www.elsevier.com/locate/jhep

Journal of Hepatology 46 (2007) 474–481

Chemoembolization of hepatocellular carcinoma with drugeluting beads: Efficacy and doxorubicin pharmacokineticsq

Marıa Varela1, Marıa Isabel Real2, Marta Burrel2, Alejandro Forner1, Margarita Sala1,Merce Brunet3, Carmen Ayuso2, Lluis Castells4, Xavier Montana2, Josep M. Llovet1,5,

Jordi Bruix1,*

1Liver Unit, IMDM, Barcelona Clınic Liver Cancer (BCLC) Group, Hospital Clınic, CIBER HEPAD, University of Barcelona,

Institut d’Investigacions Biomediques August Pi I Sunyer, IDIBAPS, Barcelona, Spain2Department of Radiology, CDI, Barcelona Clınic Liver Cancer (BCLC) Group, Hospital Clınic, CIBER HEPAD, University of Barcelona,

Institut d’Investigacions Biomediques August Pi I Sunyer, IDIBAPS, Barcelona, Spain3Pharmacology Laboratory, CDB, Barcelona Clınic Liver Cancer (BCLC) Group, Hospital Clınic, CIBER HEPAD, University of Barcelona,

Institut d’Investigacions Biomediques August Pi I Sunyer, IDIBAPS, Barcelona, Spain4Liver Unit, Hospital Valld’Hebron, Barcelona, Catalonia, Spain

5Institut Catala de Recerca Avancada (ICREA), Spain

See Editorial, pages 362–364

Background/Aims: This study assesses the safety, pharmacokinetics and efficacy of transarterial chemoembolization

using drug eluting beads (DEB), an embolizing device that slowly releases chemotherapy to decrease systemic toxicity.

Methods: Twenty-seven Child-Pugh A cirrhotics (76% male, 59% HCV) with untreated large/multifocal HCC receivedchemoembolization with doxorubicin loaded DEBs at doses adjusted for bilirubin and body surface (range: 47–150 mg).

Clinical and analytical data were recorded at 24 and 48 h, 7, 14 and 30 days after first and second TACE. Response rate

was assessed by CT at 6 months. Blood samples were obtained in 13 patients at 5, 20, 40, 60, 120 min, 6, 24, 48 and 168 h

to determine doxorubicin Cmax and AUC.

Results: DEB-TACE was well tolerated with an acceptable safety profile. Two cases developed liver abscess, one leading to

death. Response rate was 75% (66.6% on intention-to-treat). Doxorubicin Cmax and AUC were significantly lower in DEB-

TACE patients (78.97 ± 38.3 ng/mL and 662.6 ± 417.6 ng/mL min) than in conventional TACE (2341.5 ± 3951.9 ng/mL

and 1812.2 ± 1093.7 ng/mL min, p = 0.00002 and p = 0.001, respectively). After a median follow-up of 27.6 months, 1-and 2-year survival is 92.5% and 88.9%, respectively.

Conclusions: Chemoembolization using DEBs is an effective procedure with a favorable pharmacokinetic profile.

� 2007 Published by Elsevier B.V. on behalf of the European Association for the Study of the Liver.

Keywords: Hepatocellular carcinoma; TACE, Doxorubicin; Drug eluting beads; Pharmacokinetics; RECIST; Response rate

0168-8278/$32.00 � 2007 Published by Elsevier B.V. on behalf of the European Association for the Study of the Liver.

doi:10.1016/j.jhep.2006.10.020

Received 3 September 2006; received in revised form 11 October 2006; accepted 20 October 2006; available online 29 November 2006q The authors of this paper declare they have a relationship with the manufacturers of the device involved. The authors received funding from

BioCompatibles, UK which enabled them to carry out this study.* Corresponding author. Tel.: +34 93 227 98 03; fax: +34 93 227 57 92.

E-mail address: bruix@ub.edu (J. Bruix).Abbreviations: HCC, hepatocellular carcinoma; TACE, transarterial chemoembolization; DEB, drug eluting beads; CT, computed tomography;

Cmax, peak drug concentration; AUC, area under the curve; HPLC, high performance liquid chromatography; RCT, randomised controlled trial.

M. Varela et al. / Journal of Hepatology 46 (2007) 474–481 475

1. Introduction

Hepatocellular carcinoma (HCC) is increasing world-wide [1,2]. Resection, liver transplantation and ablationare the only curative treatments for HCC, but even withearly detection plans most patients are diagnosed atintermediate-advanced stage, when the sole option thatimproves survival is transarterial chemoembolization(TACE) [3]. This achieves its effect by inducing tumorischemia (usually through gelfoam injection) associatedto prior cytotoxic action of selectively injected chemother-apy aiming to prompt long-lasting intratumoral retentionof the drug. TACE elicits a high rate of objective respons-es and therapeutic success results in delayed tumor pro-gression and increased survival [4,5]. Since improvedsurvival depends on treatment response, the increase ofthe efficacy of TACE should come from an enhancementof the response together with an expansion of its duration.

Ideally, the injected chemotherapeutic should beretained in the tumor and be gradually released to avoidsystemic toxicity. However, even if suspended in lipiod-ol, its selective injection is associated to significant pas-sage into the systemic circulation.

The development of drug eluting beads (DEBs) load-ed with chemotherapy that is slowly released upon injec-tion [6] may increase the intensity and duration ofischemia while enhancing the drug delivery to the tumor.This phase 2 study was designed to establish the efficacyand safety of DEB for the TACE treatment of HCCpatients. In addition to this primary end-point, we inves-tigated in a subset of patients if DEBs modify the phar-macokinetic profile of doxorubicin.

2. Patients and methods

2.1. Patients

The study included 27 consecutive patients recruited from Febru-ary 2004 to May 2005. The protocol was approved by the Ethics Com-mittee of the Institution. Tolerance of escalating doses of doxorubicinwas assessed in pairs of patients at separate dose levels and substra-tified according to bilirubin concentration. The pharmacokinetic pro-

Table 1

Baseline characteristics of the patients enrolled in the study

Characteristics DEB-TACE (n = 27

Mean age (years) 67.9 ± 7.8Sex (male/female) 21/6Etiology of HCC (VHC/alcohol/others) 16/7/4Child-Pugh (A/B) 27/0Okuda I/II 26/1BCLC stage A/B/C 0/27/0Bilirubin (mg/dL) 1.02 ± 0.5Albumin (g/L) 40.8 ± 4.0AFP (ng/mL) (<10/10–100/>100) 0/20/7Mean tumor size (range) mm 46 (8–150)Number of nodules (1/1 + satellites/2/multifocal) 9/2/7/9

file of DEBs was done in a subset of 13 patients (with different dosesand liver function) aiming to validate the reduced passage of chemo-therapy into the systemic circulation. HCC was managed accordingto Barcelona-Clinic-Liver-Cancer (BCLC) schedule [7–9]. Patients withHCC 65 cm or 3 nodules measuring 63 cm are considered for resec-tion, ablation or transplantation. Multinodular tumors are consideredfor chemoembolization. Therefore, this study included asymptomaticpatients with untreated intermediate HCC in compensated Child-PughA cirrhosis without vascular invasion or extrahepatic spread. Exclu-sion criteria were impaired clotting tests (platelet count below50.000/mm3 or prothrombin activity below 50%), renal failure andany contraindication to doxorubicin (bilirubin levels >3 mg/dL, whiteblood cell count <3.000/mm3, cardiac ejection fraction <50%), hepato-fugal blood flow or portosystemic shunt.

Table 1 shows the characteristics of the patients. Most were male,mean age was 68 years and 59% presented hepatitis C virus. Diagnosiswas confirmed by biopsy in 13 cases (48%), by two coincidental imag-ing techniques in 11 (41%) and by one imaging technique associatedwith increased a-fetoprotein in 3 (11%). The mean diameter of the larg-est nodule was 46 mm.

2.2. DEB-TACE procedure

DEB-TACE patients underwent two TACEs separated by 2 months[10]. Response was assessed at 6 months through computed tomography(CT) according to RECIST [11] and EASL criteria [12]. All patientsunderwent baseline angiography of the celiac trunk, superior mesentericartery and hepatic artery using a peripheral arterial approach. No anti-biotic prophylaxis was given. Doxorubicin doses were adjusted for bodysurface and bilirubin. Pharmacokinetic analysis was done by a conven-tional dose escalation strategy [13], stratifying patients according to bil-irubin <1.5 mg/dL (n = 22) and between 1.5 and 3 mg/dL (n = 5). Initialdose was 25 mg/m2. If no-dose-limiting toxicity was observed, the dosewas risen to 50 mg/m2, 75 mg/m2 and 100 mg/m2 with a maximum totaldose of 150 mg per session of treatment. Four patients received 25 mg/m2 (bilirubin <1.5 mg/dL (n = 2) and bilirubin >1.5–3 mg/dL (n = 2)),four received 50 mg/m2 (bilirubin <1.5 mg/dL (n = 2) and bilirubin>1.5–3 mg/dL (n = 2)), three patients received 75 mg/m2 (bilirubin<1.5 mg/dL (n = 2) and bilirubin >1.5–3 mg/dL (n = 1)), and 16, all ofthem with normal bilirubin, received 100 mg/m2.

Highly selective catheterization was performed with a 3F micro-catheter in order to obtain complete obstruction of the nourishingarteries and avoid damage to non-tumoral liver. DEBs (BioCompati-bles Ltd., UK) [6] with a diameter ranging between 500 and 700 lmwere loaded with doxorubicin (Pharmacia-UpJohn, Barcelona, Spain)and mixed with an equal volume of contrast media. Additional unload-ed spheres were used to complete the embolization procedure.

Patients were discharged after recovery followed by physical exam-ination and blood test at 7, 14, 30 days and 1 week prior to the secondtreatment that was performed two months later [4]. Patency of thetumor arterial vessels was checked prior to the second TACE. Treat-ment efficacy was assessed by dynamic CT scan at 6 months. Responsewas evaluated by RECIST criteria [11] and following the EASL [12]amendments that take into account the amount of necrotic tumor.

) Conventional TACE (n = 5) Lipiodol–doxorubicin(n = 2)

67.7 ± 7.5 58.0 ± 2.83/2 2/05/0/0 1/15/0 2/05/0 1/10/5/0 0/0/21.28 ± 0.5 1.05 ± 0.0738.3 ± 4.7 42.5 ± 4.91/3/1 1/1/036 (15–50) 76 (43–110)1/0/2/2 0/0/0/2

476 M. Varela et al. / Journal of Hepatology 46 (2007) 474–481

2.3. Doxorubicin pharmacokinetics

This was performed in 13 patients. Peripheral blood samples weredrawn at baseline and at 5, 20, 40, 60, 120 min, 6, 24, 48 h and 7 daysafter doxorubicin administration to determine Cmax and area under thecurve (AUC). Plasma was separated immediately and stored at� 20 �C. Doxorubicin levels were measured by high performanceliquid chromatography (HPLC) with fluorescence detector [14].

HPLC was performed using an Alliance System equipped with anEmpower Network Chromatography Manager (Waters) for data cap-ture. The mobile phase is a mixture of H2O/acetonitrile/tetrahydrofu-ran (76/24/0.5 v/v/v) with the pH adjusted to 2.0 with perchloric acidand pumped at a flow rate of 1.25 mL/min. The column system includ-ed an analytical column Inertsil ODS-80A 5 lm (4.6 · 150 mm) heatedin a column oven to 50 �C. Run time between injections is 45 min. Theeluent is monitored fluorimetrically at an excitation wavelength of480 nm and an emission wavelength of 560 nm, with a band width of40 nm. The doxorubicin is eluted at 10.6 min and internal standardis eluted at 29.0 min. Calibration standards are prepared in blankhuman plasma at concentrations of 1.0, 2.5, 5.0, 10.0, 25.0, 50.0, and100 ng/mL. Sets of quality control (QC) samples are prepared at 2.0,40.0 and 70.0 ng/mL. The lower limit of quantitation is 1 ng/mL. Con-centrations versus time curves were constructed over the 7 days.

Comparison of the pharmacokinetics in DEB-TACE patients wasdone versus contemporary HCC patients with comparable hematologi-cal and biochemical profiles (Table 1). They were treated withconventional TACE (gelfoam and lipiodol) (n = 5) or with lipiodol–doxorubicin without arterial obstruction (n = 2). The procedure wasperformed through a 4–5 F catheter through the lobar, segmental or sub-segmental arteries. Doxorubicin was injected in lipiodol (LaboratoireGuerbert, Aulnay-sous-Bois, France) at a dose adjusted to bilirubin con-centrations (6 patients with bilirubin <1.5 mg/dL received 75 mg ofdoxorubicin and 1 patient with bilirubin >1.5–3 mg/dL received50 mg), while obstruction was done by gelfoam fragments injection.

2.4. Toxicity assessment

Toxicity was assessed on days 7, 14 and 30 in DEB-TACE, and ondays 7 and 30 in groups treated with conventional TACE or lipiodol–doxorubicin. Its occurrence was recorded according to World HealthOrganization (http://www.fda.gov/cder/cancer/toxicityframe.htm).

Basal angiography prior to first DEB –TACEn = 27

First DEB-TACEn = 27

3 drop o

Basal angiography prior to second DEB -TACEn = 24

1 drop o1 persisno colla

Second DEB-TACEn = 22

Assessment of response was made in 24 patients (2 odissection). Analysis of tolerance and complication incl

Fig. 1. Flowchart of patients

2.5. Statistical analysis

Comparisons among groups were made by ANOVA and U-Mann–Whitney for continuous variables and the v2 test for categorical vari-ables. AUC was calculated through the trapezoidal rule. Calculationswere made with the SPSS package (version 11.0).

3. Results

The main study includes 27 patients in whom theend-point was the assessment of safety and efficacy ofDEB-TACE. Thirteen of them were included in thepharmacokinetics analysis.

3.1. Applicability, safety and efficacy of DEB-TACE

First session of treatment was effectively performed inall patients: Arterial blood flow to all tumor sites wasfully abolished in 24 of 27 cases (89%), while obstructionwas incomplete in 3. In two of these the second DEB-TACE obtained complete arterial obstruction, whereasin the third patient obstruction was unfinished becauseHCC was fed by multiple small/narrow arteries.

Three patients were excluded before the second DEB-TACE (see flowchart, Fig. 1), one due to portal veininvasion, and 2 because of liver abscess. All threepatients had reached a total obstruction of the feedingarteries with the first DEB-TACE.

Angiography prior to the second treatment assessedthe persistence of initial obstruction or if collateralshad developed. In 4 of the 21 patients that had achieveda total occlusion during the first DEB-TACE, this was

ut:liver abscess (2)

ut: hepatic artery dissection tence of complete embolization, terals

ut due to liver abscess, 1 out due to hepatic artery uded 49 DEB-TACE sessions.

progression (1)

of DEB-TACE group.

M. Varela et al. / Journal of Hepatology 46 (2007) 474–481 477

still present. In one of these 4 patients the second treat-ment was not attempted because of persistent completearterial obstruction, while in the other three the secondDEB-TACE was performed through newly developedcollaterals. In 16 patients the initial occlusion had beenpartially permeated and the second procedure obstruct-ed again the same artery. Only one patient had fullyrecovered a patent arterial flow and complete obstruc-tion was obtained again.

Finally, one of the three patients in whom the firstDEB-TACE obtained just partial obstruction, thetumoral arteries presented a dissection of the left hepaticartery during the second angiography, and the secondDEB-TACE was cancelled. The other two patientsunderwent the second DEB-TACE to complete the arte-rial obstruction. Therefore, the analysis of tolerance andcomplications includes 49 treatment sessions.

The mean dose of doxorubicin was 128 mg (range 47–150). In one case the dose was 25% of the scheduled asblood flow was obstructed with a small amount ofDEBs. In 11 procedures (22.4%) additional unloadedbeads were used to complete the embolization. Meanprocedural time was 63 ± 19 min (range 15–103) andthe duration of hospitalization was 2.8 ± 1.4 days (range1–9).

After the first procedure, only 10 of the 27 patients(37%) presented a clinically relevant post-TACE syn-

Measurement of liver function testafter the first DEB-TACE.A.-Bilirubin

B.-SGPT

C.-γGT

D.-Alkaline Phosphatase

E.-Prothrombin activity Bilir

ubin

(mg/

dL)

5

4

3

2

1

0

A

1 month post

14 days post

7 days post

48 h post

24 h post

Baseline

48 h p24 h post

Baseline

48 h p24 h post

Baseline

GG

T (U

/L)

400

300

200

100

0

C

Alka

line

Phos

phat

e (U

/L)

600

400

200

0

D

Fig. 2. Measurement of liver function

drome. This did not require prolonged hospitalization.Three patients (11%) required acetaminophen or trama-dol due to transient abdominal pain immediately afterDEB-TACE. Six patients (22%) presented mild fever(<38 �C) and 4 (15%) nausea and vomiting. Twopatients were diagnosed with liver abscess. All but thesetwo patients returned to pretreatment status at 14 daysof follow-up. After the second procedure (n = 22), 4patients (18%) presented post-TACE syndrome, 7(32%) mild pain, and 3 (14%) nausea and vomiting.All minor secondary effects disappeared within the firstweek. None of the patients presented alopecia, bonemarrow toxicity, dyspnea or pulmonary embolism.

Liver function parameters were slightly altered imme-diately after the procedure without relevant deteriora-tion, except in the two patients with liver abscess.Fig. 2 shows the evolution of bilirubin, SGPT, cGT,alkaline phosphatase and prothrombin rate after thefirst DEB-TACE.

The first patient that developed the liver abscess washealed with antibiotics and physical status was com-pletely recovered 1 month later. He is asymptomatic at25 months follow-up. The other patient developed pro-gressive liver decompensation and died at three months,thus resulting in a 3.7% treatment-related death rate.

Response to therapy including all 27 patients isshown in Table 2. Assessment of response was not done

SGPT

(U/L

)

400

300

200

100

0

B

1 month post

14 days post

7 days postost

1 month post

14 days

7 daysost

1 month post

14 days post

7 days post

48 h post

24 h post

Baseline

1 month post

14 days post

7 days post

48 h post

24 h post

Baseline

Prot

hrom

bin

activ

ity (%

)

120

100

80

60

40

E

tests after the first DEB-TACE.

Table 2

Efficacy of the treatment according to intention-to-treat (n = 27)

RECIST-modified WHO-modified

CR 0 (0%) 7 (25.9%)PR 12 (44.4%) 11 (40.7%)SD 7 (25.9%) 1 (3.7%)P 5 (18.5%) 5 (18.5%)NA 3 (11.1%) 3 (11.1%)Total 27 27

Assessment of response rate at 6 month after the procedure, accordingto RECIST criteria taking into account the necrosis of the tumor [11]and WHO criteria modified by EASL [12].CR, complete response; PR, partial response, SD, stable disease; P,progressive disease; NA, not available; OR, objective response(CR + PR) of 66.6% according to EASL criteria according to anintention-to-treat perspective.

Group of TACE

Conventional (n = 5)DEB-TACE (n=13)

Cm

ax (n

g/m

L)

2000

1500

1000

500

0

p = 0.001

Fig. 3. Variability of Cmax in patients of DEB-TACE and conventional

TACE groups.

Table 3

Pharmacokinetic assay: doses administered, Cmax and AUC

Dose (mg) Cmax (ng/mL) AUC (ng/mL min)

DEB-TACE (n = 13)

47 57.3 396.5453 66.1 485.95

106 68.8 482.30135 128.1 1153.00110 75.1 674.19

93 31.8 251.14135 73.1 288.11150 74.0 710.61150 136.0 916.80150 27.5 250.80

56 52.5 625.7093 83.1 634.35

105 153.3 1744.30

Conventional TACE (n = 5)

75 1943.8 1788.2075 692.2 857.4050 400.0 1128.9075 1089.7 1827.1075 371.2 1567.70

Lipiodol–doxorubicin (n = 2)

75 11219.0 4163.3075 673.6 1353.00

Abbreviations: DEB-TACE, TACE performed with drug elutingbeads; Cmax, maximal doxorubicin concentration at serum afteradministration; AUC, area under the curve.

478 M. Varela et al. / Journal of Hepatology 46 (2007) 474–481

in three patients (2 due to liver abscess and 1 due to arte-rial dissection). Based on RECIST criteria [11] therewere zero complete responses, and 12 partial responses(>30% reduction was registered in 12 patients, 44%).In the remaining cases, the tumor may have experiencedextensive necrosis, but in the absence of reduction indiameter, the patients were classified as having eitherstable disease (n = 7) or disease progression (n = 5).This explains the discrepancy with the evaluation oftumor response if considering extent of necrosis. Fol-lowing the EASL [12] and AASLD [8] guidelines the rateof objective responses increased to 66.6%. Seven patients(26%) presented a complete response (disappearance ofthe enhancement of all the measurable lesions) and 11patients (41%) a partial response (>50% decrease ofenhanced lesions from baseline).

One patient died due to abscess formation and threemore patients died at 11, 18 and 25 months due to dis-ease progression. The 1- and 2-year survival is 92.5%and 88.9%, respectively, with a mean follow-up of27.6 months (range 3.3–31.3).

3.2. Pharmacokinetic profile

The pharmacokinetic profile was significantly differ-ent between the DEB-TACE and the conventionalTACE groups. The peak drug concentration (Cmax) isreached within 5 min after injection in all cases, but itis higher in patients treated by conventional TACE(895.66 ± 653.1 ng/mL) than in patients in the DEB-TACE group (78.97 ± 38.3 ng/mL) (p = 0.001). Fig. 3shows that the degree of variability of Cmax in patientsof conventional TACE group was more pronouncedthan in the DEB-TACE group, with no relationship tothe administered dose. The AUC further exposes the dif-ferences between groups. Table 3 depicts the doxorubi-cin doses administered and the Cmax and AUC of eachcase. Figs. 4 and 5 evidence that the AUC of groupDEB-TACE is significantly lower (662.6 ± 417.6 ng/mL min) than that observed in the conventional TACE,

in which AUC is 1532.98 ± 295.2 ng/mL min (p =0.005), although dose of doxorubicin was significantlyhigher in the DEB-TACE group (mean doses106.4 ± 37.2 mg vs 70.0 ± 11.2 mg, p = 0.006). The peak

DEB-TACE1000

800

600

400

200

0

conventional TACE

Time postprocedure

7 d48 h24 h

6 h2 h60 min

40 min

20 min

5 minbaseline

(ng/

mL)

Dox

orub

icin

at s

erum

(ng/

mL)

1000

800

A

BGroup of TACE

Conventional (n=5)DEB-TACE (n=13)

AUC

(ng/

mLx

min

)

2000

1500

1000

500

0

p= 0.005

Fig. 4. AUC of doxorubicin levels measured during 7 days in the DEB-

TACE and conventional TACE groups.

M. Varela et al. / Journal of Hepatology 46 (2007) 474–481 479

concentration and AUC in lipiodol–doxorubicinpatients are described in Table 3. Both values are higherthan those of conventional TACE but no statisticalcomparison is valid.

Time postprocedure

7 d48 h24 h

6 h2 h60 min

40 min

20 min

5 minbaseline

Dox

orub

icin

at s

erum 600

400

200

0

Fig. 5. Measurements of serum doxorubicin levels at different time

points and AUC in DEB-TACE patients above (A), and in the

conventional TACE group below (B).

4. Discussion

TACE is an effective option for patients with interme-diate HCC [8,15]. Improved survival is due to theachievement of an objective treatment response reflectedby extensive tumor necrosis. The strategy to improve theimpact of TACE on survival should be based on thedevelopment of new approaches that, while increasingthe antitumoral effect with longer duration of theresponse to treatment, would have a better tolerancewith a low rate of side effects.

The present data show that DEB-TACE is a promis-ing tool. It achieves major tumor necrosis, while the sideeffects of chemotherapy are reduced due to its reducedpassage into the systemic circulation. Some investigatorshave suggested that lipiodol retains chemotherapy insidethe tumor while its passage into the venous vesselsincreases tumor ischemia [16], but robust proof of thisconcept is not available. Kalayci et al. showed that thepeak concentration and the AUC were not differentbetween systemic chemotherapy and selective lipiodoli-zation [17]. Thus, conventional TACE is hampered byside effects of chemotherapy. Hepatic artery obstructioncould decrease the washout of chemotherapy by theblood stream, but the time elapsed between chemother-apy injection and gelfoam placement allows the clear-ance of a large proportion of chemotherapy. The use

of DEB-TACE sharply decreases the passage of drugto the systemic circulation even when injecting very highdoses. We injected up to 150 mg of doxorubicin, exceed-ing the common schedule and even so, the peak concen-tration and the AUC were significantly lower than thoseof conventional TACE. The better profile of DEB-TACE vs conventional TACE opens the opportunityto increase the amount of drug selectively exposed totumor cells and simultaneously, reduce toxicity. Sinceside effects may become negligible, treatment could befrequently repeated using even higher doses and haveincreased efficacy. Tris-acryl gelatine microspheres [18–

480 M. Varela et al. / Journal of Hepatology 46 (2007) 474–481

20] and alginate gel beads containing doxorubicin [21]have been investigated but no data on hepatic arterialembolization have been published. Interestingly, Cmax

is even lower than that observed in non-cirrhoticpatients treated with intravenous doxorubicin and thisis a relevant advantage to diminish early toxicity. Bycontrast, the impairment of hepatic function may favora lower extraction rate and a longer half-life in periphe-ral blood with potentially higher AUC that will main-tain the toxicity due to the total cumulative dose [22].

Two patients developed hepatic abscess after tumornecrosis. This is a complication that may develop withthe use of any embolizing device and has been reportedafter conventional TACE with gelfoam, coils, polyvinylalcohol or blood clots [23]. Some groups routinely employantibiotic prophylaxis but there is no evidence of the ben-efit [24,25]. Although the incidence of liver abscess follow-ing a TACE is low (0.26–3.12%) [26–28], the rate ofmortality is as high as 20–100%. The most important riskfactor for its development is the presence of earlier bilio-enteric anastomosis [29]. None of our patients presentedthis condition and future experience with DEB-TACEshould establish if the risk of hepatic abscess is increased.

The rate of objective responses is encouraging withthe use of DEB-TACE. Since the RECIST criteria [11]do not consider the induction of necrosis and just mea-sure tumor diameter, any study using this system as rec-ommended by the National Cancer Institute willunderestimate the efficacy of any locoregional approach.This is clearly shown in our study in which RECIST cri-teria did not register any complete response, while theinduction of massive tumor necrosis with no intratumor-al contrast uptake on dynamic CT identified seven ofsuch optimal results, and in other 11 patients the extentof tumor necrosis exceeded 50% of the baseline tumorburden. The achievement of an objective response in66.6% of the cases exceeds the response rate reportedin most of the carefully performed investigations. Inthem, the objective response rate ranges between 27%and 55% one month after the procedure [5] and 35%at 6 months [4]. Accordingly, our data indicate thatDEB-TACE appears to be able to significantly exceedthe antitumoral efficacy of conventional TACE.Whether the increase in objective response rates impactssurvival should be assessed in controlled investigations.There are no data showing that the magnitude of theobjective response after TACE results in improved sur-vival. In addition, it could be argued that the long-last-ing arterial obstruction induced by DEB-TACEprevents dynamic CT scans to detect viable tumor, whilethis would not occur if using gelfoam obstruction.According to this possibility, the comparison ofresponse rate at identical time points may be highly mis-leading and the superiority of DEB-TACE should bebetter based on the comparison of time to progression,to symptomatic progression or to death.

In summary, DEB-TACE is a highly effective treat-ment for patients with intermediate HCC. Its safetyand efficacy profile and the capacity to diminish theamount of chemotherapy that reaches the systemiccirculation warrant the development of randomized con-trolled studies to unequivocally establish its superiorityversus conventional gelfoam based TACE.

Acknowledgements

Marıa Varela is supported by Fundacion Cientıficade la Asociacion Espanola de Ayuda contra el Cancer.Spain.

Alejandro Forner is supported by a grant from Insti-tuto Sanitario Carlos III (PI 05/00645) Spain.

Josep M. Llovet is Professor of Research at InstitutCatala de Recerca Avancada (ICREA, Generalitat deCatalunya), IDIBAPS, Hospital Clınic.

This work was partially funded by grants from elInstituto Sanitario Carlos III (PI 05/150 and PI 05/1285) and performed through a research contract withBioCompatibles Ltd., UK.

References

[1] Parkin DM, Bray F, Ferlay J, Pisani P. Estimating the worldcancer burden: Globocan 2000. Int J Cancer 2001;94:153–156.

[2] Bosch FX, Ribes J, Diaz M, Cleries R. Primary liver cancer:worldwide incidence and trends. Gastroenterology2004;127:S5–S16.

[3] Brown DB, Cardella JF, Sacks D, Goldberg SN, Gervais DA,Rajan D, et al. Quality improvement guidelines for transhepaticarterial chemoembolization, embolization, and chemotherapeuticinfusion for hepatic malignancy. J Vasc Interv Radiol2006;17:225–232.

[4] Llovet JM, Real MI, Montana X, Planas R, Coll S, Aponte J,et al. Arterial embolisation or chemoembolisation versus symp-tomatic treatment in patients with unresectable hepatocellularcarcinoma: a randomised controlled trial. Lancet2002;359:1734–1739.

[5] Lo CM, Ngan H, Tso WK, Liu CL, Lam CM, Poon RT, et al.Randomized controlled trial of transarterial lipiodol chemoemb-olization for unresectable hepatocellular carcinoma. Hepatology2002;35:1164–1171.

[6] Lewis AL, Gonzalez MV, Lloyd AW, Hall B, Tang Y, Willis SL,et al. DC bead: in vitro characterization of a drug-delivery devicefor transarterial chemoembolization. J Vasc Interv Radiol2006;17:335–342.

[7] Llovet JM, Bru C, Bruix J. Prognosis of hepatocellular carcinoma:the BCLC staging classification. Semin Liver Dis1999;19:329–338.

[8] Bruix J, Sherman M. Management of hepatocellular carcinoma.Hepatology 2005;42:1208–1236.

[9] Bruix J, Sala M, Llovet JM. Chemoembolization for hepatocel-lular carcinoma. Gastroenterology 2004;127:S179–S188.

[10] Ernst O, Sergent G, Mizrahi D, Delemazure O, Paris JC,L’Hermine C. Treatment of hepatocellular carcinoma bytranscatheter arterial chemoembolization: comparison ofplanned periodic chemoembolization and chemoembolizationbased on tumor response. AJR Am J Roentgenol1999;172:59–64.

M. Varela et al. / Journal of Hepatology 46 (2007) 474–481 481

[11] Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan RS,Rubinstein L, et al. New guidelines to evaluate the response totreatment in solid tumors. European Organization for Researchand Treatment of Cancer, National Cancer Institute of the UnitedStates, National Cancer Institute of Canada. J Natl Cancer Inst2000;92:205–216.

[12] Bruix J, Sherman M, Llovet JM, Beaugrand M, Lencioni R,Burroughs AK, et al. Clinical management of hepatocellularcarcinoma. Conclusions of the Barcelona-2000 EASL conference.European Association for the Study of the Liver. J Hepatol2001;35:421–430.

[13] Ballet F, Barbare JC, Poupon R. Hepatic extraction of adriamycinin patients with hepatocellular carcinoma. Eur J Cancer ClinOncol 1984;20:761–764.

[14] de Bruijn P, Verweij J, Loos WJ, Kolker HJ, Planting AS, NooterK, et al. Determination of doxorubicin and doxorubicinol inplasma of cancer patients by high-performance liquid chroma-tography. Anal Biochem 1999;266:216–221.

[15] Llovet JM, Bruix J. Systematic review of randomized trials forunresectable hepatocellular carcinoma: chemoembolizationimproves survival. Hepatology 2003;37:429–442.

[16] Terayama N, Matsui O, Gabata T, Kobayashi S, Sanada J, UedaK, et al. Accumulation of iodized oil within the nonneoplasticliver adjacent to hepatocellular carcinoma via the drainage routesof the tumor after transcatheter arterial embolization. CardiovascIntervent Radiol 2001;24:383–387.

[17] Kalayci C, Johnson PJ, Raby N, Metivier EM, Williams R.Intraarterial adriamycin and lipiodol for inoperable hepatocellu-lar carcinoma: a comparison with intravenous adriamycin. JHepatol 1990;11:349–353.

[18] Ball DS, Heckman R, Olenick SW, Folander HL, Reed 3rd J. Invitro stability of tris-acryl gelatin microspheres in a multiphar-maceutical chemoembolization solution. J Vasc Interv Radiol2003;14:83–88.

[19] Vallee JN, Lo D, Guillevin R, Reb P, Adem C, Chiras J. In vitrostudy of the compatibility of tris-acryl gelatin microspheres withvarious chemotherapeutic agents. J Vasc Interv Radiol2003;14:621–628.

[20] Hong K, Kobeiter H, Georgiades CS, Torbenson MS, GeschwindJF. Effects of the type of embolization particles on carboplatin

concentration in liver tumors after transcatheter arterial chemo-embolization in a rabbit model of liver cancer. J Vasc IntervRadiol 2005;16:1711–1717.

[21] Kishi K, Sonomura T, Nishida N, Satoh M, Yamada R,Yamamoto T, et al. Alginate gel beads for chemoembolization:initial report. Nippon Igaku Hoshasen Gakkai Zasshi1995;55:300–304.

[22] Johnson PJ, Kalayci C, Dobbs N, Raby N, Metivier EM,Summers L, et al. Pharmacokinetics and toxicity of intraarterialadriamycin for hepatocellular carcinoma: effect of coadministra-tion of lipiodol. J Hepatol 1991;13:120–127.

[23] Kwok PC, Lam TW, Chan SC, Chung CP, Wong WK, Chan MK,et al. A randomized clinical trial comparing autologous bloodclot and gelfoam in transarterial chemoembolization for inoper-able hepatocellular carcinoma. J Hepatol 2000;32:955–964.

[24] Castells A, Bruix J, Ayuso C, Bru C, Montanya X, Boix L, et al.Transarterial embolization for hepatocellular carcinoma. Antibi-otic prophylaxis and clinical meaning of postembolization fever. JHepatol 1995;22:410–415.

[25] Geschwind JF, Kaushik S, Ramsey DE, Choti MA, Fishman EK,Kobeiter H. Influence of a new prophylactic antibiotic therapy onthe incidence of liver abscesses after chemoembolization treatmentof liver tumors. J Vasc Interv Radiol 2002;13:1163–1166.

[26] de Baere T, Roche A, Amenabar JM, Lagrange C, Ducreux M,Rougier P, et al. Liver abscess formation after local treatment ofliver tumors. Hepatology 1996;23:1436–1440.

[27] Uchida H, Ohishi H, Matsuo N, Nishimine K, Ohue S, NishimuraY, et al. Transcatheter hepatic segmental arterial embolizationusing lipiodol mixed with an anticancer drug and Gelfoamparticles for hepatocellular carcinoma. Cardiovasc InterventRadiol 1990;13:140–145.

[28] Kawai S, Tani M, Okamura J, Ogawa M, Ohashi Y, Monden M,et al. Prospective and randomized trial of lipiodol-transcatheterarterial chemoembolization for treatment of hepatocellular carci-noma: a comparison of epirubicin and doxorubicin (secondcooperative study). The Cooperative Study Group for LiverCancer Treatment of Japan. Semin Oncol 1997;24:S6-38–S6-45.

[29] Kim W, Clark TW, Baum RA, Soulen MC. Risk factors for liverabscess formation after hepatic chemoembolization. J Vasc IntervRadiol 2001;12:965–968.