CLINICAL STUDY

Metaanalysis of Survival, Complications, andImaging Response following Chemotherapy-based

Transarterial Therapy in Patients with UnresectableIntrahepatic Cholangiocarcinoma

Charles E. Ray, Jr, MD, PhD, Anthony Edwards, MD, Mitchell T. Smith, MD, MS,Stephen Leong, MD, Kimi Kondo, DO, Matthew Gipson, MD,

Paul J. Rochon, MD, Rajan Gupta, MD, Wells Messersmith, MD,Tom Purcell, MD, MBA, and Janette Durham, MD, MBA

ABSTRACT

Purpose: Unresectable intrahepatic cholangiocarcinoma represents a devastating illness with poor outcomes when treated with

standard systemic therapies. Several smaller nonrandomized outcomes studies have been reported for such patients undergoing

transarterial therapies. A metaanalysis was performed to assess primary clinical and imaging outcomes, as well as complication rates,

following transarterial interventions in this patient population.

Materials and Methods: By using standard search techniques and metaanalysis methodology, published reports (published in 2012

and before) evaluating survival, complications, and imaging response following transarterial treatments for patients with unresectable

intrahepatic cholangiocarcinoma were identified and evaluated.

Results: A total of 16 articles (N ¼ 542 subjects) met the inclusion criteria and are included. Overall survival times were

15.7 months � 5.8 and 13.4 months � 6.7 from the time of diagnosis and time of first treatment, respectively. The overall weighted

1-year survival rate was 58.0% � 14.5. More than three fourths of all subjects (76.8%) exhibited a response or stable disease on

postprocedure imaging; 18.9% of all subjects experienced severe toxicities (National Cancer Institute/World Health Organization

grade Z 3), and most experienced some form of postembolization syndrome. Overall 30-day mortality rate was 0.7%.

Conclusions: As demonstrated by this metaanalysis, transarterial chemotherapy-based treatments for cholangiocarcinoma appears to

confer a survival benefit of 2–7 months compared with systemic therapies, demonstrate a favorable response by imaging criteria, and

have an acceptable postprocedural complication profile. Such therapies should be strongly considered in the treatment of patients with

this devastating illness.

ABBREVIATIONS

HCC = hepatocellular carcinoma, RECIST = Response Evaluation Criteria In Solid Tumors

& SIR, 2013

J Vasc Interv Radiol 2013; 24:1218–1226

http://dx.doi.org/10.1016/j.jvir.2013.03.019

From the SIR 2012 Annual Meeting.

C.E.R. is a paid consultant for Nordion (Ottawa, Ontario, Canada). None of the

other authors have identified a conflict of interest.

From the Departments of Radiology (C.E.R., A.E., M.T.S., K.K., M.G., P.J.R.,

R.G., J.D.) and Medical Oncology (S.L., W.M., T.P.), University of Colorado

School of Medicine, Anschutz Medical Campus, Mail Stop C276, Leprino

Office Building, 12401 E. 17th Ave., no. 526, Aurora, CO 80045. Received

February 1, 2013; final revision received and accepted March 15, 2013.

Address correspondence to C.E.R.; E-mail: charles.ray@ucdenver.edu

Patients diagnosed with unresectable cholangiocarcinoma

have a poor prognosis, with 5-year survival rates of

approximately 5% and typical survivals times of 5–13

months (1,2). Locoregional therapies, similar to those that

have been shown to prolong survival in other unresectable

hepatic malignancies, including hepatocellular carcinoma

(HCC) (3–5), have also been evaluated in the treatment of

cholangiocarcinoma. Historically, it was difficult to differ-

entiate HCC from cholangiocarcinoma, particularly in the

absence of advanced imaging techniques or biopsy. Many

early studies evaluating outcomes of treatments offered for

primary liver cancer likely inadvertently combined data for

patients with HCC and cholangiocarcinoma. Now under-

stood to be distinct entities, HCC and cholangiocarcinoma

Volume 24 ’ Number 8 ’ August ’ 2013 1219

have been shown to respond differently to systemic

chemotherapy and radiation therapy, and indeed have

different molecular pathways believed to be responsible

for their genesis and progression (6). Cholangiocarcinoma

is a much less common malignancy, with consequently less

robust evidence on the efficacy of chemotherapy-based

transarterial therapies, but several smaller nonrandomized

studies have been published demonstrating variable effi-

cacy of such treatments (7–22).

To resolve questions regarding the survival, response per

Response Evaluation Criteria In Solid Tumors (RECIST),

and toxicity profile following chemotherapy-based trans-

arterial therapy in patients with unresectable cholangiocar-

cinoma, a metaanalysis of the literature published during or

before 2012 was performed. The hypothesis to be tested

was that transarterial therapies may confer improved tumor

response and survival benefit with acceptable toxicities

compared with standard chemotherapies.

MATERIALS AND METHODS

Literature searches were performed in PubMed and

Embase in August 2011. Two search criteria were used:

one for disease state, and the other for intervention, linked

with the Boolean operator ‘‘AND’’; within those two

Figure 1. Results of literature search.

categories, each term was linked with ‘‘OR.’’ No limitation

terms were placed on the searches; in particular, no study

design or dates of publication limitations were used. A

medical librarian performed parallel searches. The search

was repeated in March 2012, July 2012, and January 2013,

just before abstract and manuscript submission, to ensure

that the most recent literature citations were being used in

the analysis. In addition, a separate search of the Cochrane

Library reviews was performed with the single search term

‘‘cholangiocarcinoma.’’

Initial exclusion criteria included abstracts that were a

case report or letter, did not involve treatment, did not

involve subjects with cholangiocarcinoma, or could not be

translated into English. All abstracts that passed this initial

screening process were reviewed. Publications were

included in the final analysis if they met all the following

criteria (ie, final inclusion criteria): primary data (ie, not a

review) presented from a prospective or retrospective study

or published abstract, outcomes or complications reported,

treatment involved locoregional chemotherapy-based trans-

arterial therapy (ie, transarterial chemotherapeutic infusion

or transarterial chemoembolization), data reported specif-

ically reported for subjects with cholangiocarcinoma if the

study involved additional subjects; and subjects had

unresectable cholangiocarcinoma.

Table 1 . Compilation of Articles Included in Metaanalysis (7–22)

Study, Year

No. of

Pts. Chemotherapeutic Dose per Treatment Embolization Agent

Procedures per

Patient*Median Survival (mo) 1-y Survival

(%)

Severe

Toxicities†From

Diagnosis

From

Procedure

Kuhlmann et al (7), 2012 36 Irinotecan 200 mg; mitomycin C 15 mg DEBs; Gelfoam 1.5 NA �10 41.6 11

Vogl et al (8), 2012 115 Mitomycin C 8 mg/m2, gemcitabine

1,000 mg/m2,

cisplatin 35 mg/m2

DEBs 7.1 NA 13 52 0

Park et al (9), 2011 72 Cisplatin 2 mg/kg Gelfoam 2.5 12.2 NA 51 34

Schiffman et al (10), 2011 24 Irinotecan 75 mg, doxorubicin 150 mg

(median doses)

DEBs 1.8 17.5 NA �68 4

Andrasina et al (11), 2010 17 Cisplatin 45 mg/m2, 5-FU 450 mg/m2 None 1 25.2 NA 88.24 NA

Chaiteerakij et al (12), 2010‡ 32 NA Yes; NS NA 10.66 NA 46.9 NA

Kiefer et al (13), 2010 62 Mitomycin C 10 mg, doxorubicin 50 mg,

cisplatin 100 mg

PVA 2.7 20 15 75 5

Harder et al (14), 2009‡ 13 Irinotecan 200 mg DEBs 1.6 NA NA NA 4

Poggi et al (15), 2009 9 Oxaliplatin 50 mg DEBs 3.3 NA 30 �70 11

Aliberti et al (16), 2008 11 Doxorubicin 75–100 mg DEBs 2.6 NA 13 �76 0

Gusani et al (17), 2008 42 Gemcitabine 1,250–2,250 mg/m2, cisplatin

100–125 mg/m2, oxaliplatin 85–100 mg/m2

Embosphere 3.5 NA 9.1 NA 7

Kim et al (18), 2008 49 Cisplatin 2 mg/kg Gelfoam 3 12 10 46 NA

Shitara et al (19), 2008 20 Mitomycin C 2–8 mg Degradable starch

microspheres

8 NA 14.1 �60 5

Herber et al (20), 2007 15 Mitomycin C 10 mg None 3.9 16.3y 16.3y 54.5 2

Burger et al (21), 2005 17 Cisplatin 100 mg, doxorubicin 50 mg,

mitomycin C 10 mg

PVA or Embosphere NA 23 NA NA 1

Kirchhoff et al (22), 2005 8 Cisplatin 50 mg/m2, doxorubicin 50 mg/m2 Degradable starch

microspheres

2 12y 12y NA 0

DEB ¼ drug-eluting bead, 5-FU ¼ 5-fluorouracil, NA ¼ not applicable, NCI ¼ National Cancer Institute, NS ¼ not specified, PVA ¼ polyvinyl alcohol, WHO ¼ World HealthOrganization.nMedian or mean.†Total NCI/WHO grade 4 3 toxicities from any cause.‡Study published only in abstract form.yDoes not specifically state if survival from time of diagnosis or time of procedure.

Ray

et

al’

JV

IR1220

’M

eta

an

aly

sis:T

ran

sarte

rial

Th

era

py

for

Ch

ola

ng

ioca

rcino

ma

Volume 24 ’ Number 8 ’ August ’ 2013 1221

The following data were recorded from each included

study: subject demographic data, study design, median

survival from the time of diagnosis and time of therapy,

Kaplan–Meier 1-year survival data, response categorized

per RECIST, length of time between therapy and postther-

apy imaging evaluation, number and type of toxicities of

treatment, and type of locoregional therapy, including dose,

embolization agent, and number of treatments per subject.

By using standard metaanalysis methodology, data were

pooled and analyzed. Because of a lack of comparative

arms in the published reports, tests for heterogeneity (I2 or

Cochran Q) were not performed.

Based on the nature of the present study, institutional

review board review was not sought.

RESULTS

The search criteria resulted in an initial 369 citations; 16 of

these studies, which included 542 subjects, fulfilled the

inclusion/exclusion criteria and are included in the meta-

analysis (Fig 1, Table 1) (7–22).

Direct comparisons between studies were limited by

differences in reporting survival (from time of diagnosis vs

from time of procedure) and differences in chemothera-

peutic and embolization agents used. Across the 16 studies,

subjects had similar Child–Pugh disease class (0% with

class C disease) and Eastern Cooperative Oncology Group

performance status (4 95% with performance status o 2).

For studies presenting such data, the weighted cumulative

median overall survival from the date of diagnosis was 15.7

months � 5.8 (Fig 2) (9–13,18,21); from the dates of

initiation of transarterial therapy, the weighted median overall

survival was 13.4 months � 6.7 (Fig 3) (7–13,15,16,18–20).

The overall weighted average for Kaplan–Meier-calculated

1-year survival was 58.0% � 14.5 (Fig 4). There was not

a substantial difference in survival between studies report-

ing these data. The single exception to this was the study by

Poggi et al (15), in which the median overall survival from

the time of the first procedure was 30 months (Fig 3).

Although this median far exceeds the remainder of the

reported survival times from the other studies used in this

analysis, the small number of subjects in this report (N ¼ 9)

had little influence on the overall average weighted survival

rate; the adjusted weighted median after excluding this report

only decreased from 13.4 months to 12.2 months.

The reported length of time between transarterial chemo-

embolization treatment and the postprocedure imaging eval-

uation was relatively consistent. Of the 10 studies that

reported such data, nine described initial imaging 1–3 months

following therapy (four studies reporting imaging at 3 mo,

three studies reporting imaging at 1–2 months, and two

studies providing a range of 1–3 months) (7,9,10,13,15,16,

18,19,21). The single remaining study (14) described initial

imaging performed 2–4 months after transarterial chemoembo-

lization. The vast majority of studies reported patients being

evaluated with computed tomography (CT) (9,10,15,18–20)

or with CT or magnetic resonance imaging (7,13,14,16). Only

one study (22) mentioned the use of ultrasound (in addition to

CT) to evaluate response.

Nearly one fourth of the reported subjects (22.8%) had a

complete or partial response on imaging based on RECIST,

whereas more than half (53.9%) had stable disease based on

the same criteria (Fig 5). The cumulative responses on

imaging follow-up per RECIST were complete response in

1.6%, partial response in 21.2%, stable disease in 53.9%, and

progressive disease in 23.2% (Fig 5). It is important to note

that the imaging response criteria used was largely RECIST,

not modified RECIST. The former criteria, which were used

in 10 of the 12 studies that reported imaging response

(modified RECIST was used in one (12), and a nonvalidated

scoring system was used in one [22]), reports response as a

change in diameter of the lesion, not the degree of residual

contrast enhancement. However, of the eight total patients

who exhibited a complete response, five were patients from

the one study that used modified RECIST (12). Extracting

these patients from the data set, and using only studies in

which RECIST were used, would provide a complete

response rate of only 0.8% (three of 399 patients).

In determining complication rates, 84 severe toxicities

(National Cancer Institute/World Health Organization gra-

de Z 3) were reported (18.9% of all subjects in those

studies that reported complications; Table 2). There

appeared to be a relatively even distribution of compli-

cation types, including hematologic and nonhematologic

complications. Interestingly, there were higher reported

serious complication rates in the more recent reports

(published in 2010 or later) compared with the earlier

reports (2009 or earlier; 17.1% vs 12.3%; P ¼ .16, w2

analysis). One observation that bears evaluation was the

relatively high severe complication rate when irinotecan was

used as the chemotherapy agent during transarterial

chemoembolization; two of the three studies in which

irinotecan was used (7,14) showed higher severe complica-

tion rates (30.8% and 30.6%) than the 18.8% average.

However, when all three studies in which irinotecan was

used (7,10,14) were pooled, there was no significant increase

in overall severe complication rate compared with studies in

which irinotecan was not used (20.9% vs 18.2%, respec-

tively; P = .673, w2 analysis).

There were four deaths within 30 days of transarterial

chemoembolization, corresponding to a reported mortality

rate of 0.7%. Many subjects experienced postembolization

syndrome; however, because of diverse reporting practices

by the various authors, a specific rate could not be

calculated. When reported, the incidences of postemboli-

zation syndrome ranged from 16% to 100%, but the

majority of studies subjectively described postembolization

syndrome occurring in ‘‘most’’ of their subjects.

For the studies that presented such data, there were a

total of 1,453 procedures performed cumulatively, averag-

ing 4.1 treatments per subject. However, this mean is

skewed by the presence of one very large study (N = 115)

(8) in which the average number of procedures performed

Figure 3. Forest plot of overall survival from the time of first treatment.

Figure 2. Forest plot of overall survival from the time of diagnosis. **Published only in abstract form.

Ray et al ’ JVIR1222 ’ Metaanalysis: Transarterial Therapy for Cholangiocarcinoma

was more than seven per subject. If this single study is

excluded, the mean number of procedures decreases to 2.7

per subject.

Only two studies did not use an embolization agent but

rather performed transarterial chemotherapeutic infusion

therapy (11,20). Indwelling port systems and intermittent

arterial access were used in these studies.

In all 16 studies, chemotherapy agents varied widely, and

included irinotecan, mitomycin C, gemcitabine, cisplatin,

doxorubicin, 5-fluorouracil, and oxaliplatin. Embolic agents

Figure 5. Cumulative weighted response per RECIST. RECIST = Response Evaluation Criteria In Solid Tumors.

Figure 4. Forest plot of 1-year overall survival by Kaplan–Meier method. **Published only in abstract form.

Volume 24 ’ Number 8 ’ August ’ 2013 1223

were also diverse, and included Gelfoam (Upjohn, Kala-

mazoo, Michigan), polyvinyl alcohol particles, Embosphere

particles (Biosphere, Rockland, Massachusetts), drug-elut-

ing beads, degradable starch microspheres, and Lipiodol

(Guerbet, Roissy, France; Table 1). In a separate analysis,

overall survival rates from the time of first treatment with

drug-eluting beads (with irinotecan and/or doxorubicin)

(7,10,14,16) was compared versus those in studies in which

drug-eluting beads were not used. There was no significant

difference in overall survival between those subjects treated

with drug-eluting beads (16.5 mo) and those not treated

with this technique (12.0 mo; P ¼ .41, Student t test).

Reported overall 1-year survival for the patients treated

with drug-eluting beads was also not significantly different

Table 2 . Toxicities following Intraarterial Therapies in Published Reports (7–22)

Study, Year Chemotherapeutic AgentsToxicities*

PES Complications

Severe Minor

Kuhlmann et al (7), 2012 Irinotecan, mitomycin 11 53 ‘‘Almost all’’ PES; 2 deaths from biliary sepsis

Vogl et al (8), 2012 Multiple groups (mitomycin C, gemcitabine,

and both, with/without cisplatin)

0 NA NA –

Park et al (9), 2011 Cisplatin 34 NA NA All severe: anemia (n ¼ 3), thrombocytopenia (n ¼ 6),

neutropenia (n ¼ 1), increased INR (n ¼ 1);

nonhematologic: increased AST (n ¼ 2,

increased ALT (n ¼ 1), increased ALP (n ¼ 1),

increased bilirubin (n ¼ 1), decreased albumin (n ¼ 1),

pain (n ¼ 3), nausea (n ¼ 1)

Schiffman et al (10), 2011 DEBIRI (n ¼ 35), DEBDOX (n ¼ 7) 4 7/42 4 (16.7) Atrial fibrillation (grade 2; n ¼ 2), hepatic insufficiency

(grade 3; n ¼ 2), and sepsis (grade 4; n ¼ 1), fatal HRS

(n ¼ 1), pneumonia (grade 2; n ¼ 1)

Andrasina et al (11), 2010 Cisplatin and 5-FU NA NA 13 (80) NA

Chaiteerakij et al (12), 2010† NA NA NA NA NA

Kiefer et al (13), 2010 Mitomycin C, doxorubicin, cisplatin 5 NA 107 (65) Pulmonary infarct (grade 2; n ¼ 1), pulmonary edema

(grade 4; n ¼ 1), elevated cardiac enzymes (n ¼ 1), ARF

(n ¼ 1), severe PES (n ¼ 1), hyperglycemia (n ¼ 1)

Harder et al (14), 2010† Irinotecan 4 NA 4 (31) RUQ pain (grade 3; n ¼ 4), RUQ pain (grade 1/2; n ¼ 6)

Poggi et al (15), 2009 Oxaliplatin 11 NA NA Abdominal pain (grade 3; n ¼ 9), hypertensive crisis

(grade 3; n ¼ 1), cholangitis (grade 3; n ¼ 1)

Aliberti et al (16), 2008 Doxorubicin 0 85 29 (100) Hepatic abscess (n ¼ 1), 95% grade 2 nausea/vomiting

(n ¼ 27), 100% neoplastic fever

Gusani et al (17), 2008 Gemcitabine (100%), cisplatin, oxaliplatin‡ 7 NA ‘‘Most’’ MI (grade 4; n ¼ 1), hepatic abscess (grade 4; n ¼ 1),

thrombocytopenia (n ¼ 1), sepsis (n ¼ 1)

Kim et al (18), 2008 Cisplatin NA NA ‘‘Most’’ Hepatic abscess (n ¼ 1), ‘‘most’’ patients with fever, nausea,

or vomiting (low grade)

Shitara et al (19), 2008 Mitomycin C 5 62/20 NA Duodenal ulcer (grade 3; n ¼ 3), RUQ pain (grade 3; n ¼ 2),

nausea (grade 3; n ¼ 1), anorexia (grade 3; n ¼ 1)

Herber et al (20), 2007 Mitomycin C 2 12/58 6 (40) Anaphylactic shock (grade 4; n ¼ 1), gastric ulceration (grade 4;

n ¼ 1), grade 1/2 nausea/vomiting (n ¼ 4), RUQ pain (n ¼ 6),

hepatic artery spasm (n ¼ 2)

Burger et al (21), 2005 Cisplatin, doxorubicin, mitomycin C 1 2/17 NA Fatal massive UGI bleeding (n ¼ 1)

Kirchhoff et al (22), 2005 Cisplatin, doxorubicin 0 NA NA Common nausea and fever

Values in parentheses are percentages where appropriate.ALP ¼ alkaline phosphatase, ALT ¼ alanine transaminase, ARF ¼ acute renal failure, AST ¼ aspartate transaminase, DEBDOX ¼ drug-eluting beads with doxorubicin, DEBIRI ¼drug-eluting beads with irinotecan, 5-FU ¼ 5-fluorouracil, HRS ¼ hepatorenal syndrome, INR ¼ International Normalized Ratio, MI ¼ myocardial infarction, NA ¼ not reported,NCI ¼ National Cancer Institute, PES ¼ postembolization syndrome, RUQ ¼ right upper quadrant, UGI ¼ upper gastrointestinal, WHO ¼ World Health Organization.nToxicities from any cause, events/subjects or procedure. Severe toxicities are NCI/WHO grade 4 3; minor are NCI/WHO grade o 3.†Study published only in abstract form.‡Gemcitabine was administered in all patients: with/without cisplatin (n ¼ 2) or with/without oxaliplatin (n ¼ 4), or gemcitabine with/without cisplatin in combination (n ¼ 14) orgemcitabine and cisplatin followed by oxaliplatin (n ¼ 4).

Ray

et

al’

JV

IR1224

’M

eta

an

aly

sis:T

ran

sarte

rial

Th

era

py

for

Ch

ola

ng

ioca

rcino

ma

Volume 24 ’ Number 8 ’ August ’ 2013 1225

than in those patients treated with standard transarterial

chemoembolization therapy (61.8% vs 59.7%, respectively;

P ¼ .339, Student t test).

DISCUSSION

In the present metaanalysis of 16 studies examining out-

comes following transarterial therapies in unresectable

cholangiocarcinoma, heterogeneity and small sample sizes

across studies made some comparisons difficult. However,

some important conclusions may be drawn.

The median survival from the time of diagnosis of

15.7 months (13.4 mo from time of treatment) following

transarterial therapies is generally higher than rates

reported with the use of only systemic treatments. Histor-

ically, most studies of single-agent chemotherapy regimens

report overall survival rates ranging from 5 to 8 months

(1,23–26). As always, however, one must be mindful of

selection bias when comparing specific disease interven-

tions with historical controls, as many of the subjects

represented in the historical studies may have been

excluded from transarterial therapies for various reasons.

Two recent studies (27,28) have reported improved sur-

vival rates for subjects undergoing combination systemic

chemotherapy (gemcitabine and cisplatin) in this popula-

tion. Although somewhat confusing because of how the

data were reported, by combining the results following

treatment of various forms of cholangiocarcinoma (eg,

intrahepatic, extrahepatic, and gallbladder carcinoma), both

these studies (27,28) demonstrated mean survival times

(11.7 mo and 13.0 mo) for intrahepatic and/or nongall-

bladder cholangiocarcinoma that are more favorable than

historical survival rates for the same disease processes.

Regarding tumor response per RECIST, more than three

fourths of all subjects showed a response or stable disease;

however, the majority of these subjects (53.9% of all

patients) exhibited stable disease by imaging criteria. It

is important to note that, with one exception (N ¼ 32), all

the studies included in this analysis used RECIST to eval-

uate response (12). With the increasing use of newer

imaging response criteria (eg, modified RECIST, European

Organization for Research and Treatment of Cancer), in

which enhancement patterns rather than overall size of the

lesion are measured, it is expected that imaging responses

will increase as a result of the use of these less stringent

criteria. Therefore, care should be taken when comparing

studies that use different imaging response criteria.

Complication rates were relatively high, with severe

complications (ie, National Cancer Institute/World Health

Organization grade 4 3) seen in a total of 18.9% of

patients. In addition, most studies reported minor toxicities

in the majority of subjects. These minor toxicities were

largely those associated with postembolization syndrome

(minor pain, fever, nausea/vomiting, fatigue), whereas the

major toxicities were highly variable but tended to be

nonhematologic. Finally, there was a very low overall 30-

day mortality rate (0.7%).

The limitations of the present study include insufficient

data to allow comparative conclusions to be drawn in terms

of efficacy based on type of chemotherapeutic agent(s)

used for transarterial therapies. In addition, many other

technical factors—such as the type of embolic agent

used and how selective the embolization procedures

were—potentially confound the results of this analysis as

a result of the substantial heterogeneity and lack of power

in the reported cases. Many of the studies did not report

whether their patients were treatment-naive or had received

chemotherapy previously. Of those studies that did report

on previous treatment (7,9–11,13,15,16,18–22), previous

exposure to therapy was highly variable (0%–100%), and

the data were typically not presented in such a way to

allow stratification of the groups. The fact that the included

studies were published over a relatively long period of time

(8 y) also adds to the potential for selection bias and

technical heterogeneity. Two of the studies (N ¼ 32 and

N ¼ 13) included in the final analysis (12,14) were

published in abstract form only, raising concerns that these

publications did not undergo a robust peer-review process.

The subjects included in the final analysis were from

manuscripts published around the world, including Asia,

Europe, and North America; this raises the concern that the

underlying disease process, techniques used in treatment,

indication for repeat treatment, and selection bias for

patients undergoing transarterial therapy may skew the

results, or perhaps make the overall results less general-

izable to any specific patient population. The lack of

standardization of reporting findings poses notable limi-

tations when evaluating the entire dataset. Finally, although

care was taken to identify all pertinent published articles, it

is possible that some published data were missed in the

initial and follow-up searches.

Although not included in the present metaanalysis, there

have been several recent studies evaluating selective internal

radiation therapy with yttrium-90 (90Y) in the treatment of

cholangiocarcinoma (29–31). These early data are promis-

ing, and demonstrate median survival rates from the time of

first 90Y treatment ranging from 9.3 to 22.0 months. As is

common with early studies, 90Y therapy was often used as

salvage therapy in these publications, with many subjects (as

many as 79%) having received previous systemic chemo-

therapy or presenting with extrahepatic disease (as many as

48%) (29,30). Further studies are needed to more fully

elucidate the role of selective internal radiation therapy in

the treatment of cholangiocarcinoma.

As demonstrated by the present metaanalysis, trans-

arterial chemotherapy–based treatments for cholangiocar-

cinoma appears to confer a survival benefit compared with

systemic therapies. This survival benefit appears to be most

substantial when compared versus single-agent systemic

chemotherapy, although the data are less compelling

for combination chemotherapy for intrahepatic cholangio-

carcinoma and nongallbladder cholangiocarcinoma. As

Ray et al ’ JVIR1226 ’ Metaanalysis: Transarterial Therapy for Cholangiocarcinoma

demonstrated by this metaanalysis, transarterial chemotherapy–

based treatments for cholangiocarcinoma demonstrated a

favorable response by imaging criteria, and has an accept-

able postprocedural complication profile. Such therapies

should be strongly considered in the treatment of patients

with this devastating illness.

REFERENCES

1. Khan SA, Thomas HC, Davidson BR, Taylor-Robinson SD Cholangiocar-

cinoma. Lancet 2005; 366:1303–1314. Erratum in: Lancet 2006;

367:1656.

2. Zechlinski JJ, Rilling WS. Transarterial therapies for the treatment of

intrahepatic cholangiocarcinoma. Semin Intervent Radiol 2013; 30:21–27.

3. Lo CM, Ngan H, Tso WK, et al. Randomized controlled trial of trans-

arterial lipiodol chemoembolization for unresectable hepatocellular carci-

noma. Hepatology 2002; 35:1164–1171.

4. Llovet JM, Real MI, Montana X, et al. Arterial embolization or chemo-

embolization versus symptomatic treatment in patients with unresect-

able hepatocellular carcinoma: a randomized controlled trial. Lancet 2002;

359:1734–1739.

5. Marelli L, Stigliano R, Triantos C, et al. Transarterial therapy for

hepatocellular carcinoma: which technique is more effective? A system-

atic review of cohort and randomized studies. Cardiovasc Intervent

Radiol 2007; 30:6–25.

6. Patt YZ, Hassan MM, Aguayo A, et al. Oral capecitabine for the

treatment of hepatocellular carcinoma, cholangiocarcinoma, and gallblad-

der carcinoma. Cancer 2004; 101:578–586.

7. Kuhlmann JB, Euringer W, Spangenberg HC, et al. Treatment of unre-

sectable cholangiocarcinoma: conventional transarterial chemoembolization

compared with drug eluting bead-transarterial chemoembolization and

systemic chemotherapy. Eur J Gastroenterol Hepatol 2012; 24:437–443.

8. Vogl TJ, Naguid NNN, Nour-Eldin NEA, et al. Transarterial chemo-

embolization in the treatment of patients with unresectable cholangio-

carcinoma: results and prognostic factors governing treatment success.

Int J Cancer 2011; 131:733–740.

9. Park SY, Kim JH, Yoon HJ, Lee IS, Yoon HK, Kim KP. Transarterial

chemoembolization versus supportive therapy in the palliative treatment

of unresectable intrahepatic cholangiocarcinoma. Clin Radiol 2011; 66:

322–328.

10. Schiffman SC, Metzger T, Dubel G, et al. Precision hepatic arterial

irinotecan therapy in the treatment of unresectable intrahepatic chol-

angiocellular carcinoma: optimal tolerance and prolonged overall survival.

Ann Surg Oncol 2011; 18:431–438.

11. Andrasina T, Valek V, Panek J, et al. Multimodal oncological therapy

comprising stents, brachytherapy, and regional chemotherapy for chol-

angiocarcinoma. Gut Liver 2012; 4(suppl):S82–S88.

12. Chaiteerakij R, Schmit G, Mettler TA, Andrews J, Roberts LR. Trans-

arterial chemoembolization (TACE) improved survival in unresectable

intrahepatic cholangiocarcinoma. 61st Annual Meeting of the American

Association for the Study of Liver Diseases (abstr.). Hepatology 2010;

52(suppl):1137A.

13. Kiefer MV, Albert M, McNally M, et al. Chemoembolization of intra-

hepatic cholangiocarcinoma with cisplatinum, doxorubicin, mitomycin C,

Ethiodol, and polyvinyl alcohol: a 2-center study. Cancer 2011; 117:

1498–1505.

14. Harder J, Euringer W, Langer M, Thimme R, Blum HE,

Spangenberg HC. Transarterial chemoembolization of intrahepatic

cholangiocarcinomas with irinotecan-eluting beads: preliminary results.

60th Annual Meeting of the American Association for the Study of Liver

Diseases (abstr.). Hepatology 2009; 50(suppl 4):1116A.

15. Poggi G, Amatu A, Montagna B, et al. OEM-TACE: a new therapeutic

approach in unresectable intrahepatic cholangiocarcinoma. Cardiovasc

Intervent Radiol 2009; 32:1187–1192.

16. Aliberti C, Benea G, Massimo T, Giorentini G. Chemoembolization

(TACE) of unresectable intrahepatic cholangiocarcinoma with slow-

release doxorubicin-eluting beads: preliminary results. Cardiovasc Inter-

vent Radiol 2008; 31:883–888.

17. Gusani NJ, Balaa FK, Steel AL, et al. Treatment of unresectable

cholangiocarcinoma with gemcitabine-based transcatheter arterial che-

moembolization (TACE): a single institution experience. J Gastrointest

Surg 2008; 12:129–137.

18. Kim JH, Yoon HK, Sung KB, et al. Transcatheter arterial chemoembo-

lization or chemoinfusion for unresectable intrahepatic cholangiocarci-

noma: clinical efficacy and factors influencing outcomes. Cancer 2008;

113:1614–1622.

19. Shitara K, Ikami I, Munakata M, Muto O, Sakata Y. Hepatic arterial

infusion of mitomycin C with degradable starch microspheres for unre-

sectable intrahepatic cholangiocarcinoma. Clin Oncol 2008; 20:241–246.

20. Herber S, Otto G, Schnieder J, et al. Transarterial chemoembolization

(TACE) for inoperable intrahepatic cholangiocarcinoma. Cardiovasc Inter-

vent Radiol 2007; 30:1156–1165.

21. Burger I, Hong K, Schulick R, et al. Transcatheter arterial chemo-

embolization in unresectable cholangiocarcinoma: initial experience in a

single institution. J Vasc Interv Radiol 2005; 16:353–361.

22. Kirchhoff T, Zender L, Merkesdal S, et al. Initial experience from a

combination of systemic and regional chemotherapy in the treatment of

patients with nonresectable cholangiocellular carcinoma in the liver.

World J Gastroenterol 2005; 11:1091–1095.

23. Alberts SR, Gores GJ, Kim GP, et al. Treatment options for hepatobili-

ary and pancreatic cancer. Mayo Clin Proc 2007; 82:628–637.

24. Mazhar D, Stebbing J, Bower M. Chemotherapy for advanced chol-

angiocarcinoma: what is standard treatment? Future Oncol 2006; 2:

509–514.

25. Okusaka T, Ishii H, Funakoshi A, et al. Phase II study of single-agent

gemcitabine in patients with advanced biliary tract cancer. Cancer

Chemother Pharmacol 2006; 57:647–653.

26. Dingle BH, Rumble RB, Brouwers MC. The role of gemcitabine in the

treatment of cholangiocarcinoma and gallbladder cancer: a systematic

review. Can J Gastroenterol 2005; 19:711–716.

27. Valle JW, Wasan H, Johnson P, et al. Gemcitabine alone or in combina-

tion with cisplatin in patients with advanced or metastatic cholangiocarci-

nomas or other biliary tract tumours: a multicentre randomised phase II

study – the UK ABC-01 Study. Br J Cancer 2009; 101:621–627.

28. Okusaka T, Nakachi K, Fukutomi A, et al. Gemcitabine alone or in

combination with cisplatin in patients with biliary tract cancer: a com-

parative multicentre study in Japan. Br J Cancer 2010; 103:469–474.

29. Hoffmann RT, Paprottka PM, Schon A, et al. Transarterial hepatic

yttrium-90 radioembolization in patients with unresectable intrahepatic

cholangiocarcinoma: factors associated with prolonged survival. Cardio-

vasc Intervent Radiol 2012; 35:105–116.

30. Saxena A, Bester L, Chua TC, Chu FC, Morris DL. Yttrium-90 radio-

therapy for unresectable intrahepatic cholangiocarcinoma: a preliminary

assessment of this novel treatment option. Ann Surg Oncol 2010; 17:

484–491.

31. Ibrahim SM, Mulcahy MF, Lewandowski RJ, et al. Treatment of

unresectable cholangiocarcinoma using yttrium-90 microspheres: results

from a pilot study. Cancer 2008; 113:2119–2128.