Defining the Role of Hepatic Arterial Infusion Chemotherapy in Metastatic Colorectal Cancer

Defining the Role of Hepatic Arterial Infusion Chemotherapy in Metastatic Colorectal Cancer

ABSTRACT: The use of hepatic arterial infusion (HAI) chemotherapy in patients with liver-only colorectal metastases is based on the pharmacologic principle that the regional administration of some drugs can lead to higher drug concentrations at the site of the metastases and avoid systemic toxicity. Early randomized trials resulted in high response rates but did not lead to a survival advantage with HAI. More recent trials have utilized improved surgical techniques and strict guidelines regarding dose reduction or cessation of HAI chemotherapy, resulting in a significant reduction in toxicity. In patients with unresectable liver metastases, two recent European trials using HAI fluorouracil (5-FU) again failed to demonstrate an improvement in survival, but both were plagued by a high complication rate with an associated high proportion of patients failing to receive the assigned treatment. In contrast, the preliminary results of a recent Cancer and Leukemia Group B trial did demonstrate a survival advantage with HAI floxuridine when compared to systemically administered 5-FU. Trials investigating the use of HAI chemotherapy in the adjuvant setting have yielded mixed results. Moreover, in light of improved response rates and overall survival with newer more active chemotherapeutic and novel agents, the absolute role of HAI chemotherapy remains undefined.

In the United States, approximately
140,000 people will be diagnosed
with colorectal cancer in
2004, with one-third ultimately dying
of metastases from that disease. Most
metastases occur in the liver, and
around 80% of patients with hepatic
metastases will die of progressive liver

In most instances, colorectal cancer
metastasizes to multiple sites.
However, metastases may be limited
to the liver in a significant subset of
patients. This is evidenced by longterm
disease-free survival in many
patients who undergo resection of isolated
hepatic metastases. Unfortunately,
although one-third of patients with
hepatic metastases have disease confined
to the liver, resection is possible
in only 10% of these patients due to
extensive infiltrative disease. Of this
small group, roughly 30% will realize
long-term survival with resection.[1]

Although normal hepatocytes receive
their blood supply from the portal
circulation, liver metastases receive
80% of their blood supply via the
hepatic artery.[2] The recognition of
this feature of hepatic tumors led to
regional delivery of chemotherapy by
way of the hepatic artery, preferentially
delivering drug to the tumor
while sparing normal hepatocytes.

Extensive investigation of hepatic
arterial infusion (HAI) chemotherapy
began in the 1970s using percutaneously
placed catheters. With the introduction
of a totally implantable
pump in the 1980s,[3] further investigation
suggested a marked improvement
in response rates but failed to
show a significant survival benefit.
Techniques employed in pump placement
and administration of chemotherapy
were then further refined in
an effort to improve safety and optimize
response and survival.

Not every chemotherapeutic agent
has the pharmacologic properties or
activity in colorectal cancer to justify
its HAI use. Floxuridine (FUDR) has
become the agent of choice because
of a relatively high hepatic-to-systemic
ratio (400) and high rate of hepatic
extraction (94% to 99%). Fluorouracil
(5-FU) has also been commonly used,
although it has a lower hepatic-tosystemic
ratio (10 to 100) and rate of
hepatic extraction (20% to 55%).[4,5]
Additionally, the relative insolubility
of that 5-FU precludes administration
using an implantable pump (since
it must be diluted in a relatively large
volume), requiring the use of a port
or percutaneously placed catheter

Randomized studies have employed
HAI chemotherapy as primary
therapy in patients with unresectable
hepatic metastases as well as
in the adjuvant setting after resection
of liver metastases. This article will
summarize these studies and address
current issues related to the practical
administration and future role of HAI
chemotherapy in an era of improved
efficacy and survival with newer chemotherapeutic
and novel agents.

Primary Treatment for
Unresectable Liver Metastases

In the early 1980s, with only 5-FU
or its analogs available, four major
prospective trials involving 375 patients
compared systemic fluoropyrimidine
chemotherapy with HAI
FUDR. Although the response rate was consistently superior in the HAI
arm of all four trials, none was able to
demonstrate a statistically significant
survival advantage.[6-9] Two European
trials in the late 1980s randomized
patients with unresectable liver
metastases to receive HAI FUDR or
best supportive care. Both trials demonstrated
an improved survival advantage
among patients treated with
HAI chemotherapy. However, most
patients in the control arms did not
receive systemic chemotherapy, which
would be considered suboptimal treatment
by today's standards (see Table

Two meta-analyses have been performed
to address the possibility that
the lack of a survival benefit in these
studies reflected sample size limitations
rather than treatment failure.[
12,13] Both confirmed the
superior response rate with HAI chemotherapy
compared to systemic chemotherapy.
However, while there was
a trend toward improved survival in
the HAI arms, this survival benefit
only reached statistical significance
when including those patients who
received no treatment in the control

Data from these trials and the subsequent
meta-analyses consistently
demonstrate a significant improvement
in response with HAI chemotherapy.
However, this improvement
in response rate also consistently failed
to translate into a significant survival
advantage. Although several factors
may be responsible for this apparent
disconnect, surgical issues associated
with pump placement and hepatobiliary
toxicity associated with chemotherapy
administration were found to
be significant contributors to HAI
treatment's failure to achieve a survival

Refinement of Surgical Technique
and Chemotherapy Delivery

The goal of pump placement is
twofold: (1) to achieve complete hepatic
perfusion of chemotherapy, and
(2) to prevent misperfusion of chemotherapy
to the stomach or duodenum
with its resultant inflammation
and ulcer formation. Initial studies
suggested a relatively high incidence
of misperfusion and lack of complete
hepatic perfusion. This problem was
addressed with an attempt to standardize
pump placement technique in
a manner that would minimize these

Several safeguards are now routinely
implemented to ensure complete
hepatic arterial patency and
perfusion and to exclude misperfusion.
Preoperative angiography is performed
to evaluate anomalous hepatic
perfusion and confirm portal vein patency.
Angiography is now followed
by exploratory laporatomy to exclude
extrahepatic disease with elective
cholecystectomy to prevent postoperative
chemical cholecystitis. Intraoperatively,
the pump port is injected
with fluorescein and the liver, stomach,
and duodenum are examined using
a Wood's lamp to ensure
appropriate perfusion. Finally, prior
to using the pump, radionuclide angiography
(with technetium macroaggregated
albumin) is employed
through the pump catheter, again to
ensure adequate hepatic perfusion and
exclude misperfusion.

In addition to standardization of
surgical techniques, different chemotherapeutic
regimens have been employed
to minimize drug-induced
hepatic toxicity. The dose of FUDR
used in most of the above studies (0.3
mg/kg/d) resulted in a high proportion
of patients experiencing permanent
biliary damage. Further phase II
studies explored different dosing
schedules as well as using dexamethasone
and/or leucovorin to minimize
toxicity or increase efficacy.
Combining HAI chemotherapy with
systemic chemotherapy has also been
addressed with similar goals.

HAI Therapy:
Maximizing FUDR?
One phase II trial conducted at
University of California, San Francisco
(UCSF), used a much lower infusional
FUDR dose alternating with
HAI bolus 5-FU, seeking to take advantage
of the differing toxicities and
pharmacokinetics of these drugs. In
none of the 64 patients treated was
administration of this regimen limited
by hepatobiliary toxicity. In this
group, where 30 had not responded to
prior systemic chemotherapy, the HAI
regimen described resulted in a 50%
objective response rate with a median
survival of 22.4 months. However,
progressive liver tumor was the initial
site of failure and cause of death in
the majority of patients.[17]

A series of studies performed at
Memorial Sloan-Kettering Cancer
Center (MSKCC) sought to better define
the optimal administration of HAI
FUDR. One randomized phase II
study attempted to minimize biliary
inflammation with its associated hepatobiliary
toxicity by administering
dexamethasone with FUDR. This
study failed to show a decrease in
toxicity or increase in drug delivery,
but surprisingly suggested an improvement
in response rate and median
survival in the dexamethasone arm
compared to FUDR alone.[18] A similar
attempt to increase efficacy and
reduce toxicity included varying doses
of leucovorin in combination with
varying doses of FUDR. Biliary sclerosis
remained high, but again, longterm
survival was better than

Ultimately, MSKCC created a hybrid
regimen combining FUDR, dexamethasone,
and leucovorin in an
effort to both reduce toxicity and maximize
response. In phase II trials, the
response rate using this regimen in
previously untreated patients was 78%
with a median survival of 24.8 months.
Using strict criteria for dose reduction
or discontinuation of treatment,
the incidence of biliary sclerosis was
also reduced to 3%.[20]

Employing these techniques, surgical
complications and chemotherapeutic
toxicity have improved
markedly. Prior to these refinements,
the incidence of biliary sclerosis
ranged from 10% to 40% with an incidence
of ulcer formation or gastritis
frequently over 20%. With the safeguards
described above, the incidence
of biliary sclerosis is now less than
5% with a similar reduction in ulcer
formation or gastritis. In addition to
the above-mentioned surgical safeguards,
European investigators have
employed 5-FU rather than FUDR
with similarly stringent guidelines regarding
dose reduction and discontinuation
of treatment. This has
resulted in similar improvement in
treatment-related toxicity, although
the use of external catheters and ports
results in a greater incidence of catheter
occlusion and hepatic artery thrombosis.

Recent Studies
Two large European studies and
one United States study have recently
been reported using HAI chemotherapy
as the primary treatment modality
for patients with unresectable
colorectal hepatic metastases (see
Table 2).

The German Cooperative Group
on Liver Metastases published results
from a multicenter randomized study
where 168 patients were assigned to
one of three arms: (1) HAI 5-FU plus
leucovorin, (2) IV 5-FU plus leucovorin,
or (3) HAI FUDR. In terms of
response and survival, there were no
statistically significant differences
among the three treatment arms. Intrahepatic
response was similar in the
two HAI treatment arms (45% with
5-FU, 43% with FUDR) and, as expected
from previous trials, lower in
the IV 5-FU arm (20%). On the other
hand, median survival was similar in
the HAI and IV 5-FU arms (18.7 and
17.6 months, respectively), with median
survival in the HAI FUDR arm
lower at 12.7 months (not statistically

The Medical Research Council and
European Organization for Research
and Treatment of Cancer (MRC/
EORTC) colorectal cancer study
groups randomized 290 patients to
receive either HAI 5-FU plus leucovorin
or systemic 5-FU plus leucovorin.
In contrast to previous studies,
response rates were similar between
the HAI and IV arms (22% vs 19%).
Median survival also did not differ
between the two groups (14.7 vs 14.8
months for the HAI and IV groups,

Both European trials used 5-FU as
the chemotherapeutic agent delivered
via the hepatic artery. The reasons
cited for choosing 5-FU were twofold:
(1) to reduce hepatobiliary toxicity
associated with FUDR; and (2)
to take advantage of 5-FU's lower
hepatic extraction rate with its associated
systemic spillover, ostensibly
treating systemic micrometastatic disease.
Because of relative insolubility
and the associated necessary large
volume, administering HAI 5-FU requires
using percutaneously placed
catheters or ports rather than implantable
pumps. In multiple reviews, ports
have been associated with a significantly
higher complication rate compared
to implantable pumps.[23,24]
Additionally, without the use of continuous
heparin infusion as is used
with implantable pumps, the use of
ports or catheters has resulted in a
higher incidence of hepatic artery

In the German study, 32% of patients
assigned to the HAI arms failed
to receive their assigned therapy. Similarly,
the MRC/EORTC trial authors
point out that 37% of patients assigned
to the HAI arm did not receive their
designated treatment in this trial. Additionally,
only two cycles were given
on average in the HAI arm of this
trial because of the associated toxicity.
This high rate of patients failing to
receive their assigned treatment has
been cited as a criticism regarding the
reported response rates and survival
data in these two trials.

The first trial to demonstrate a survival
advantage with HAI compared
to systemic chemotherapy in patients
with unresectable liver metastases was
presented at the American Society of
Clinical Oncology (ASCO) Annual
Meeting in 2003. This multi-institution
Cancer and Leukemia Group B
(CALGB) trial randomized 135 patients
to receive either HAI FUDR
plus leucovorin plus dexamethasone
or systemic 5-FU plus leucovorin.
Median survival was superior in the
HAI group compared to the systemic
chemotherapy group (22.7 vs 19.8
months, P = .027). Similar to previous
trials, superiority with HAI chemotherapy
was also demonstrated in
control of hepatic metastatic disease
and overall response rate (48% vs
25%, P = .009). Additionally, preliminary
quality-of-life data suggest improvement
in the HAI group at 1 year
compared to the control arm.[25]

This CALGB study is encouraging
in that it appears that with the safeguards
mentioned above, investigators
are able to reduce toxicity, thus
demonstrating a statistically significant
survival benefit and an improvement
in quality of life when comparing
HAI FUDR to systemic 5-FU. As
mentioned above, the two European
studies used 5-FU with ports and both
studies had a large number of patients
not receiving their assigned treatment.
Although cited as a criticism, this observation
stands out demonstrating that even with highly monitored chemotherapy
administration, this modality
of treatment continues to be
plagued by issues regarding its practical
incorporation into standard care.

Alternative Strategies
In addition to modifying the dosing
regimen of FUDR and adding steroids
and/or leucovorin, an alternative
strategy has been to alternate HAI
chemotherapy with systemic chemotherapy.
When using strict administration
and dosing guidelines as
mentioned above, this approach would
in theory achieve high local response
rates with reduced hepatobiliary complications
while simultaneously treating
extrahepatic micrometastases, thus
decreasing the risk for extrahepatic
disease progression. A number of
smaller phase II studies have investigated
this concept utilizing a variety
of chemotherapy schedules, most using
systemic FUDR or 5-FU.[26,27].
To date, there have been mixed results,
some suggesting improved tolerance
while others are complicated
by significant systemic and hepatobiliary
toxicity. Currently, systemic
oxaliplatin (Eloxatin) and/or irinotecan
(Camptosar) are being added to
HAI FUDR regimens to achieve these
goals.[28,29] This theory is yet to be
evaluated in a randomized phase III

HAI as Primary Therapy
in Setting of New
Chemotherapeutic Options

The above-mentioned CALGB trial
demonstrating improved survival
with HAI chemotherapy compared
with systemic 5-FU plus leucovorin
began enrollment in 1994. Since that
time, newer chemotherapeutic agents
have demonstrated increased activity
with resultant improvement in survival
in the setting of metastatic
colorectal cancer.

Currently, with the availability of
irinotecan, oxaliplatin, and 5-FU, median
survival for all patients with metastatic
colorectal cancer approaches
or exceeds 20 months.[30] No randomized
trials have been performed
comparing HAI therapy with these
newer agents. When utilizing these
newer, more active agents, survival is
similar to the median survival demonstrated
in studies using HAI chemotherapy
as the primary therapeutic
modality. This brings into question
any survival benefit with HAI compared
to today's standard systemic
chemotherapy options. Current investigation
into HAI chemotherapy in the
setting of unresectable hepatic metastases
involves combining this modality
with systemic chemotherapy
utilizing these newer, more active
agents,[27] as well as investigating
these drugs as the primary chemotherapy
agent delivered by the hepatic
artery pump.[31,32]

HAI in the Preoperative Setting

With the remarkably high response
rate to HAI chemotherapy, one obvious
consideration is to implement this
approach in the neoadjuvant setting.
One retrospective study at the M. D.
Anderson Cancer Center attempted to
address this question. Over 15 years,
383 patients with unresectable liver
metastases from colorectal cancer
were treated with HAI chemotherapy.
A total of 22 (0.6%) patients underwent
exploratory surgery, of which
18 received some form of treatment
(10 underwent resection, 6 resection
and radiofrequency ablation [RFA] or
cryotherapy, and 2 RFA alone). At a
median follow-up of 17 months, 15
(83%) of the 18 patients who had received
treatment had developed recurrent
disease. The other three
patients had died of other causes within
7 months of surgery.[33] Similarly,
of over 300 patients at UCSF
treated with HAI chemotherapy, only
3 have been explored for surgical resection,
without any long-term survivors.[

There are no prospective randomized
trials using HAI chemotherapy
in the neoadjuvant setting. However,
data such as mentioned above have
not been encouraging. Lacking reliable
clinical benefit, administering
HAI chemotherapy with a neoadjuvant
intent should be used cautiously.
Furthermore, the biliary toxicity associated
with HAI chemotherapy may
diminish hepatic reserve and render
patients less able to undergo further
hepatic resections.


Loading comments...
Please Wait 20 seconds or click here to close