Int. J. Radiation Oncology Biol. Phys., Vol. 69, No. 5, pp. 1448–1455, 2007Copyright � 2007 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/07/$–see front matter

doi:10.1016/j.ijrobp.2007.05.009

CLINICAL INVESTIGATION Liver

INTRA-ARTERIAL RHENIUM-188 LIPIODOL IN THE TREATMENT OFINOPERABLE HEPATOCELLULAR CARCINOMA: RESULTS OF AN

IAEA-SPONSORED MULTINATION STUDY

PATRICIA BERNAL, M.D.,* JEAN-LUC RAOUL, M.D., PH.D.,y GAJ VIDMAR, PH.D.,z

ERDENECHIMEG SEREEGOTOV, M.D.,x FELIX X. SUNDRAM, M.D.,k AJAY KUMAR, M.D.,{

JAE MIN JEONG, M.D., PH.D.,** PAWANA PUSUWAN, M.D.,yy CHAITANYA DIVGI, M.D., PH.D.,zzxx

PAT ZANZONICO, M.D., PH.D.,zz JANEZ STARE, M.D., PH.D.,z JOHN BUSCOMBE, M.D.,kk CHAU TRINH THI

MINH, M.D.,{{ MAUNG MAUNG SAW, PH.D.,k SHAOLIANG CHEN, M.D.,*** RUBEN OGBAC, M.D.,yyy

AND AJIT K. PADHY, M.D.kzzz

*Fundacion Santa Fe De Bogota, Bogota, Colombia; yCentre Regional de Lutte Contre le Cancer, Centre Eugene Marquis, University ofRennes, Rennes, France; z IBMI, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia; xFirst State Central Clinic, National

Medical University of Mongolia, Ulaanbaatar, Mongolia; kSingapore General Hospital, Singapore; {All India Institute of MedicalSciences, New Delhi, India; **Seoul National University Hospital, Seoul, South Korea; yySiriraj Hospital, Mahidol University,

Bangkok, Thailand; zzMemorial Sloan-Kettering Cancer Center, New York, NY; xxDivision of Nuclear Medicine & Clinical MolecularImaging, University of Pennsylvania, Philadelphia, PA; kkRoyal Free Hospital, London, United Kingdom; {{Cho Ray Hospital, Ho Chi

Minh City, Vietnam; ***Zhongshan Hospital, Fudan University, Shanghai, China; yyySt. Luke’s Medical Centre, Quezon City,Philippines; and zzz International Atomic Energy Agency, Vienna, Austria

Purpose: Intra-arterial injections (IAI) of 131I-lipiodol is effective in treating hepatocellular carcinoma patients,but is expensive and requires a 7-day hospitalization in a radioprotection room. 188Re is inexpensive, requires nopatient isolation, and can be used with lipiodol.Methods and Materials: This International Atomic Energy Agency–sponsored phase II trial aimed to assess thesafety and the efficacy of a radioconjugate 188Re + lipiodol (188Re-Lip) in a large cohort of hepatocellular carci-noma patients from developing countries. A scout dose is used to determine the maximal tolerated dose (lungs <12Gy, normal liver <30 Gy, bone marrow <1.5 Gy) and then the delivery of the calculated activity. Efficacy wasassessed using response evaluation criteria in solid tumor (RECIST) and alpha-feto-protein (aFP) levels and severeadverse events were graded using the Common Toxicity Criteria of the National Cancer Institute scale v2.0.Results: The trial included 185 patients from eight countries. The procedure was feasible in all participating cen-ters. One treatment was given to 134 patients; 42, 8, and 1 received two, three, and four injections, respectively. Theinjected activity during the first treatment was 100 mCi. Tolerance was excellent. We observed three complete re-sponses and 19 partial responses (22% of evaluable patients, 95% confidence interval 16–35%); 1- and 2-year sur-vivals were 46% and 23%. Some factors affected survival: country of origin, existence of a cirrhosis, Cancer of theLiver Italian Program score, tumor dose, absence of progression, and posttreatment decrease in aFP level.Conclusions: IAI of 188Re-Lip in developing countries is feasible, safe, cost-effective, and deserves a phase IIItrial. � 2007 Elsevier Inc.

Hepatocellular carcinoma, Radiolabeled lipiodol, 188 Rhenium, Internal radiotherapy.

INTRODUCTION

Hepatocellular carcinoma (HCC) is frequent in Asia and in

Africa, and its frequency is increasing in developed countries.

Most HCCs are unresectable when diagnosed, so the patients

receive palliative treatment (1). Several different palliative

treatments have been developed, but until now, only chemo-

embolization has been validated in randomized controlled

studies. Chemoembolization efficacy is related to ischemia

Reprint requests to: Jean-Luc Raoul, M.D., Ph.D., CentreRegional de Lutte Contre le Cancer Eugene Marquis, 35062 CedexRennes, France. Tel: (+33) 299253172; Fax: (+33) 299253108;E-mail: raoul@rennes.fnclcc.fr

14

and effect of the cytotoxic agent (2). This chemotherapeutic

agent is usually coupled to lipiodol because it is considered

to be a carrier vehicle. Lipiodol could also be coupled with

a radionuclide. Studies with I-131-lipiodol were encouraging:

this treatment was better than best supportive care in HCC

with portal vein thrombosis (3), and comparable to chemoem-

bolization when the portal vein is patent (4). But I-131-lipio-

dol is expensive, has high gamma energy, short beta (i.e.,

Conflict of interest: none.Received March 9, 2007, and in revised form May 2, 2007.

Accepted for publication May 7, 2007.

48

188Re-Lipiodol treatment for hepatocellular carcinoma d P. BERNAL et al. 1449

cytotoxic) range, and requires radiation protection for several

days; hence, there is a need for another radionuclide. Rhe-

nium-188 has many advantages: it can be produced cheaply

and its gamma energy is low, thus reducing the need for

hospitalization, and its beta emission is more energetic with

a consequently greater cytotoxic range (5).

A stable conjugation of Re-188 to Lipiodol was achieved

using HDD (6, 7). Preliminary feasibility studies performed

in Belgium (8) and in Asia (9) confirmed the good tolerance

and gave some promising efficacy results. Using this method-

ology, a therapy trial was conducted under the auspices of

one of the International Atomic Energy Agency’s (IAEA)

Thematic Coordinated Research Projects, entitled ‘‘Manage-

ment of liver cancer using radionuclide methods with special

emphasis on transarterial radioconjugate therapy and internal

dosimetry.’’ The trial was unique in that a single protocol us-

ing a common labeling procedure and dosimetric methodol-

ogy was conducted in eight countries across two continents.

Here we present the results of this multination trial with

emphasis on tolerance and efficacy.

PATIENTS AND METHODS

Patient eligibility criteriaPatients were eligible for this trial if they met the following crite-

ria: age >18 years; histopathologic diagnosis of HCC or association

of an hypervascularized tumor in a cirrhotic/fibrotic liver and alpha-

feto-protein (aFP) higher than 400 mg/L; at least one measurable

tumor greater than 1 cm (unidimensional); acceptable general

condition: Karnofsky performance status (KPS) $60%; serum

creatinine #2 mg/dL; absolute neutrophil count $1,500/mL and

platelets $100 G/L; international normalization ratio #1.5. All

participant patients were required to provide written informed con-

sent and the protocol was approved by each participating center’s

Institutional Medical Ethics Committee.

Patients were excluded if they were pregnant or lactating; had poor

liver function (Child C); extrahepatic metastasis, severe chronic pul-

monary disease, or other serious illness; estimated survival expec-

tancy less than 1 month; or allergy to intravenous contrast media.

MethodsPatient eligibility was confirmed by clinical examination, serum

aFP level, abdominal computed tomography (CT); chest X-ray; bio-

logic tests (blood count; serum creatinine, international normalization

ratio, aspartate serum transglutaminase (AST), alanine transglut-

amines (ALT), gamma glutamyl transpeptidase (GGT), alkaline phos-

phatase, albumin; serology for hepatitis B virus or hepatitis C virus

infection). Child’s score (combination of clinical data: ascites and en-

cephalopathy and prothrombin time, albumin and bilirubin serum

levels) and Cancer of the Liver Italian Program (CLIP) score (10) (a

combination of four parameters: Child’s score, macroscopic aspect

of the tumor, serum aFP level, and portal vein thrombosis) were deter-

mined from these data.

Treatment procedureThe tungsten-188/rhenium-188 generator, which has a long

useful shelf-life of several months, provides a good yield of car-

rier-free rhenium-188 on a routine basis, making possible in-house

preparation of rhenium-labeled radiopharmaceuticals for clinical

use. Rhenium-188 in the form of sodium prephenate was obtained

by eluting the W-188/Re-188 generators (Oak Ridge National Lab-

oratory, TN). The Re-188 prephenate solution was concentrated by

passing through ion exchange columns, as previously described

(11, 12). Re-188 lipiodol was prepared according to the procedures

described previously (6, 7, 13) using the lyophilized HDD kits

(Seoul National University Hospital, Korea).

Total liver and hepatic tumor mass measurements were made from

CT scans. The patient was catheterized in the interventional radiol-

ogy suite, and the catheter placed in a location suitable for selective

hepatic lesion intra-arterial therapy. After the catheter position had

been verified, approximately 150 MBq (4 mCi) Re-188 Lipiodol

was injected (‘‘scout’’ dose). The patient was then taken immediately

to the nuclear medicine department. Anterior and posterior whole

body scans at 20 cm/min or 5-min spot chest and abdomen images

were obtained to permit conjugate-view methodology. Images

were obtained with the photopeak energy window centered at 155

keV, with a symmetric 20% window. Geometric means of total

counts in the liver, the liver tumor, and the lungs were obtained.

These were entered into a spreadsheet developed by the IAEA Coor-

dinated Research Project (CRP) group that allowed estimates of

source and target region activities, cumulative activities and ab-

sorbed doses to be calculated.

The patient then returned to the interventional radiology suite,

where catheter tip position was confirmed before injection of thera-

peutic Re-188 lipiodol was carried out. The therapeutic amount of

radioactivity was no more than that calculated to deliver 1.5 Gy to

marrow or 30 Gy to liver or 12 Gy to lung. Then the patient was

transferred to a radiation isolation area until his or her general con-

dition and radiation safety regulations permitted discharge (usually

1 day). Immediately before discharge, anterior and posterior images

similar to those obtained after the scout dose were carried out.

Follow-upDuring the hospitalization, all side effects were checked. Biologic

tests were repeated on Day 7 and at 2 months, and CT scan at 2 and 6

months. Patients could receive further injections after the third

month. The first patient was enrolled in the study at the end of the

year 2000; patient vital status was followed until the end of 2005.

EfficacyEfficacy was assessed regarding tumor size (using Response

Evaluation Criteria in Solid Tumor [RECIST] criteria) and aFP

level. Unfortunately, in the majority of responders, the response

could not be confirmed after 4 weeks (no CT scan was performed).

Regarding aFP, if initial level was more than five times the local

upper normal limit, patients were classified into four categories of

biochemical response: normalization, complete biochemical re-

sponse; partial biochemical response, decrease by more than 50%;

stable, change between –50% and +50%; and biochemical progres-

sion, more than 50% increase.

In addition, a physician’s biochemical efficacy rating on the

equivalent four-point scale was given by the treating physician for

each patient, regardless of the initial aFP value.

Efficacy regarding tumor size was assessed in the patients that re-

ceived at least one treatment and had the CT scan after 2 months,

whereas aFP analysis was based on the patient’s best result.

ToxicityToxicity was assessed using the Common Toxicity Criteria of the

National Cancer Institute (CTC-NCI) scale v2.0 (Grade 0 to 4) based

on the patient’s worse result (either immediately after treatment or at

follow-up). Adverse effects were also clinically classified as mild,

1450 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 5, 2007

moderate (requiring simple treatment), severe (requiring hospitali-

zation or important treatment), or life-threatening. All side effects

occurring within the first 2 months were considered as possibly re-

lated to the treatment and analyzed; after 2 months, side effects were

presumed to be related to tumor progression or to liver disease and

were not recorded as treatment-related toxicities. Global toxicity

was rated by the physician on a per-patient basis as none, mild, mod-

erate, severe, or life-threatening.

Statistical analysisThis trial was designed at the first expert meeting in Singapore

(1999); it was decided that the major measure of efficacy was to be

a simple binary variable (response/no response). It was further esti-

mated that the response rate would be around 40% and that we

wanted to estimate this proportion with a 10% precision. This gave

us a required sample size of 234; the first patient was entered at the

end of 2000. We decided to prematurely stop the trial in mid-2005

after inclusion of 185 patients because of the slow accrual rate.

Frequency tables were produced for all the available patient char-

acteristics and outcome variables, and descriptive statistics were cal-

culated. For dichotomous outcomes, exact binomial confidence

intervals were calculated. Population survival curves were estimated

using the Kaplan-Meier procedure, and the differences in survival

regarding prognostic factors were tested using the Cox regression

model. Association between categorical variables was tested using

Fisher’s exact test, and association of a binary outcome with two

or more prognostic factors was tested using logistic regression;

association between categorical and numerical variables was tested

using the exact version of the Mann-Whitney test or the Kruskal-

Wallis test. Statistical analyses were performed with SPSS 13.0

(SPSS Inc., 2004) and Cytel Studio 6.3.0 (Cytel Corp., 2004) soft-

ware.

RESULTS

The trial comprised 185 patients from eight countries: 13

from China, 10 from Colombia, 41 from India, 31 from Mon-

golia, 8 from the Philippines, 17 from Singapore, 13 from

Thailand, and 52 from Vietnam. Patient age ranged from

22 to 84 years (median 55, mean 55.4, SD 11.8 years). There

were 146 (79%) men and 39 (21%) women. Of the 145

patients with available information, 77 (53%) had cirrhosis;

of the 112 patients with available information, 18 (16%)

had portal vein thrombosis. Child score was recorded for

115 patients, 88 (77%) of whom were classified as A. KPS

was recorded in 170 patients; it was 100 or 90 in 75 (44%)

patients, and 70 or 60 in 46 (27%). Underlying liver pathol-

ogy was evaluated in 177 patients: hepatitis B virus was pres-

ent in 77 (44%), hepatitis C virus in 22 (12%), combination of

both in 7 patients (4%), alcoholism in 4 (2%), and other pa-

thology in 3 (2%); none was recorded in 64 patients (36%).

One tumor was found in 155 patients (84%); 15 (8%) had

two, 6 (3%) three, and 9 patients (5%) had four tumors. The

largest tumor diameter ranged from 1 to 23 cm (median 9.2

cm, mean 9.1 cm, SD 4.4 cm), and from 1 to 2,600 cm3 in

volume. The ratio of tumor volume to liver volume ranged

from less than 0.01 to 6.27 (median ratio 0.28), and in 24%

of the patients the principal tumor represented more than

50% of the liver volume. aFP level was above higher than

mg/L in 34% of patients, and it exceeded the upper normal

limit by a factor larger than 5 in 61% of the patients. CLIP

score (95 patients) was 0 in 7% of the patients, 1 in 17%, 2

in 44%, 3 in 28%, and 4 in 4% of those patients.

There were large differences between the countries regard-

ing cirrhosis (p < 0.001), Child score (p < .001), and CLIP

score (p = .001). The percentage of cirrhotic patients varied

from 0% to 80% and, among cirrhotic patients, from 0% to

50% were assigned Child B score. The average CLIP score

varies regarding the countries from 1.4 to 2.4; the highest

scores were recorded in countries having the lowest rate of

cirrhotic patients.

A single treatment was given to 134 patients (72%), 42

(23%) received two treatments, 8 (4%) three, and 1 patient

four. Estimates of radiation absorbed dose were calculated

after the scout dose, and in 32% of patients the dose limiting

organ was the lungs, whereas in the remaining 68%, it was

the liver. The total injected activity (including the scout

dose) during the first treatment ranged from 21.2 to 363.4

mCi (mean 108.1 mCi, median 100.0 mCi, SD 53.8 mCi).

Radiation absorbed dose to the primary tumor (index lesion)

with the first treatment ranged from 1 to 304.7 Gy (median

46 Gy, mean 63.4 Gy, SD 59.9 Gy).

ToxicityImmediately after the injection, 25 patients experienced

pain in the right hypochondrium and 10 experienced vomit-

ing; in the first few days, 34 had mild or moderate fever

(Table 1). Within the first 2 months, 28 patients (15%)

died; according to the physician opinion, the death was

directly due to the tumor in 18 cases, and possibly related

to other factors in 10 patients.

After the first injection, Grade 3 or 4 toxicity was found in

22 patients. Liver toxicity was the most frequent. However,

an increase in bilirubin level was not considered as reflecting

therapeutic toxicity: according to the CTC-NCI scale, 33 pa-

tients had initial values corresponding to bilirubin grade 3 or

4 toxicity (>3.0 upper limit of normal [ULN]) before the

treatment and 40 after the injections (only these 7 cases

were considered as developing a Grade 3-4 toxicity). Hema-

tologic toxicity was observed in 6 patients. Some other tox-

icities were noted including one case of pneumopathy. Two

further cases of pneumopathy were recorded among the 42

patients who received a second injection.

Grade 3 or 4 toxicity incidence differed between countries,

ranging from 0% to 26%. The association of Grade 3 or 4 tox-

icity with cirrhosis (toxicity present in 17% of cirrhotic and

6% in noncirrhotic patients) was not statistically significant;

among the cirrhotic patients, presence of toxicity was mar-

ginally associated with Child score B as compared with A

(logistic regression, adjusting for country: p = 0.071). There

was no association between toxicity and total injected activ-

ity, but KPS >90 was marginally associated with lower odds

for toxicity (logistic regression, adjusting for country: p =

0.057). Finally, that liver was the dose limiting organ in

more than two-thirds of the patients did not increase the

odds for hepatic toxicity as measured by changes from base-

line in serum bilirubin, AST, or ALT (p = 0.087).

188Re-Lipiodol treatment for hepatocellular carcinoma d P. BERNAL et al. 1451

When clinically rated on a 5-point scale, adverse effects

were rated as none in 68% of the cases, mild in 25%, moder-

ate in 3%, severe in 3%, and life-threatening in 1% of the

cases.

EfficacyData on largest tumor diameter after therapy and tumor re-

sponse as evaluated from CT scans were available for 88 pa-

tients (48%). Complete response was recorded in 3 patients

(3%; 95% CI, 1–9%), and partial response was reported in

19 patients (22%; 95% CI, 14–32%); so the objective re-

sponse rate was 25% (95% CI, 16–35%), which corresponds

to 12% on the intention-to-treat basis; stable disease was re-

ported for 47 patients (53%; 95% CI, 43–64%), and tumor

progression for 19 (22%; 95% CI, 14–32%).

Among the 86 patients with elevated aFP level and avail-

able follow-up, complete biochemical response was recorded

in 7 patients (8%; 95% CI, 4–16%), partial biochemical re-

sponse in 18 patients (21%; 95% CI, 13–30%), stable disease

in 55 patients (64%; 95% CI, 53–74%), and progression in 6

patients (7%; 95% CI, 3–14%). Because only a small and se-

lected group of patients with available tumor response data

had elevated initial aFP level, the comparison between tumor

response and biochemical response is not informative. There

exists a significant agreement between tumor response

and physician’s biochemical response rating (p < 0.001)

(Table 2).

Although total injected activity during the first treatment

was not associated with tumor progression (p = 0.995), prin-

cipal tumor dose during first treatment was p = 0.014 among

the patients receiving more than 30 Gy, the proportion of tu-

mor progression was lower (9%) than the proportion (35%)

Table 1. Recorded toxicity after the first therapeutic injectionof Re-188 lipiodol according to CTC-NCI scale 2.0 in the

treated population (n = 185)

Grade 3n

Grade 4n

Hematologic toxicityWBC 2 0Platelets 2 0Hemoglobin 0 3

Hepatic toxicityAST/ALT 14 10Bilirubin (pretreatment abN) 5 (27) 2 (6)

Digestive toxicityUlcer 1 0Diarrhea 0 1

Other toxicitiesHypotension 1 0pneumopathy 0 1

Clinically significant SAE 5 2

The number of patients who had abnormal pretreatment values ofbilirubin is in parentheses. Clinically significant side effects weredescribed in only 7 patients.

Abbreviations: CTC-NCI = Common Toxicity Criteria of theNational Cancer Institute; WBC = white blood cell count; AST =aspartate serum transglutaminase; ALT = alanine transglutamines;abN = abnormal value; SAE = severe adverse event.

among those with lower tumor dose (estimated odds ratio

0.18; 95% CI, 0.03–0.85).

CLIP score, Child class, tumor presentation (three cate-

gories: uninodular tumor less than 50% of liver size, uninod-

ular >50%, or multinodular), and KPS (90 or more vs. less)

were not statistically significantly associated with (lack of)

tumor progression (p = 0.368, 1.000, 0.521, and 0.295,

respectively).

SurvivalThe Kaplan-Meier survival curve for the treated popula-

tion is given in Figure 1. Estimated median survival was

256 days (95% CI, 133–79 days). Estimated 3-, 6-, 12-, and

24-month survival was 93%, 60%, 46%, and 23%, respec-

tively. Median follow-up was 455 days (95% CI, 407–502

days). There were differences between countries regarding

patient survival (log–rank test: p = 0.006; estimated median

survival per country ranged from 175 days to 518 days).

Cirrhosis was a significant negative prognostic factor (Cox

regression, adjusting for country and tumor size: p = 0.021);

observed median survival for cirrhotic patients was 218 days

vs. 246 for noncirrhotic patients. Among cirrhotic patients,

Child score had statistically significant effect on survival

(Cox regression, adjusting for country and tumor size: p =

0.004); patients with score A had better survival (median

625 days) than patients with score B (median 161 days).

Survival was also related to CLIP score (Cox regression, ad-

justing for country: p = 0.015), with survival decreasing with

increasing score. Principal tumor dose during first treatment

significantly influenced survival: patients receiving more

than 30 Gy had significantly better survival (Cox regression

adjusting for country and tumor size: p = 0.006) than patients

receiving less than 30 Gy (median survival, 471 days vs. 146

days; estimated hazard ratio, 0.40; 95% CI, 0.21–0.78). As

illustrated in Figure 2, lack of tumor progression was also

a significant predictor of survival (Cox regression adjusting

for country: p < 0.001); patients with objective response or

stable disease had much better survival than those with

progressive disease (median survival, 532 days vs. 198

days; estimated hazard ratio, 0.24; 95% CI, 0.12–0.47). Fi-

nally, among the patients with initial elevated aFP, partial

and complete responders survived longer than the remaining

patients (Cox regression adjusting for country: p < 0.001;

median survival, 518 days vs. 205 days; estimated hazard

ratio, 0.37; 95% CI, 0.18–0.77).

Table 2. Agreement between tumor response rating andphysician’s biochemical response rating

Physicians’ biochemicalresponse rating

Tumor response Complete Partial Stable Progression Total

Complete 2 0 0 0 2Partial 2 9 4 3 18Stable 5 5 27 9 46Progression 0 3 2 11 16Total 9 17 33 23 82

1452 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 5, 2007

DISCUSSION

Hepatocellular carcinoma is relatively very frequent in de-

veloping countries. By contrast, most therapeutic series of

cases come from developed countries, so the relevance of

their conclusions for developing countries remains question-

able. In a palliative setting, only chemoembolization has been

demonstrated as better than best supportive care, in two ran-

domized controlled trials (2, 14); unfortunately side effects

are frequent and could need intensive care for some days.

In another randomized trial, 131I-labeled lipiodol injections

have been demonstrated as efficient but better tolerated

than chemoembolization (4). But this treatment requires for

radioprotection purposes an isolation of some days (depend-

Fig. 1. Estimated population survival curve (black line) with 95%confidence interval limits (gray lines); the 1 and 2-year overallsurvival were, respectively, 46% and 23%.

Fig. 2. Estimated survival curve for patients without (dashed grayline) and with tumor progression (solid black line). Vertical linesdenote censored cases; labels on horizontal axis denote mediansurvival time for the two groups.

ing on local regulations). Both options are attractive and

could be use in developed countries.

The advantages of the use of intra-arterial 188Re-lipiodol

injection as compared with other intra-arterial treatments are

as follows.

1. It is easier to deliver compared with chemoembolization

(injection within the common hepatic artery) or compared

with radioembolization using 90Yttrium microspheres

(15). The technique was feasible in all the participating

countries after a short learning period. Moreover, its

good tolerance avoids a long hospitalization period and

expensive care.

2. It is cheaper than I-131-lipiodol and requires no patient

isolation.

3. The generator can be used for other kinds of treatments.

Thus this cost-effective therapeutic option is of potential

interest in developing countries.

This study was the first multicenter study on treatment of

liver cancer built and conducted in developing countries. It in-

cluded very distant and different countries, with the conse-

quence of considerable heterogeneity in patients and thus

results. Nevertheless, the study shows that it is possible to

conduct a scientific medical study in developing countries.

This great heterogeneity between the different countries

seems to be related to the disease (e.g., etiology, stage at diag-

nosis) as well as to caregiver access. The treatment was per-

formed properly in a large majority of the patients and the

final outcome (dead or alive and date of death) could be ob-

tained for all the included patients. Obviously, as presumed,

the quality of the data is lower than in studies conducted in

richer countries. An important number of initial or follow-

up data were missing. That was mainly related to the vague-

ness of some reports and the absence of clear normal values

from some laboratories, contrasting with the strictness of

data managers (J.L.R., G.V.) (e.g., if the portal trunk was

not described as thrombosed or free of thrombus, these data

were considered as missing) but also, to the difficulties in

some countries to perform a good initial work-up and to care-

fully follow-up patients because of distance and costs. But this

was improving, and for the last included patients, most data

could be retrieved. This learning curve was observed in all

countries. It is likely that future trials in these hospitals will

be of better quality. This weakness in reliability was compen-

sated by the large number of included patients in this phase II

trial. This trial also taught the experts that the fewer data are

required, the better the quality. In the next trials, we will

have to shorten the number of required data to have a simpler

and more reliable study. That means that with some very sim-

ple precautions phase II and phase III studies can be per-

formed in developing countries.

The included population has some unique features of interest:

� underlying liver disease is due to a viral hepatitis (B more

frequently than C) in more than two thirds of the cases and

only half of the population had a liver cirrhosis; this profile

is usual in Asian patients;

188Re-Lipiodol treatment for hepatocellular carcinoma d P. BERNAL et al. 1453

� multinodular tumors are infrequent and the tumor is unique

in most of the cases;

� in most cases, poor prognosis factors (16) are present: large

size (mean diameter: 9 cm), volume greater than half the

liver volume in one fourth of the cases, elevated aFP level

(>400 ng/mL) in one third, portal vein thrombosis in one in

six, severe hyperbilirubinemia in 18%, and poor KPS in

27%. This is reflected by the CLIP scoring system: one

third of the patients had a score of 3 or more corresponding

in the Italian experience (17) to a median predicted sur-

vival of less than 4.5 months.

The treatment was easy to deliver, no major problem was

recorded; repeated injections have been performed in 51 pa-

tients and 9 patients received three or four injections without

evidence of cumulated toxicity.

The clinical tolerance of this treatment was good. For the

physician in charge, the adverse effects were classified as

none or mild in more than 90% of the cases. Immediate side

effects (e.g., pain, vomiting, fever) were mild and directly re-

lated to the arterial infusion. These side effects were also ob-

served after injection of 131I-lipiodol and are usual and could

be severe after chemoembolization (postembolization syn-

drome) (18). As expected, the late toxicities are infrequent

and, except pneumopathies could be unrelated to the treat-

ment; they tended to be more frequent in Child B patients

and in patients with lowest KPS. Three cases of pneumopathy

have been described. Such side effects have been recorded af-

ter intra-arterial injections of radiolabeled lipiodol (19) and

labeled microspheres (20). After use of microspheres, pneu-

mopathies seemed the result of an increased intrahepatic

shunting (greater than 13%) conducting to a lung irradiation

of more than 10 Gy (10–36 Gy; median 25 Gy) and then to

a radiation pneumonitis. The explanation does not seem so

univocal with radiolabeled lipiodol, and pneumopathies

could correspond either to a radiation pneumonitis or to an al-

lergic reaction to lipiodol. In favor of the latter hypothesis are:

they occurred in patients with low pulmonary uptake, that

most cases occurred after a second injection, and responded

well to corticosteroids (21). We observed the same lung tox-

icity profile in this trial. One patient (of 185) who received

more than the limit for pulmonary toxicity (12 Gy) developed

pulmonary toxicity after one injection. This case illustrates

the importance of pulmonary absorbed dose measurements

and the importance of using imaging for assessment of biodis-

tribution. Two patients among the 51 who received repeated

injections (4%) also had pulmonary toxicity; the cumulative

dose in these 2 patients was less than the presumed maximum

permissible absorbed dose, suggesting that insult to the lungs

may well be additive or even synergistic or related to another

factor (allergic manifestation ?).

Biologic tolerance was also excellent. As it could be sup-

posed from lipiodol distribution (22), hematologic toxicity

was very infrequent. Liver toxicity was usually infrequent

and mild, but the NCI-CTC grading is difficult to use in

patients with severe chronic liver disease; a Grade 3 transam-

inases toxicity is reached when the transaminase level is

greater than 5 ULN and this is common in cases of chronic

viral hepatitis (60% of our population). Bilirubin could also

be elevated from chronic liver disease or the tumor rather

than to a toxic event; a Grade 3 is reached when the bilirubin

level is 3 ULN. This was the case in 40% of our population

during the initial 2-month follow-up, but half of these pa-

tients had at least a Grade 3 before injection. We have noticed

that liver toxicity does not seem to be strongly related to liver

irradiation because it was not more frequent in patients in

whom the liver was the dose limiting organ.

The deaths in the first 2 months seemed, according to the

physician assumption, to be more related to liver disease

and tumor than to an unexpected toxicity. Namely, among

the patients who died within the first 2 months, the proportion

of Child B (53%) was significantly higher (p = 0.009) than

among the other patients (20%).

Regarding the usual aggressive behavior of HCC and the

rarity of spontaneous stability, we have data to conclude

that this treatment by Re-188 lipiodol is efficient. Our statis-

tics were based on a supposed response rate of around 40% as

observed in first trial with radiopharmaceutics. Our data were

far lower. But, the objective response rate of 25% in evalu-

able patients is of value taking into account that most in-

cluded patients had huge lesions at inclusion and received

only one injection. In previous studies with chemoemboliza-

tion, the response rate ranged from 60% to 15%, but was

dependent on inclusion criteria (better responses in case of

small tumors or measurable tumors); it was 25% in both

arms among evaluable patients included in the comparison

of I131-Lip vs. chemoembolization (4). In large tumors

treated by chemoembolization, an objective response rate

of 10% has been observed by Ryder et al. (23) contrasting

with an excellent response rate of 56% in small lesions in

the same retrospective series. In our study, the therapeutic

control of tumor size (one of the main therapeutic objectives

in a palliative setting) could be achieved in more than three

quarters of the patients. This tumor control was clearly re-

lated to the treatment: a tumor radiation absorbed dose less

than 30 Gy (obtained with the first treatment) corresponded

to a fourfold increase in the progression rate compared with

patients whose tumors received >30 Gy. This relationship

between tumor delivered dose and response rate has been

hypothesized in another study using 131-Lip (24), which

showed an inverse relationship between tumor size and intra-

tumoral retention of lipiodol and a direct relationship

between intratumoral retention and response rate.

It is obviously very difficult to draw any conclusions on

survival data from a phase II study. We do not have any

idea of survival with only best supportive care in these

different populations of HCC patients and as noticed in

the analysis, survival is very different from one country to

another. Surprisingly, the 1- and 2-year survivals of 46%

and 23% are in accordance with most results obtained after

chemoembolization (25) in Western countries. A better

1-year survival was observed when inclusion criteria in ran-

domized studies became more stringent, as in the Spanish

trial (26).

1454 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 5, 2007

Some factors of prognostic value were found in this study.

The presence and severity of cirrhosis had an adverse impact;

the prognostic value of the CLIP score was confirmed in

these Asian patients. Tumor control and biochemical re-

sponse were also associated with improved survival. The

principal tumor dose received during the first injection influ-

enced the prognosis; if tumor received less than 30 Gy the

median survival was less than 5 months as compared with

a median survival of more than 15 months if more than 30

Gy was delivered. But the real effect of this treatment on sur-

vival will only be assessed by a randomized controlled study

versus best supportive care.

CONCLUSIONS

Hepatocellular carcinoma is a disease predominantly ob-

served in developing countries; until now, almost all thera-

peutic studies had been conducted in developed countries

and their conclusions could not be proposed to high preva-

lence countries because of their cost or of the need for some

specific skills. This study is the first large-scale phase II study

conducted in the field of HCC in developing countries and

demonstrates that a multicenter study of this nature is feasible

in a range of developing countries and could participate to

develop some efficient and easy to perform treatments.

The main weakness of this study is the large number of

missing data, but this was compensated by the huge number

of patients included in this phase II study; such results in

a phase II study including some dozen of patients, as ob-

served in Western countries, would have forbidden any

conclusion. This study demonstrated that this treatment is

efficient and well tolerated, particularly in patients with

well-compensated liver cirrhosis and good general status.

That will deserve to go further in a phase III randomized trial

having as the primary objective overall survival. We have

found some factors of prognostic value: toxicity was more

frequent if initial general status was poor; response rate

was better if the principal tumor dose was high and higher

than 30 Gy; and survival was dependent not only on usual

prognostic factors found in HCC (Child’s score, CLIP score),

but also to factors associated with treatment efficacy (im-

provement in aFP level, tumor control) and directly to the

delivered treatment (tumor dose).

At last, it will be of importance to take advantage of the in-

terest showed by participating physicians to conduct some

other studies using this network of motivated hospitals, and

in the field of HCC to conduct a randomized phase III clinical

trial comparing the presented treatment with chemoemboliza-

tion, because the latter is a therapeutic option widely used in

developing countries. A pragmatic phase III randomized

study involving a large number of patients and focused in

some major end points of that type will certainly be possible

and the conclusions drawn from it, would be more applicable

to developing countries (and to patients in nonacademic cen-

ters from developed countries) than those coming from some

selected academic US, European, or Japanese centers.

REFERENCES

1. Llovet JM, Burroughs A, Bruix J. Hepatocellular carcinoma.Lancet 2003;362:1907–1917.

2. Llovet JM, Real MI, Montana X, et al. Arterial embolisation orchemoembolisation versus symptomatic treatment in patientswith unresectable hepatocellular carcinoma: A randomisedcontrolled trial. Lancet 2002;359:1734–1739.

3. Raoul JL, Guyader D, Bretagne JF, et al. Randomized con-trolled trial for hepatocellular carcinoma with portal vein throm-bosis: Intra-arterial iodine-131-iodized oil versus medicalsupport. J Nucl Med 1994;35:1782–1787.

4. Raoul JL, Guyader D, Bretagne JF, et al. Prospective random-ized trial of chemoembolization versus intra-arterial injectionof 131I-labeled-iodized oil in the treatment of hepatocellularcarcinoma. Hepatology 1997;26:1156–1161.

5. De Ruyck K, Lambert B, Bacher K, et al. Biologic eosimetry of188Re-HDD/lipiodol versus 131I-lipiodol therapy in patientswith hepatocellular carcinoma. J Nucl Med 2004;45:612–618.

6. Lee YS, Jeong JM, Kim YJ, et al. Synthesis of 188Re labeledlong chain alkyl diaminedithiol for therapy of liver cancer.Nucl Med Comm 2002;23:237–242.

7. Paeng JC, Jeong JM, Yoon CL, et al. Lipiodol solution of188Re-HDD as a new therapeutic agent for transhepatic arterialembolization in liver cancer: Preclinical study to rabbit livecancer model. J Nucl Med 2003;44:2033–2038.

8. Lambert B, Bacher K, De Keukeleire K, et al. 188Re-HDD/Lipiodol for treatment of hepatocellular carcinoma: A feasibilitystudy in patients with advanced cirrhosis. J Nucl Med 2005;46:1326–1332.

9. Sundram F, Chau TC, Onkhuudai P, et al. Preliminary results oftransarterial rhenium-188 lipiodol in the treatment of inoperable

primary hepatocellular carcinoma. Eur J Nucl Med Mol Imaging2004;31:250–257.

10. The Cancer of the Liver Italian Program (CLIP) investigators. A

new prognostic system for hepatocellular carcinoma: A retro-

spective study of 435 patients. Hepatology 1998;28:751–755.11. Guhlke S, Beets AL, Mirzadeh S, et al. Tandem cation/anion

column treatment of ammonium acetate eluants of the

alumina-based tungsten-188/rhenium-188 generator: A new

simple method for effective concentration of rhenium-188

prephenate solutions. J Nucl Med 2000;41:1271–1278.12. Knapp FF Jr., (Russ), Turner JH, Jeong J-M, et al. Issues asso-

ciated with the use of tungsten-188/rhenium-188 generator and

concentrator system and preparation of Re-188 HDD: A report.

World J Nucl Med 2004;3:137–143.13. Jeong JM, Kim YJ, Lee YS, et al. Lipiodol solution of a lipo-

philic agent, 188ReTDD, for the treatment of liver cancer.

Nucl Med Biol 2001;28:197–204.14. Lo CM, Ngan H, Tso WK, et al. Randomised controlled trial of

transarterial lipiodol chemoembolization for unresectable hepa-

tocellular carcinoma. Hepatology 2002;35:1164–1171.15. Salem R, Thurston KG. Radioembolization with 90Yttrium

microspheres: A state of the art brachytherapy treatment for pri-

mary and secondary liver malignancies. J Vasc Interv Radiol2006;17:1251–1278.

16. Bruix J, Llovet JM. Prognostic prediction and treatment strate-

gies in hepatocellular carcinoma. Hepatology 2002;35:519–

524.17. The Cancer of the Liver Italian Program (CLIP) investigators.

Prospective validation of the CLIP score: A new prognostic

188Re-Lipiodol treatment for hepatocellular carcinoma d P. BERNAL et al. 1455

system for patients with cirrhosis and hepatocellular carcinoma.Hepatology 2000;31:840–845.

18. Leung DA, Goin JE, Sickles C, Raskay BJ, Soulen MC. Deter-minants of postembolization syndrome after hepatic chemo-embilization. J Vasc Interv Radiol 2001;12:321–326.

19. Raoul JL, Boucher E, Olivie D, et al. Association of cisplatinand intra-arterial injection of (131)I-lipiodol in treatment ofhepatocellular carcinoma: Results of a phase II trial. Int J RadiatOncol Biol Phys 2006;64:745–750.

20. Leung TW, Lau WY, Ho SK, et al. Radiation pneumonitis afterselective internal radiation treatment with intraarterial 90Yt-trium-microspheres for inoperable hepatic tumors. Int J RadiatOncol Biol Phys 1995;33:919–924.

21. Raoul JL, Bretagne JF, Caucanas P, et al. Internal radiationtherapy for hepatocellular carcinoma. Results of a Frenchmulticenter phase II trial of transarterial injection of iodine131-labeled-lipiodol. Cancer 1992;69:346–352.

22. Raoul JL, Bourguet P, Bretagne JF, et al. Hepatic artery injec-tion of I131 labelled lipiodol. I. Biodistribution study resultsin patients with hepatocellular carcinoma and liver metastases.Radiology 1988;168:541–545.

23. Ryder SD, Rizzi PM, Metivier E, et al. Chemoembolisationwith lipiodol and doxorubicin: Applicability in British patientswith hepatocellular carcinoma. Gut 1996;38:125–128.

24. Raoul JL, Messner M, Boucher E, et al. Preoperative treatmentof hepatocellular carcinoma with intra-arterial injection of 131I-labelled Lipiodol. Br J Surg 2003;90:1379–1383.

25. Geschwind JFH, Ramsey DE, Choti MA, et al. Chemoemboli-zation of hepatocellular carcinoma. Results of a meta-analysis.Am J Clin Oncol 2003;26:344–349.

26. Llovet JM, Real MI, Montana X, et al. Arterial embolisation orchemoembolisation versus symptomatic treatment in patientswith unresectable hepatocellular carcinoma: A randomisedcontrolled trial. Lancet 2002;359:1734–1739.