recorded (Table I).These 3 parameters did not differ distinctly beforeand after PVL in cirrhotic rats, nor did they differ

Fig 1. (A) Comparison of BrdU assay of cirrhotic andnoncirrhotic rats after PVL. (B) Comparison of resti-tuted liver mass of cirrhotic and noncirrhotic rats afterPVL. (C) Comparison of redistributed volume ratio ofcirrhotic and noncirrhotic rats after PVL.

in noncirrhotic rats. The changes of ICG-R15 be-tween baseline and day 7 after PVL in cirrhoticand noncirrhotic rats are shown in Fig 2. TheICG-R15 of noncirrhotic rats at day 7 after PVLdid not statistically differ from that of baseline(2.0 ± 1.1% vs 2.6 ± 1.5%, P > .05), whereas theICG-R15 of cirrhotic rats at day 7 after PVL was dis-tinctly improved compared with that of baseline(6.0 ± 4.1% vs 15.8 ± 4.6%, P < .01).

Histologic evolution of FLR after PVL. Thegross morphology of FLR of cirrhotic rats at day 7after PVL remained of uneven surface as shown inFig 1, whereas the severity of necroinflammation oftheir FLR under microscopic examination wasmuch attenuated. The total Knodell score (HAI)of FLR in cirrhotic rats at day 7 after PVL was de-creased compared with that of baseline (7.6 ± 0.8vs 17.4 ± 2.3, P < .01) (Fig 3, A and B). In contrast,the sham-operated cirrhotic group exhibited a sim-ilar HAI before and at day 7 after the operation(17.6 ± 2.1 vs 17.9 ± 2.6, P = .88). Specifically, theextensive bridging necrosis of the hepatic paren-chyma was diminished at day 7 after PVL (4.2 ±0.6 vs 10.0 ± 0, P < .01). The periportal fibroticbands detected by Masson-trichrome stainingwere decreased at day 7 after PVL (0.022 ± 0.001cm2/cm2 vs 1.135 ± 0.009 cm2/cm2, P < .01)(Fig 3, C and D). The number of positive a-SMAHSCs was decreased at day 7 after PVL comparedwith that at the baseline (14 ± 3/cm2 vs 434 ±71/cm2, P < .01) (Fig 3, E and F).

In vivo liver function image study. In anotherindependent study, the experimental rats under-went Tc-99m sulfur-colloid liver SPECT and dy-namic Tc-99m DISIDA liver SPECT examinationbefore PVL and at day 1 and day 7 after PVL.

The representative features of Tc-99m sulfur-colloid liver SPECT in cirrhotic and noncirrhoticrats after PVL were shown (Fig 4); based on thesedata, the functional liver mass of FLR of experi-mental rats after PVL could be chronologically ob-served and quantified in vivo. Accordingly, theredistributed volume ratio of noncirrhotic and cir-rhotic rats after PVL was calculated (at day 7, 79 ±4% vs 64 ± 4%, P < .01) (Fig 5), which was generallycorrelated with that of the autopsy data asdescribed above. The representative features ofdynamic Tc-99m DISIDA liver SPECT in cirrhoticand noncirrhotic rats after PVL were shown(Fig 6). Based on these images, the overall DISIDAclearance rates of noncirrhotic and cirrhotic ratsbefore and after PVL was demonstrated, and thelobar DISIDA clearance rates could also be ana-lyzed separately (Fig 7). The DISIDA clearancecurve of noncirrhotic rats was shifted to the right

SurgeryFebruary 2009

206 Lin et al

Table I. Parameters of liver function after PVL in noncirrhotic and cirrhotic rats

Noncirrhosis Cirrhosis

Parameters Baseline Day 7 post PVL Baseline Day 7 post PVL

Albumin (g/dL) 2.1 ± 0.3 1.5 ± 0.3 1.8 ± 0.3 1.3 ± 0.3ALT (U/L) 45.8 ± 13.5 58.3 ± 10.8 48.8 ± 14.5 59.4 ± 14.8Bilirubin (mg/dL) 0.5 ± 0 0.5 ± 0 0.5 ± 0 0.4 ± 0.1

at day 1 and then returned to that of baseline atday 7 after PVL. In contrast, the DISIDA clearancecurve of cirrhotic rats was shifted to the right with asmall magnitude at day 1, and it dramaticallyshifted to the left with a considerable magnitudeat day 7 after PVL. Of utmost significance, the DIS-IDA clearance curve of FLR in cirrhotic rats wasshifted to the left at day 7 after PVL with a magni-tude equivalent to that of the global liver. To quan-tify the hepatocyte uptake and excretion functionof DISIDA, the time to peak and excretion T1/2

of radioactivity of cirrhotic and noncirrhotic ratsbefore and after PVL were measured (Table II).In brief, the uptake of DISIDA (time to peak) bycirrhotic rats at baseline was more delayed thannoncirrhotic rats (2.6 ± 1.0 min vs 2.0 min). Again,the clearance T1/2 of DISIDA by cirrhotic rats atbaseline was more delayed than noncirrhotic rats(9.0 ± 3.7 min vs 1.3 ± 0.8 min). Finally, the clear-ance T1/2 of FLR of cirrhotic rats at day 7 after PVLwas less compared with baseline data (5.2 ± 1.9 minvs 8.6 ± 3.1 min), although it is still greater thannoncirrhotic rats (1.3 ± 0.6 min).

DISCUSSION

In the current study, we demonstrated for thefirst time that the technique of PVL not onlyincreases the functional volume of FLR of cirrhoticrats but also improves the hepatocyte function ofFLR through tissue remodeling. Based on ourdata, the restituted liver mass (1.8 ± 0.2-fold vs

Fig 2. Changes of ICG-R15 in cirrhotic and noncirrhoticrats before and day 7 after PVL.

3.0 ± 0.7-fold, P < .01) and redistributed volume ra-tio (63.2 ± 4.2% vs 80.5 ± 3.9%, P < .01) of cirrhoticrats at the end point after PVL were less than innoncirrhotic rats. These findings are consistentwith those observed from clinical settings.6-9 Thisregeneration discrepancy among cirrhotic andnoncirrhotic subjects certainly should be kept inmind because the functional volume of FLR deter-mines the maximum extent that the targeted livercan be resected safely.16 A detailed histologic eval-uation of FLR after portal vein embolization on ahuman being is intriguing but not feasible in gen-eral practice. At most, for example, liver biopsyspecimens could be obtained from the nonembol-ized lobe during hepatic resection, and the cell ki-netics and morphometric analysis could beinvestigated,17 which did not represent the realityof entire FLR. Using a murine model, we demon-strated that the technique of PVL promotes tissueremodeling of FLR from a preexisting nodulararchitecture to a healthier microenvironment. Asshown in our results, the periportal necroinflam-mation and fibrosis bands were progressively atten-uated after PVL, which is reflected by decreasedscoring of HAI and Masson-trichrome staining. Inother words, the fibrotic status of FLR is somehowresolved partially after PVL. A similar assumptionthat an increase in liver volume correlates with in-creased function in postportal vein embolizationhas been demonstrated in a study that showed anincrease in asialoglycoprotein receptor bindingsites in the FLR before resection.18 The reportsof the reversibility of liver fibrosis in humans in-clude abstinence from alcohol, surgical reversalof jejunoileal bypass, immunosuppressive therapyfor autoimmune hepatitis, phlebotomy for hemo-chromatosis, long-term treatment with lamivudinefor chronic hepatitis B, treatment of hepatitis Cand hepatitis D with interferon, and treatment ofprimary biliary cirrhosis with ursodiol plus metho-trexate.19 Over the past 2 decades, it is clear thatthe accumulation of extracellular matrix and scar-ring in the fibrotic/cirrhotic liver is not a static orunidirectional event but a dynamic and regulatedprocess that is amenable to intervention. The acti-vation of HSCs is a central event in hepatic

SurgeryVolume 145, Number 2

Lin et al 207

Fig 3. Histologic examination of FLR of cirrhotic rats before (A) and day 7 (B) after PVL. (A and B) Hematoxylin andeosin stain, original magnification 3 200. Expression of collagen fibers of cirrhotic rats detected by Masson-trichromebefore (C) and day 7 (D) after PVL. Original magnification 3 40. Expression of a-SMA--positive HSCs in cirrhotic ratsbefore (E) and day 7 (F) after PVL. Immunohistochemical staining, original magnification 3 100.

fibrosis.20 In all forms of liver injury, these perisi-nusoidal HSCs undergo a conversion from quies-cent cells that store vitamin A to cells that arecontractile, proliferative, and fibrogenic. In animalmodels, as liver injury resolves, the activated HSCsare cleared by apoptosis, and the expression of tis-sue inhibitors of metalloprotinases decreases to al-low active enzymes to resorb the extracellularmatrix.21,22 In the current study, the number of ac-tivated HSCs in the FLR of cirrhotic rats after PVLwas largely diminished along with the time periodafter PVL. Therefore, the altered histologic archi-tecture that evolved from the preexisting nodularmicrostructure after PVL might be substantiallycaused by a decreased number of activated HSCs.Obviously, the human liver regenerative capacityafter several years of chronic-evolving cirrhoticchange is not likely to be as responsive as that ofa rat treated for 16 weeks with an acute toxin

that induces microcirrhotic change. Nevertheless,the procedure-related beneficial effect on histo-logic evolution of FLR other than volume shiftingalone is encouraging and justifies the currently in-creasing use of portal vein embolization employedin those with cirrhosis and borderline liver reservewho need an extensive hepatectomy.

Although the serum levels of albumin, ALT, andbilirubin of cirrhotic rats at day 7 after PVL did notsignificantly differ from those at baseline, it isgenerally considered that these parameters cannotreflect the hepatocyte function precisely. The clin-ical evaluation of liver function in cirrhotic sub-jects is largely based on the Child-Pugh scoring.Nevertheless, many Child-Pugh class-A cirrhoticpatients with hepatocellular carcinoma developpostoperative liver failure. Quantitative liver func-tion tests are designed to measure either theblood-flow-dependent hepatocyte function, such

SurgeryFebruary 2009

208 Lin et al

Fig 4. Tc-99m sulfur colloid liver SPECT images of noncirrhotic (A--C) and cirrhotic (D--F) rats. The volume renderingimages represent functional liver mass acquired at baseline (A, D), day 1 (B, E), and day 7 (C, F) after PVL. R, Right lobe;LM, left and median lobe; C, caudate lobe.

as ICG clearance, lidocaine clearance, and galac-tose elimination capacity, or the blood-flow-inde-pendent hepatocyte functional capacity, such asthe aminopyrine breath test.23-26 The ICG-R15 isconsidered to be one of the most valuable and re-liable tests for assessing hepatic functional reserveand predict posthepatectomy liver failure in cir-rhotic patients. In the current study, we assessedthe liver function reserve using a clinical scoringsystem ICG-R15 and compared it with SPECTthereafter. We chose the ICG-R15 as our ‘‘goldstandard’’ measure because it is very popular clini-cally and because it is convenient for 1-time bloodsampling. Based on our data, the ICG-R15 of cir-rhotic rats at day 7 after PVL was consistentlyimproved compared with that at baseline (6.0 ±4.1% vs 15.8 ± 4.6%, P < .01). The ICG-R15 test,however, represents the global liver function only,and the lobar function of interest is hardly mea-sured separately. Uesaka et al27 have demonstratedthat right portal vein embolization results in en-hancement of the functional capacity as well asthe volume of the left (nonembolized) lobe bymeasuring biliary ICG excretion and volumetry, re-spectively. The relative functional capacity of theleft lobe increased by 20% after right portal veinembolization, which was greater than that ob-served in the relative size of the left lobe (8%).

The total function and volume of the liver didnot change after portal vein embolization, whichindicates that the function and volume of the em-bolized lobe and the nonembolized lobe were inbalance. Nevertheless, the technique of measuringbiliary ICG secretion to predict lobar function of

Fig 5. Comparison of redistributed volume ratio amongnoncirrhotic and cirrhotic rats after PVL detected by Tc-99m sulfur-colloid SPECT.

SurgeryVolume 145, Number 2

Lin et al 209

Fig 6. Coronal view of dynamic Tc-99m DISIDA liver SPECT images in noncirrhotic (A) and cirrhotic rats (B) acquiredat 2, 4, 6, and 8 min after tracer injection. The radioactivity intensity was normalized to its maximum for comparison.

Fig 7. Whole-liver DISIDA clearance in (A) noncirrhotic and (B) cirrhotic rats. FLR DISIDA clearance in (C) noncir-rhotic and (D) cirrhotic rats. FLR is indicated in the right and caudate lobes.

SurgeryFebruary 2009

210 Lin et al

Table II. Tc99m DISIDA SPECT examination in cirrhotic and noncirrhotic rats after PVL

Baseline Day 1 post PVL Day 7 post PVL

Control subjects (n = 6)Time to peak (min) 2 2 2RCL clearance T1/2 (min) 1.33 ± 0.77 1.90 ± 0.50 1.30 ± 0.63LML clearance T1/2 (min) 1.42 ± 0.83 2.06 ± 0.54 1.10 ± 0.32WL clearance T1/2 (min) 1.38 ± 0.80 1.98 ± 0.52 1.19 ± 0.46

Cirrhotic subjects (n = 6)Time to peak (min) 2.67 ± 1.03 3.67 ± 1.51 3.00 ± 1.15RCL clearance T1/2 (min) 8.68 ± 3.19 8.78 ± 2.98 5.29 ± 1.93LML clearance T1/2 (min) 9.15 ± 4.50 11.51 ± 5.38 6.60 ± 2.73WL clearance T1/2 (min) 9.01 ± 3.72 10.07 ± 3.79 6.16 ± 2.68

RCL, Right and caudate lobes; LML, left and median lobes; WL, whole liver.

the FLR is not feasible for those with hepatocellu-lar carcinoma and cirrhosis who are not equippedwith percutaneous transhepatic biliary drainage.

To investigate the hepatic function of FLRindividually after PVL, we conducted SPECT ex-amination using 2 different tracers before andafter PVL in cirrhotic subjects. The Tc-99m sulfur-colloid liver SPECT enabled us to detect thedynamic change of functional liver mass of cir-rhotic rats after PVL.

According to our data, the redistributed volumeratio of noncirrhotic and cirrhotic rats after PVLwere 79 ± 4% and 64 ± 4%, respectively, which wasgenerally correlated with that of the autopsy data.The Tc-99m sulfur-colloid liver SPECT is believed tobe superior to conventional CT volumetry becausethe former identifies only the viable and function-ing mass as scintigraphically positive, whereas thelatter would cover all liver tissues regardless of theirviability.23-26 This discrimination is clinically impor-tant because an infracted liver does not function.On the other hand, dynamic Tc-99m DISIDASPECT confers an excellent chance to delineatethe hepatocyte uptake and secretion function ofindividual lobe. When the liver is diseased, the ex-cretion function of hepatocytes for DISIDA is im-paired and delayed; thus, the DISIDA clearancecurve of the diseased animals will be accordinglyshifted to right. On the contrary, if the diseasedliver is improved functionally, then the DISIDAclearance curve will be shifted to left.

As shown in our data, the DISIDA clearancecurve of the cirrhotic rats was shifted to right with asmall magnitude at day 1, and then it was dramat-ically shifted to left at day 7 after PVL. Of moresignificance, the DISIDA clearance curve of FLR inthe cirrhotic rats was shifted to left at day 7 afterPVL with a magnitude equivalent to that of theglobal liver. This phenomenon can be interpretedas the improved hepatocyte secretion function of

cirrhotic rats after PVL detected by DISIDA SPECTis nearly almost contributed by FLR componentalone. Taken together, we speculate that the im-proved ICG-R15 in cirrhotic rats after PVL waspredominately attributed to the improved hepato-cyte function of FLR.

In conclusion, the regenerated functional livermass of cirrhotic rats after PVL is less than non-cirrhotic rats; whereas the hepatocyte function ofFLR in cirrhotic rats is improved relevant to tissueremodeling, which is an additional advantageother than volume shifting.

REFERENCES

1. De Baere T, Roche A, Elias D, Lasser P, Lagrange C, Bous-son V. Preoperative portal vein embolization for extensionof hepatectomy indications. Hepatology 1996;24:1386-91.

2. Hemming AW, Reed AI, Howard RJ, Fujita S, Hochwald SN,Caridi JG, et al. Preoperative portal vein embolization forextended hepatectomy. Ann Surg 2003;237:686-93.

3. Imamura H, Shimada R, Kubota M, Matsuyama Y, Na-kayama A, Miyagawa S, et al. Preoperative portal vein embo-lization: an audit of 84 patients. Hepatology 1999;29:1099-105.

4. Nagino M, Kamiya J, Nishio H, Ebata T, Arai T, Nimura Y.Two hundred forty consecutive portal vein embolizationsbefore extended hepatectomy for biliary cancer. Surgicaloutcome and long-term follow-up. Ann Surg 2006;243:364-72.

5. Farges O, Belghiti J, Kianmanesh R, Regimbeau JM, San-toro R, Vilgrain V, et al. Portal vein embolization beforeright hepatectomy. Prospective clinical trial. Ann Surg2003;237:208-17.

6. Chik BH, Liu CL, Fan ST, Lo CM, Poon RT, Lam CM, et al.Tumor size and operative risks of extended right-sided he-patic resection for hepatocellular carcinoma. Arch Surg2007;142:63-9.

7. Wakabayashi H, Ishimura K, Okano K, Karasawa Y, Goda F,Maeba T, et al. Application of preoperative portal vein em-bolization before major hepatic resection in patients withnormal or abnormal liver parenchyma. Surgery 2002;131:26-33.

8. Azoulay DA, Castaing D, Krissat J, Smail A, Hargreaves GM,Lemoine A, et al. Percutaneous portal vein embolization in-creases the feasibility and safety of major liver resection for

SurgeryVolume 145, Number 2

Lin et al 211

hepatocellular carcinoma in injuried liver. Ann Surg 2000;232:665-72.

9. Giraudo G, Greget M, Oussoultzoglou E, Rosso E, Bachel-lier P, Jaeck D. Preoperative contralateral portal vein embo-lization before major hepatic resection is a safe and efficientprocedure: a large single institution experience. Surgery2008;143:476-82.

10. Nagino M, Nimura Y, Kamiya J, Kondo S, Uesaka K, Kin Y,et al. Changes in hepatic lobe volume in biliary tract cancerpatients after right portal vein embolization. Hepatology1995;21:434-9.

11. Yeh TS, Ho YP, Huang SF, Yeh JN, Jan YY, Chen MF. Thalid-omide salvages lethal hepatic necroinflammation and accel-erates recovery from cirrhosis in rats. J Hepatol 2004;41:606-12.

12. Nagano Y, Nagahori K, Kamiyama M, Fujii Y, Kubota T,Endo I, et al. Improved functional reserve of hypertrophiedcontralateral liver after portal vein ligations in rats. J Hepa-tol 2002;37:72-7.

13. Liu EH, Chen MF, Yeh TS, Ho YP, Wu RC, Chen TC, et al. Auseful model to audit liver resolution from cirrhosis in ratsusing functional proteomics. J Surg Res 2007;138:214-23.

14. Knodell RG, Ishak KG, Black WC, Chen TS, Craig R, Kaplo-witz N, et al. Formulation and application of a numericalscoring system for assessing histological activity in asymp-tomatic chronic active hepatitis. Hepatology 1981;1:431-5.

15. Desmet VJ, Gerber M, Hoofnagle JH, Manns M, Scheuer PJ.Classification of chronic hepatitis: diagnosis, grading andstaging. Hepatology 1994;19:1513-9.

16. Kubota K, Makuuchi M, Kusaka K, Kobayashi T, Miki K,Hasegawa K, et al. Measurement of liver volume and hepa-tic functional reserve as a guide to decision-making inresectonal sugery for hepatic tumors. Hepatology 1997;26:1176-81.

17. Harada H, Imamura H, Miyagawa S, Kawasaki S. Fate ofhuman liver after hemihepatic portal vein embolization:

cell kinetic and morphometric study. Hepatology 1997;26:1162-70.

18. Kudo M, Todo A, Ikekubo K, Yamamoto K, Vera DR, Stadal-nik RC. Quantitative assessment of hepatocellular functionthrough in vivo radioreceptor imaging with technetium99m galactosyl human serum albumin. Hepatology 1993;17:814-9.

19. Desmet VJ. Comments on cirrhosis reversal. Dig Liver Dis2005;37:909-16.

20. Bonis PAL, Friedman SL, Kaplan MM. Is liver fibrosis revers-ible? N Engl J Med 2001;344:452-4.

21. Iradale JP, Benyon RC, Pickering J, McCullen M, Northrop M,Pawley S, et al. Mechanism of spontaneous resolution of ratliver fibrosis. Hepatic stellate cell apoptosis and reducedhepatic expression of metalloproteinase inhibitors. J ClinInvest 1998;102:538-49.

22. Friedman SL. Liver fibrosis-from bench to bedside. J Hepa-tol 2003;38:S38-53.

23. Carrasquillo JA, Rogers JV, Williams DL, Shuman WP, OlsonDO, Larson SM. Single-photon emission computed tomog-raphy of the normal liver. AJR Am J Roentgenol 1983;141:937-41.

24. Zuckerman E, Slobodin G, Sabo E, Yeshurun D, Naschitz JE,Groshar D. Quantitative liver-spleen scan using single pho-ton emission computerized tomography (SPECT) for assess-ment of hepatic function in cirrhotic patients. J Hepatol2003;39:326-32.

25. Oppenheim BE, Krepshaw JD. Dynamic hepatobiliarySPECT: a method for tomography of a changing radioactiv-ity distribution. J Nucl Med 1988;29:98-102.

26. Hoefs JC, Chen PT, Lizotte P. Noninvasive evaluation ofliver disease severity. Clin Liver Dis 2006;10:535-62.

27. Uesaka K, Nimura Y, Nagino M. Changes in hepatic lobarfunction after right portal vein embolization. An appraisalby biliary indocyanine green excretion. Ann Surg 1996;223:77-83.