Annals of Oncology II: 1563-1570.2000.
S 2000 Kluwer Academic Publishers. Printed in the Netherlands.
Original article
Hepatic arterial 5-fluorouracil in patients with liver metastases of
colorectal cancer: Single-centre experience in 145 patients
J. M. G. H. van Riel,1 C. J. van Groeningen,1 S. H. M. Albers,1 M. Cazemier,1 S. Meijer,2
R. Bleichrodt,2 F. G. van den Berg,3 H. M. Pinedo1 & G. Giaccone1
Departments of ^Medical Oncology, 2Surgtcal Oncology, 3Radiology, University Hospital Vrije Universtteit, Amsterdam, The Netherlands
shorter compared to patients without extrahepatic disease or
5-FU-based pretreatment (9.7 vs. 19.3 months and 10.1 vs. 17.4
Background: Hepatic arterial chemotherapy for liver metastases months, respectively), forty-seven percent of patients stopped
of colorectal cancer is still under discussion. Mainly because of treatment because of a complication. Complications most
the technical complications of this mode of treatment and the often seen in patients with arterial ports were hepatic artery
lack of a survival benefit in randomized studies. We performed thrombosis (48%) and dislocation of the catheter (22%).
an analysis of hepatic arterial 5-fluorouracil (5-FU) chemoConclusions: The results of our analysis are in line with
therapy in 145 consecutive patients treated at a single institu- previous phase III studies. Extrahepatic disease and i.v. 5-FUtion.
based pretreatment were prognostic for reduced OS. The
Patients and methods: One hundred forty-five patients with complication rate of hepatic arterial delivery was worrisome,
inoperable liver metastases from colorectal cancer were in- although, no negative impact on survival could be established.
cluded. 5-FU, 1000 mg/m2/day continuous infusion for five There is a strong need for improvement of hepatic arterial
days every three weeks, was delivered in the hepatic artery by delivery methods before further evaluation of hepatic arterial
5-FU will be worthwhile.
percutaneous catheter or arterial access device.
Results: The response rate was 34% for all patients, 40% in
patients with extrahepatic disease, and 15% in patients with i.v.
5-FU-based pretreatment. TTP and OS for all patients were 7.5 Key words: 5-fluorouracil, arterial access device, chemotherapy,
and 14.3 months, respectively. In patients with extrahepatic colorectal cancer, hepatic arterial chemotherapy, liver metasdisease or i.v. 5-FU-based pretreatment, OS was significantly tases, port-a-cath
Summary
Introduction
The liver is the major site of metastasis in colorectal
cancer. Over 60% of patients with metastatic colorectal
cancer have disease localized in the liver, and the liver is
the only metastatic site in approximately half of these
patients. The one- and three-year survivals of untreated
patients with liver metastases are poor, 31% and 2.6%,
respectively [1]. Surgical resection, which is the only
curative treatment, results in a five-year survival of
20%-40% [2]. However, in only 10%-20% of patients
resection is possible.
In patients with irresectable disease the use of hepatic
arterial chemotherapy is a therapeutic possibility which
has been investigated for many years. The rationale for
hepatic arterial chemotherapy is based on the fact that
liver metastases derive most of their blood supply (more
than 80%) from the hepatic artery and on the high first
pass hepatic extraction of the drugs used for this
approach, i.e., floxuridine (FUDR) and 5-fluorouracil
(5-FU) [3, 4]. Both factors facilitate high local drug
concentrations with reduced systemic toxicity, allowing
treatment with relatively high dosages compared to
intravenous treatment. Phase III studies comparing intravenous 5-FU with hepatic arterial FUDR or 5-FU,
showed significantly higher response rates with hepatic
arterial treatment than with intravenous treatment (40%60% vs. 10%-20%, respectively) [5-10]. Disappointingly,
this has not been reflected in a significant survival
advantage even when meta-analyses were performed
[11, 12]. Many reasons for this discrepancy have been
proposed: 1) The number of patients included in the
studies was small, illustrated by the low total number of
patients (391 and 579) included in the two meta-analyses.
2) A high number of randomized patients included in
the studies (intention-to-treat) failed to receive hepatic
arterial treatment because of extra-hepatic disease
detected after randomization, or due to technical
complications of the procedure. 3) Some studies allowed
patients who progressed on intravenous treatment to
cross over to hepatic arterial treatment. 4) And finally,
extrahepatic disease progression was the reason of treatment failure in many patients.
Because hepatic arterial chemotherapy is gaining
popularity as adjuvant treatment after surgical resection
[13, 14], and neoadjuvant treatment is also being con-
1564
sidered, the hazards of this treatment modality need to
be weighted against the potential therapeutic benefit.
Here we report our experience of hepatic arterial chemotherapy in patients with liver metastases from colorectal
cancer. Over a period of almost 17 years, a total of
145 patients received hepatic arterial 5-FU, to our
knowledge the largest single institution series.
Patients and methods
Inclusion criteria
The patients included in this analysis were identified from the hospital
database by screening for patients who had liver metastases from
colorectal cancer and received hepatic arterial treatment with 5-FU
between November 1982 and April 1999. Patients included had histologically documented colorectal cancer with liver metastases only, not
amenable to surgery, or minimal extrahepatic disease.
Table 1 Patient characteristics.
Age (years)
Median
Range
60
34-83
Sex
Male
Female
Site of primary tumor
Colon
Rectum
Liver metastases
Synchronous
Metachronous
Primary tumor - time of resection
Before start
During treatment
Site of metastases
Liver alone
Liver and other sites
Previous i.v. 5-FU-based therapy
Yes
No
81
64
108
37
92
53
134
11
117
28
32
113
Hepatic arterial infusion
Hepatic arterial chemotherapy was delivered by an arterial access
system (arterial port) connected with a portable pump or a percutaneous catheter with a bedside pump. The choice between the percutaneous catheter and the arterial access system at the beginning of
treatment was mainly based on the degree of liver involvement, the
presence of extrahepatic disease, and the performance status Patients
with extensive liver metastases, known extrahepatic disease and poor
performance score, started treatment with a percutaneous catheter.
Subsequently, patients with stable or responding disease were considered for access system implantation. For the placement of the arterial
access system, patients underwent a laparotomy and the catheter was
positioned in the ligated gastroduodenal artery with the catheter tip
located at the junction of the gastroduodenal and hepatic artery.
Preoperative angiograms were made to assess hepatic arterial anatomy
prior to system implantation. A cholecystectomy was routinely performed to prevent potential chemical cholecystitis during chemotherapy. The port was placed subculaneously on the left ventral chest
wall. Shortly after surgery, a technetium-99 macroaggregate albumin
(MAA) scan of the liver was performed via the port to ensure adequate
perfusion of the entire liver. The percutaneous catheter was angiographically placed with the tip located preferentially in the common
hepatic artery after the branching of the gastroduodenal artery. Hospitalization and immobilization were required during each infusion
period and during immobilization patients received prophylactic subcutaneous heparin to prevent deep venous thrombosis.
defined as a ^ 25% increase in any measurable lesion or the appearance of a new lesion. Stable disease was defined as < 25% increase and
< 50% decrease of the lesion. In patients who had also disease outside
the liver, all measurable sites were included in the evaluation.
Time to progression was defined as the interval from the onset of
treatment (i.e, placement of the arterial port or catheter) until disease
progression or death. Thefindingof extrahepatic disease at the time of
arterial access device placement, in patients who had arterial chemotherapy by percutaneous catheter, was also regarded as progression.
Overall survival was defined as the interval from the onset of treatment
until death if no progression was observed.
Statistical analysis
The median follow-up for all patients was 57 months. We performed
response evaluation both on intention to treat bases as well as in
patients with evaluable disease who received at least two cycles of
chemotherapy. Time to progression and overall survival were estimated
by Kaplan-Meier survival curves. In eight patients we were not able to
define the time of progression and in eight patients the time of death
was not known. These patients were censored and the date of their last
visit was used. For comparison between survival curves the log rank
test was performed.
Chemotherapy
All patients received hepatic arterial 5-FU by continuous infusion.
Patients with a percutaneous catheter received 1000 mg/m2/day for five
days every three weeks and patients with an arterial access system
received an equivalent total dose over six instead of five days every
three weeks. If treatment was tolerated well, the next cycle the dose was
escalated by 200 mg/m2/day.
Study parameters
Before start of treatment all patients had a complete history taking, a
physical examination, and laboratory tests (hematology. liver function,
renal function). Whenever indicated other tests were performed. Evaluation was performed every two to three cycles by CT scan or ultrasound of the liver, while also a chestfilmwas made. Complete response
was defined as the disappearance of all tumor. Partial response was
defined as a ^ 50% decrease of the lesions, and progressive disease was
Results
Patient characteristics
From November 1982 to April 1999, 145 patients with
liver metastases from colorectal cancer received hepatic
arterial 5-FU. The patient characteristics are depicted in
table 1. Ninety-two patients had synchronous and 53
patients had metachronous liver metastases. At the start
of treatment 28 patients presented also extrahepatic
disease, and 32 patients had received prior i.v. 5-FUbased treatment for metastatic disease. The primary
tumor was resected before start of chemotherapy in 134
patients. In 17 of these 134 patients the arterial access
system was implanted during surgery for the primary
1565
Table 2. Response rates in all patients and subgroups.
Response
No pretreatment and
hepatic disease only
All patients
Extrahepatic disease
1.1.1.
CR
PR
SD
PD
NE
(n = 127)
(%)
(« = 145)
(%)
Evaluable
(n = 84)
I.T.T.
(n = 95)
3
31
51
15
3
27
45
13
12
4
32
52
12
3
28
46
11
12
Prior 5-FU-based therapy
Evaluable
(« = 25)
(%)
I.T.T.
(n =28)
(%)
Evaluable
(n = 27)
(%)
4
36
40
20
3
32
36
18
11
15
59
26
0
1.1.1.
(n = 32)
(%)
0
12
50
22
16
Abbreviations: I.T.T. - intention to treat: CR - complete response, PR - partial response. SD - stable disease, PD - progressive disease, NE
not evaluable.
145 patients
[EHD 28 (19%)]
[pretreated 32 (22%)]
52 patients
percutaneous catheter
[EHD 19 (37%)]
[pretreated 17 (33%)]
93 patients
Arterial port
[EHD 9 (9%)]
[pretreated 15 (15%)]
20 patients
Arterial port
[EHD 4 (20%)]
[pretreated 4 (20%)]
Figure 1. Treatment flow chart. Abbreviations: EHD - extrahepatic
disease; pretreated - 5-FU-based treatment for metastatic disease.
tumor. Of the 11 patients starting hepatic arterial 5-FU
in the presence of the primary tumor, all except one,
who had rectal cancer treated by cryosurgery and had an
arterial port placed, started treatment by a percutaneous
catheter. In three patients the arterial port placement
was planned between chemotherapy cycles at time of
surgery for the primary tumor. Because of unexpected
extrahepatic disease found at laparotomy, the port was
not placed in two of these three patients. The remaining
8 patients never had the primary tumor resected.
Of the 145 patients treated, 93 patients started treatment with port implantation and 52 with a percutaneous
catheter. In 20 of these 52 patients, an arterial port was
implanted subsequently (Figure 1). A total of 1014 cycles
of arterial chemotherapy were given, 845 by arterial port
and 169 by percutaneous catheter. The median number
of cycles per patients was 6 (range 0-31), 6 for ports
(range 0-28) and 3 for percutaneous catheters (range
0-10). Five patients who started treatment with a port
implantation never received treatment. In four because
of surgical complications. Of these four, one patient died
of pulmonary embolism and three patients developed
postoperative intra-abdominal infection and had deterioration of their performance score preventing further
therapy. The other patient did not receive hepatic arterial treatment because only minimal flow was directed
to the liver as shown on the postoperative MAA-scan
with preferential flow to the splenic artery for which
embolization was believed to be of no help. Three patients who underwent port placement after hepatic arterial treatment by percutaneous treatment never received
therapy by port. In one patient catheterization of the
hepatic artery by the femoral route was not possible. In
this patient, because of extensive liver enlargement due
to metastases, the gastroduodenal artery could not be
identified at laparotomy and the catheter was placed in
the umbilical vein. This patient died of rapidly progressive disease before chemotherapy could be started. In
one patient no adequate liver perfusion was established
and another patient had a bowel perforation during port
placement and at re-operation the port was removed.
After recovery this patient received hepatic arterial
5-FU by a percutaneous catheter.
Response evaluation
Eighteen of the one hundred forty-five patients (12%)
were not evaluable for response for the following reasons:
five patients never received hepatic arterial chemotherapy; six patients had only one chemotherapy cycle because treatment-related complications precluded further
treatment, two patients had disease not evaluable for
response on CT-scan or ultrasound, and in five patients
no adequate evaluation was performed.
Of the 127 evaluable patients 4 had a complete
response (3%), 39 (31%) patients had a partial response,
65 (51%) had stable disease and progressive disease was
observed in 19 (15%). In the evaluable non-pretreated
patients, with disease limited to the liver, the overall
response rate was 32%, with a 4% complete response
rate and progressive disease in 12%. In evaluable patients with extrahepatic disease the overall response
(including hepatic and extrahepatic sites) was evaluated.
Thirty-six percent had a partial response and four percent had a complete response, resulting in an overall
response rate of forty percent. Of the evaluable 5-FUpretreated patients, 15% had a partial response, and
complete responses were not observed. Table 2 shows
1566
Table 3. Response, time to progression and survival depending on
route.
Response
CR
PR
SD
PD
TTP
OS
All patients
PC only
PC and AP
APonly
(77 = 145)
(77 = 32)
(77 = 20)
(77 = 93)
(/o)
(/»)
(/o)
(/o)
3
31
51
15
7.5
0
11
58
31
4.9
5
58
26
10
4
32
54
11
7.3
(0.7-82.5)
14.3
(11.1-17.5)
(1.4-24.2)
6.4
(4.3-8.6)
12.9
(2.1-32.6)
23.9
(14.3-33.5)
(1.5-82.5)
17.4
(13.9-20.9)
0
X
1
w
0)
c
.2
tl>
O)
0
Abbreviations: PC - percutaneous catheter; AP - arterial port: CR complete response; PR - partial response; SD - stable disease; PD progressive disease; TTP - median time to progression in months
(95% CI); OS - median overall survival in months (95% CI).
the response rates for evaluable patients and the response
rates in the intention to treat analysis.
Because the decision to start treatment with a percutaneous catheter or arterial port was based on the
extension of disease and a port was subsequently placed
depending on initial response to therapy, we performed
a separate response evaluation of these selected groups.
Table 3 depicts the results. Fourteen of the fifty-two patients (27%) who started treatment with a percutaneous
catheter had a partial response and thirteen of these had
a laparotomy for port placement. Two of these thirteen
patients had evidence of intra-abdominal extrahepatic
disease and finally had no port implanted. In the one
patient with a partial response who did not have a
laparotomy for port placement, a recent operation for
ileus contraindicated repeat surgery.
Time to progression (TTP) and overall survival (OS)
(Figure 2)
At the time of evaluation, 16 patients were still alive, and
6 patients were still on treatment. The median TTP for
all patients was 7.5 months (95% CI: 8.1-12.5), the
median OS 14.3 months. The median TTP was significantly prolonged (P = 0.0001) in patients who had to
stop treatment because of a complication (67 patients, 9
months) compared to patients who stopped for progressive disease (58 patients, 5.3 months), respectively. This
difference was even more pronounced for the median OS
which was 10 months in patients who stopped for progressive disease and 19.9 in patients with complicated
treatment (P = 0.0002). In the 42 responding patients
(partial or complete response), the median TTP was 10.3
months in the 11 patients (1 patient with disease outside
the liver, three pretreated patient) who stopped because
of progressive disease, and 11.0 months in the 29 patients
who stopped because of a complication (5 patients with
disease outside the liver, no pretreated patients) (P - 0.1).
Median OS for both groups was 18.5 and 27.2 months
respectively (P = 0.09). Of the other 2 out of 42 responding patients 1 is still in complete remission and 1 had the
Months
8
g
CO
1
60
months
Figure 2. Overall progression-free survival and overall survival.
liver metastases resected. OS was significantly longer
in patients without extrahepatic disease (17.3 months)
compared to patients with extrahepatic disease (9.7
months, P - 0.004) (Table 4).
The decision to start treatment with a percutaneous
catheter with responding patients subsequently receiving
an arterial port, or to start treatment with an arterial
port, was merely based on the patient condition and
the presence or absence of extrahepatic disease. For this
reason we also evaluated these subgroups. The results
are depicted in Table 3. The group of patients who had
an arterial port implanted after starting treatment with
a percutaneous catheter had the longest TTP and OS of
11.6 and 23.9 months, respectively.
Reasons for treatment discontinuation in patients with
arterial ports and percutaneous catheters
Progressive disease was the reason for treatment discontinuation in 58 of the 145 patients (40%). Progression
was localized in the liver only in 24 (41%), outside the
liver in 20 (35%), and both in the liver and at distant
sites in 14 patients (24%). Sixty-eight of one hundred
forty-five patients (47%) had to stop treatment because
1567
Table 4. Time to progression for different patient groups.
All patients
No pretreatment hepatic
disease only
Extrahepatic disease
Prior 5-FU based therapy
Yes
Yes
No
MedianTTP (95% CI)
/•-value
7.5 (6.2-8.8)
8.5(6.3-10.7)
6.7(3.0-10.4)
7.6(6.2-9.0)
0 070
Median OS (95% CI)
/'-value
14.3(11.1-17.5)
18.3(15.9-20.7)
9.7(4.1-15.3) 17.3(14-20.5)
0.004
6.2(4.7-7.7)
No
8.1(6.3-9.9)
0.040
10.1(8.0-12.2) 17.4(13.9-21.0)
0.009
Abbreviations: TTP - median time to progression in months (95% CI); OS - median overall survival in months (95% CI).
Table 5. Arterial ports and percutaneous catheters: reasons for treatment discontinuation.
Reason for treatment
discontinuation
Arterial port
(%)
Percutaneous
catheter (%)
Progressive disease
Complication
Hepatic artery thrombosis/
coehac trunc thrombosis
Hepatic artery stenosis
Catheter dislocation
Upper intestinal bleeding/
ulcus ventricuh
Deep venous thrombosis
Liver abscess
Pocket abscess
Complicated surgery
Complaints
Other reasons
Arterial port placement
Extrahep. dis. at port placement
Patient refusal
Metastasectomy
Not related to treatment
Unknown
Still on treatment
42 (37)
60 (53)
16(31)
7(13)
29 (26)
3
1
2
13(12)
3
1
1
1
4
8(7)
5(4)
1
2
2
1(1)
5(4)
26 (50)
20 (38)
4
2
3(6)
of a complication. Nineteen patients stopped due to
other reasons (Table 5).
Arterial ports
Hepatic artery thrombosis (in 29 patients, 26%) and
dislocation of the catheter (in 13 patients, 12%) were the
complications most frequently observed in patients with
an arterial port. The median number of treatment cycles
before a complication led to discontinuation of chemotherapy was 6 in both patient groups. The presenting
symptoms of these complications were abdominal pain
and obstruction of the catheter. Hepatic artery thrombosis was confirmed by angiography in 17 patients; in
the other 12 patients the diagnosis was based on MAAscan and/or Doppler-ultrasound examination of the
hepatic arterial flow. Complaints, most often pain, disappeared after infusion was stopped and no other sequelae were observed. In three patients a stenosis of the
hepatic artery was observed. Dislocation occurred as a
result of migration of the catheter-tip from the gastroduodenal artery. Catheter migration to the abdominal
space, the biliary ducts, or intestines was observed.
Dislocation was observed in 13 patients, and based on
angiography combined with findings on abdominal CTscan and MAA-scan in 11 patients. Often dislocation
was accompanied by serious symptoms, which required
removal of the access system in 6 of these 13 patients.
Two of these six patients presented with upper intestinal
bleeding due to perforation of the stomach and duodenum with ulceration, as confirmed at surgery. One
patient in whom the port was removed had duodenal
perforation without bleeding. Another patient had abdominal cramps and icterus, caused by a small abscess
in the hepatoduodenal ligament with an extensive
fibrotic reaction, resulting in stenosis of the choledochal
duct. The port was removed and aT-drain was left in the
choledochus. Of the other two patients who had the port
removed, in one the catheter was located in an abdominal abscess and in the other in the abdominal space
surrounded by extensive fibrotic tissue. In 7 of 13 patients the access system was left in situ but no longer
used. In one patient there was no bleeding focus found
despite an extensive work-up consisting of a gastroduodenoscopy and colonoscopy. Transient jaundice due to
dislocation to the biliary system caused hemobilia. One
patient developed abdominal cramps and jaundice, and
the catheter tip was dislocated to the biliary system. This
catheter was left in situ but no longer used. In the additional five patients there was dislocation to the abdominal space in four and perfusion of the splenic artery
in one.
In eight patients treatment was ended because of
other complaints consisting mostly of unexplained abdominal pain during treatment. Four patients had complicated surgery and did not receive hepatic arterial
chemotherapy. One patient who had a klebsiella septicaemia short after port placement developed a liver
abscess after three cycles of hepatic arterial 5-FU, which
was drained and treated with antibiotics. This patient
developed a fatal liver bleeding at the drain site for
which it was decided not to operate. One patient developed an abscess of the subcutaneous port pocket, with
formation of an enterocutaneous fistula for which the
system was removed. This patient, who received a total
of 28 hepatic arterial treatments, died of septic complications.
1568
Percutaneous catheters
Of the 52 patients who started treatment with a percutaneous catheter, 24 patients had subsequently an arterial
port placement which was actually placed in 20 patients.
In four patients, extrahepatic disease became evident at
laparotomy, and in these patients the arterial port was
not placed, and no further arterial treatment was given.
Other reasons for treatment discontinuation were stenosis of the hepatic artery (2 patients), thrombosis of the
truncus coeliacus (1 patient), upper gastro-intestinal
bleeding (2 patients), ulcus ventriculi (1 patient), deep
venous thrombosis (1 patient) and patient refusal in two.
Discussion
The lack of an improvement of survival with hepatic
arterial chemotherapy when compared to systemic treatment, the morbidity associated with hepatic arterial
delivery, and extrahepatic disease progression are
reasons for hesitation by many clinicians. This analysis
evaluates hepatic arterial 5-FU-infusion chemotherapy
in a large number of patients treated consecutively at a
single institution. The results observed in our population
(ORR: 34%,TTP: 7.5 months and OS: 14.3 months) are
similar to those obtained in phase III studies despite the
inclusion of patients with poor prognostic characteristics [5-10]. Patients who received prior 5-FU-based
chemotherapy for metastatic disease had a response rate
of only 15% and a significantly shorter TTP and OS
compared to non-pretreated patients (Table 4). Although
earlier reports suggested that patients with prior treatment did as well as non-pretreated patients, recently, poor
response rates (14%—25%) and survival (10-12 months)
in patients refractory to 5-FU-based chemotherapy for
metastatic disease have been reported [15-17]. Based on
these observations it may be concluded that there is no
rationale for the use of single-agent 5-FU HAI in patients progressive on i.v. 5-FU based chemotherapy or in
patients with extrahepatic disease. Other treatments,
like systemic CPT-11 or oxaliplatin, possibly in combination with 5-FU HAI may be preferred in this setting.
Forty-seven percent of patients had to terminate
treatment because of complications related to arterial
delivery. Hepatic arterial thrombosis and dislocation of
the catheter were the reasons observed most often. An
important difference with most studies is that we used
5-FU which had to be delivered by an external pump
connected to an arterial access system or percutaneous
catheter, because of the relatively high volumes needed.
Most other studies used FUDR delivered by an implantable pump and hepatic artery thrombosis was not reported to be a problem [18, 19]. In a study of Rougier et
al. who also used 5-FU via an arterial port, the median
patency of the hepatic artery was 6.5 months with 50%
of the patients developing thrombosis of the hepatic
artery, very similar to our results [20]. Lorenz and
Miiller described in a study comparing i.v. 5-FU-leucovorin, HAI 5-FU-leucovorin and HAI FUDR, a tech-
nical complication rate in 16 of 40 (40%) patients
receiving HAI 5-FU-leucovorin (all arterial ports) and
in 3 of 37 (8%) in patients on HAI FUDR treatment (7
implantable pumps/30 arterial ports) [21]. The pathogenesis of these complications is not clear. Direct vessel
irritation by 5-FU and alterations of arterial flow by
catheter implantation are likely to be involved. The
intermittent character of the infusion of 5-FU applied
with port plus external pump and the continuous infusion with heparinized saline in-between FUDR treatments with implantable pumps may also be important
factors accounting for the high rate of thrombosis found
with arterial ports as compared with implantable pumps.
Theoretically, there is the possibility that with FUDR
the symptoms of vascular occlusion are less overt because of the relatively low volume infused, accounting
for the less frequent diagnosis of thrombosis compared
to 5-FU and arterial ports. Studies with systematic
follow-up of hepatic artery patency are lacking. The
occurrence of complications related to the arterial port
do not seem to affect treatment outcome. Lorenz and
Miiller reported a prolonged TTP for patients with HAI
5-FU-leucovorin compared to patients with HAI FUDR
(9.2 vs. 5.9 months; P - 0.033), despite the much higher
complication rate in the 5-FU-leucovorin group [21].
Doci et al. found no survival difference between patients
treated with a port or implantable pump. The median
patency time was much longer for pumps (28 months)
than for arterial ports (9 months). Because TTP was
shorter than device patency there was no survival difference between patients with arterial port or pump [22], In
our study we found that the time to progression for
patients who had to stop because of a complication was
significantly longer than the time to progression for
patients who stopped because of progressive disease (9.0
and 5.3 months, respectively; P - 0.0001). TTP and OS
were prolonged in patients responding to HAI who
stopped because of a complication compared to patients
responding to HAI who stopped because of progressive
disease (11 vs. 10.3 months and 27.2 vs. 18.5 months,
respectively, P = NS). Based on these findings it does not
seem to be likely that treatment complications have a
negative impact on outcome of treatment. Moreover, it
cannot be ruled out that hepatic arterial occlusion prolongs TTP an OS because of devascularization of the
tumor. However, no increase in response or survival has
been reported with chemoembolization [23, 24]. In our
study there was no significant difference in the TTP and
OS between responding patients who stopped treatment
because of catheter dislocation and those who stopped
for hepatic artery thrombosis.
A major reason for failure of HAI is the development
of extrahepatic disease, which occurs as first failure in
55% of patients [12]. In our analysis disease progression
outside the liver alone was observed in 35%, and in 24%
progression was both in the liver and at other sites.
In order to overcome extrahepatic failure different approaches have been studied in phase I and II studies.
Systemic FUDR combined with hepatic arterial FUDR
1569
[25], and high dose 5-FU infusion (with leucovorin
biomodulation) in order to cause 'overflow' into the
systemic circulation [26, 27] are most appealing. Furthermore, in systemic therapy biomodulation of 5-FU with
leucovorin could double the response rates. Although
in some studies a reduction of extrahepatic progression
has been reported, phase III studies comparing these
methods to systemic i.v. treatment with 5-FU-leucovorin are needed to answer the question if a survival
advantage is offered with combined arterial and systemic treatment. Randomized trials comparing hepatic
arterial 5-FU and systemic leucovorin with systemic
5-FU-leucovorin, are ongoing (MRC sponsored study,
UK and CALGB, USA).
Regarding the palliative character of hepatic arterial
chemotherapy, the results of quality of life studies are of
interest. Allen-Mersh et al. found a prolonged survival
with a normal quality of life in patients receiving hepatic
arterial FUDR as compared with patients receiving
symptom palliation only [7]. However, in this study,
device-related complications occurred in only 8 of 51
patients and were reason for treatment termination in
only 2 patients. The median treatment duration was 12
cycles. Most other published studies reported higher
complication rates. A later report of the same group
showed a better-sustained quality of life with similar
survival for HAI as compared with systemic chemotherapy [28]. In our study, complications led to serious
sequelae in only a small number of patients, although
they resulted in treatment ending in most of the patients.
Unsuccessful treatment after abdominal surgery for
access system placement is an important issue in these
patients with a limited life expectancy. Avoiding unnecessary surgery was an important argument to start
treatment with a percutaneous catheter in patients for
whom there was serious doubt if they would derive any
benefit from hepatic arterial chemotherapy. In our
opinion this approach was justified by the fact that in
this type of patients survival will mainly depend on
hepatic disease control. Although TTP and OS in patients who ultimately received a port suggested adequate
selection, only 20 of 52 patients had finally a port placed
and in patients who never had an access system placed
TTP and OS were only 3.0 and 6.4 months, respectively.
These results do not justify the morbidity of treatment
with percutaneous catheters and better methods for
patient selection are required.
There is still much discussion whether hepatic arterial
chemotherapy has a place in the treatment of patients
with liver metastases. The development of antineoplastic
agents resulted in new systemic opportunities to which
the results of hepatic arterial chemotherapy have to be
compared. With the introduction of CPT-11 and oxaliplatin, new chemotherapeutic regimens result in promising response rates and prolonged survival, even in pretreated patients [29-32]. On one hand this may challenge
the use of HAI and on the other new combinations
of HAI 5-FU/FUDR with CPT-11 or oxaliplatin may
improve the results. Complications related to the mode
of delivery as we have described are reason for contempt
for many. With continuation of HAI it is of importance
to seek for better, less inconvenient modes of delivery.
New methods for port placement by laparoscopic or
angiographic procedures can be contemplated [33, 34].
Recent studies showed that hepatic arterial chemotherapy may be effective in the adjuvant setting after
resection of liver metastases or other local therapies
such as cryosurgery or radiofrequency ablation [13, 14].
The role of neoadjuvant HAI is being studied [35]. In
both the adjuvant and neo-adjuvant setting, complications will have more repercussions than in advanced
disease. Occurrence of hepatic artery thrombosis, catheter dislocation or biliary sclerosis might have major
consequences, causing liver insufficiency, infectious
complications, or making metastasectomy impossible.
In conclusion: The high complication rate, and the
lack of a survival benefit, make hepatic arterial 5-FU
treatment for patients with colorectal cancer metastatic
to the liver a disputable alternative to systemic treatment.
Especially since new treatment options have evolved.
Based on our results HAI 5-FU cannot be regarded as
standard treatment for colorectal cancer confined to the
liver. Before conducting further studies on this mode of
hepatic arterial treatment, better methods of delivery
in order to reduce the complication rate and morbidity
are needed.
References
1. Stangl R, Altendorf-Hofmann A, Charnley RM et al Factors
influencing the natural history of colorectal liver metastases.
Lancet 1994; 343: 1405-10.
2. Fong Y, Cohen AM, Former JG et al. Liver resection for colorectal metastases. J Clin Oncol 1997; 15: 939-46.
3. Ensminger WD, Rosoowsky A, Raso V et al. A clinical pharmacological evaluation of hepatic arterial infusions of 5-fluoro-2'deoxyuridine and 5-fluorouracil. Cancer Res 1978; 38: 3784-92.
4. Chen H-SG, Gross JF. Intra-arterial infusion of anticancer drugs:
Theoretic aspects of drug delivery and review of responses.
Cancer Treatment Reports 1980; 64: 31-40.
5. Kemeny N, Daly J, Reichman B et al. Intrahepatic or systemic
infusion offluorodeoxyuridinein patients with liver metastases
from colorectal carcinoma. Ann Intern Med 1987; 107: 459-65.
6. Rougier P, Laplanche A, Huguier M et al. Hepatic arterial
infusion of floxuridine in patients with liver metastases from
colorectal carcinoma: Long-term results of a prospective randomized trial. J Clin Oncol 1992; 10: 1112-8.
7. Allen-Mersh TG, Earlam S, Fordy C et al. Quality of life and
survival with continuous hepatic-artery floxuridine infusion for
colorectal liver metastases. Lancet 1994; 344: 1255-60.
8. Hohn DC, Stagg RJ, Friedman MA et al. A randomized trial of
continuous intravenous versus hepatic intraarterial floxuridine
in patients with colorectal cancer metastatic to the liver: The
Northern California Oncology Group Trial. J Clin Oncol 1989; 7:
1646-54.
9. Chang AE, Schneider PD, Sugarbaker PH et al. A prospective
randomized trial of regional versus systemic continuous 5-fluorodeoxyuridine chemotherapy in the treatment of colorectal liver
metastases. Ann Surg 1987; 206: 685-93.
10. Martin JK, O'Connel MJ,Wieand HS et al. Intra-arterial floxuridine vs. systemicfluorouracilfor hepatic metastases from colorectal cancer. Arch Surg 1990; 125: 1022-7.
1570
11. Harmantas A, Rotstein LE, Langer B. Regional versus systemic
chemotherapy in the treatment of colorectal carcinoma metastatic to the liver. Is there a survival difference? Meta-analysis of
the published literature. Cancer 1996; 78: 1639-45
12. Meta-Analysis Group in Intestinal Cancer. Reappraisal of
hepatic infusion in the treatment of liver metases from colorectal
cancer. J Natl Cancer Inst 1996: 88: 252-8.
13. Kemeny N, Huang Y, Cohen AM et al. Hepatic arterial infusion
of chemotherapy after resection of hepatic metastases from colorectal cancer. N Engl J Med 1999; 341: 2039-48.
14. Kemeny MM, Adak S, Lipsitz S et al. Results of the Intergroup
(Eastern Cooperative Oncology Group (ECOG) and Southwest
Oncology Group (SWOG) prospective randomized study of surgery alone versus continuous hepatic artery infusion of FUDR
and continuous systemic infusion of 5FU after hepatic resection
for colorectal liver metastases. Am Soc Clin Oncol 1999; 18: 264,
1999 (Abstr).
15. Fordy C, Glover C, Davies MM et al. Hepatic arterial floxuridine
as second-line treatment for systemic fluorouracil-resistant colorectal liver metastases. Br J Cancer 1998; 78: 1058-60.
16. Zamparelh G, Pancera G, Barm S et al. Intraarterial infusion
(HAI) of floxuridine (FUDR) as second-line therapy of liver
metastases from colorectal cancer. A GISCAD phase II study.
Proc Am Soc Clin Oncol 1998; 17: all81.
17. Kemeny N, Cohen A, Seiter K et al. Randomized trial of hepatic
arterial floxuridine, mitomycin and carmustine versus floxuridine
alone in previously treated patients with liver metastases from
colorectal cancer. J Clin Oncol 1993; 11: 330-5.
18. Hohn DC, Rayner AA, Economou JS et al. Toxicities and
complications of implanted pump hepatic arterial and intravenous floxuridine infusion. Cancer 1986; 57: 465-70.
19. Campbell KA, Burns RC, Sitzmann JV et al. Regional chemotherapy devices: Effect of experience and anatomy on complications. J Clin Oncol 1993; 11: 822-6.
20. Rougier P, Ducreux M, Pignon JP et al. Prognostic factors in
patients with liver metastases from colorectal carcinoma treated
with discontinuous intra-arterial hepatic chemotherapy. Eur J
Cancer 1991; 27: 1226-30.
21. Lorenz M, Miiller H-H. Randomized, multicenter trial of fluorouracil plus leucovorin administered either via hepatic arterial
or intravenous infusion versusfluorodeoxyuridineadministered
via hepatic arterial infusion in patients with nonresectable liver
metastases from colorectal carcinoma. J Clin Oncol 2000; 18:
243-54.
22. Doci R, Bignami P, Montalto F et al. Prognostic factors for
survival and disease-free survival in hepatic metastases from
colorectal cancer treated by resection. Tumori 1995; 81 (Suppl):
143-6.
23. Hunt TM, Flowerder AD, Birch SJ et al. Prospective randomized
controlled trial of hepatic arterial embolisation or infusion
chemotherapy with 5-fluorouracil and degradable starch microspheres for colorectal liver metastases. Br J Surg 1990; 77:779-82.
24. Leichman CG. Jacobson JR, Modiano M et al. Hepatic chemoembolization combined with systemic infusion of 5-fluorouracil
and bolus leucovorin for patients with metastatic colorectal
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
carcinoma: A Southwest Oncology Group pilot trial. Cancer
1999; 86: 775-81.
Safi F, Bittner R, Roscher R et al. Regional chemotherapy for
hepatic metastases of colorectal carcinoma (continuous intraarterial versus continuous intra-arterial/intravenous therapy).
Results of a controlled clinical trial. Cancer 1989; 64: 379-87.
Warren HW, Anderson JH, O'Gorman Pet al. A phase II study of
regional 5-fluorouracil infusion with intravenous folinic acid for
colorectal liver metastases. Br J Cancer 1994; 70: 677-80.
Kerr DJ, Ledermann JA, McArdle CS et al. Phase I clinical and
pharmacokinetic study of leucovorin and infusional hepatic arterialfluorouracil.J Clin Oncol 1995; 13: 2968-72.
Earlam S, Glover C, Davies M et al. Effect of regional and
systemic fluorinated pyrimidines chemotherapy on quality of life
in colorectal liver metastasis patients. J Clin Oncol 1997; 15:
2022-9.
Rougier P, Cutsem Ev, Bajetta E et al. Randomised trial of
irinotecan versusfluorouracilby continuous infusion after fluorouracil failure in patients with metastatic colorectal cancer.
Lancet 1998; 352: 1407-12.
Cunningham D, Pyrhonen S, James RD et al. Randomised trial
of irinotecan plus supportive care versus supportive care alone
after fluorouracil failure for patients with metastatic colorectal
cancer. Lancet 1998; 352: 1423-8
Gramont Ad, Vignoud J, Tournigand C et al. Oxaliplatin with
high-dose leucovorin and 5-fluorouracil 48-hour infusion in
pretreated metastatic colorectal cancer. Eur J Cancer 1997; 33.
214-9.
Giacchetti S, ltzhaki M, Gruia G et al. Long-term survival of
patients with unresectable colorectal cancer liver metastases following infusional chemotherapy with 5-fluorouracil. leucovorin,
oxaliplatin and surgery. Ann Oncol 1999; 10: 663-9.
Seki H, Kimura M, Yoshimura N et al. Hepatic arterial infusion
chemotherapy using percutaneous catheter placement with an
implantable port: Assessment of factors affecting patency of the
hepatic artery. Clin Radiol 1999; 54.
Jung HY, Shim HJ, Kwak BK et al. Percutaneously implantable
catheter-port system for chemotherapeutic infusion through the
hepatic artery. Am J Radiol 1999; 172: 641-4.
Link KH, Pillasch J, Formentini A et al. Downstaging by regional
chemotherapy of non-resectable isolated colorectal liver metastases. Eur J Surg Oncol 1999; 25: 381-8.
Received 31 May 2000; accepted 24 August 2000.
Correspondence to"
J. M. G. H. van Riel, MD
Department of Medical Oncology
University Hospital Vrije Universiteit
DeBoelelaanlll7
1081 HV Amsterdam
The Netherlands
E-mail: [email protected]