|
Introduction Colorectal
carcinoma is the second most frequent cause of cancer death with increasing frequency in
the western world. Almost half of the patients with colorectal carcinoma are cured by
radical surgical resection of the primary tumour together with the corresponding lymphatic
drainage area. At the time of initial diagnosis approximately twenty percent of patients
present with synchronous liver metastases and approximately 10-15 percent develop
metachronous liver metastases. Some 20 percent of patients with synchronous or
metachronous liver metastases from colorectal cancer can be offered liver resection with
curative intent. Five-year survival of these patients has been reported as 25 percent
although recurrence-free 5-year survival is less than 20 percent [1]. Since
perioperative mortality is consistently very low there is little room for improvement. Any
improvement of longterm survival must, therefore, be attributed to patient selection
rather than improved surgical therapy. Adjuvant treatment has therefore been administered
in an attempt to improve longterm survival.
Metaanalysis of randomized trials studying adjuvant chemotherapy in patients with
primary colorectal cancer and lymph node metastases (UICC stage III) has concluded that
adjuvant 5-fluorouracil (5-FU) based chemotherapy confers a small survival benefit of some
2-5 percent at 5 years. [2]. Adjuvant chemotherapy has been applied in a similar fashion
to patients with curatively resected liver metastases of colorectal origin by the
systemic, intraperitoneal intraarterial or intraportal route [3]. From some series
using these treatment concepts 5-year survival of almost 40 percent has been reported, but
rarely more than 20 patients were studied [4,5]. While the majority of studies were
non-randomized and often retrospective, some randomized and controlled prospective trials
have been announced almost a decade ago. However, to date little information has become
available from these efforts. Either the respective trials have not been completed or they
have not been formally published. Clear treatment strategies regarding adjuvant
chemotherapy after potentially curative resection of colorectal liver metastases have not
been established.
Since 1981, we have performed an increasing number of liver resections for metastatic
disease. Although adjuvant chemotherapy has not been routinely recommended after
potentially curative liver resection, a number of our patients has received adjuvant 5-FU
based chemotherapy during the past 15 years. This was either because the patient
specifically asked for such treatment or because physicians responsible for the patient
after liver resection in their local community recommended such therapy. We are,
therefore, able to examine in our patients the experience with adjuvant chemotherapy after
resection of colorectal liver metastases.
Patients and Methods
Patient characterisation
Between 1981 and 1995, 286 patients had liver resection for metastatic colorectal
cancer at the Department of Surgery, University of Heidelberg. Surgical procedures
included hemihepatectomies (27%), segmentectomies of various extent (38%), and wedge
resections (27%). To be included in this study, liver resection had to be curative with
macroscopically and microscopically free resection margins. Patients who had received
neoadjuvant chemotherapy prior to liver resection as well as patients that had received
adjuvant chemotherapy after resection of the primary tumor were excluded. Therefore, 75
patients did not meet the inclusion criteria of having curative liver resection and being
chemotherapy 'naive'. All patients that died within 2 months postoperatively were excluded
from survival analysis [n=9 (3,7%)].
Overall, 202 patients were eligible for the analysis. Due to the long observation
period (1981-1995) and because a number of physicians were involved, chemotherapy regimens
varied. However, all chemotherapy regimens included 5-fluorouracil. All study patients
were followed until death or the end of the observation period for a minimum of 6 months
and a median of 24 months.
Potential prognostic factors were identified by a previous review of the literature and
examined by univariate analysis. They included age, sex, stage and location of the primary
tumour, time until treatment of liver metastasis, surgical margin (more or less than 10
mm), amount of blood lost during liver resection, grading, size and number of liver
metastasis and other factors [1]. They were tested in the current study population and, if
confirmed, subsequently entered into a multivariate analysis to identify risk factors
independently associated with overall survival. Since the relative risks associated with
each of the identified risk factors were comparable (TABLE 1), the relative risk of death
could be described by the number of risk factors present. All patients were then assigned
to three risk groups: one group with an expected low risk for treatment failure
characterised by the presence of none, one or two risk factors, a group with a medium risk
of treatment failure with patients carrying three risk factors, and finally a high risk
group characterised by the presence of four or all five confirmed risk factors. After
stratification for relative risk of death into three groups the survival of patients
receiving adjuvant chemotherapy after complete resection of colorectal liver metastasis
was compared to survival of patients not receiving postoperative 5-FU-based chemotherapy
separately for each risk level.
For statistical analysis SAS software, release 6.12, was used. The statistical end
point of this study was overall survival from the date of liver resection. Cancer-related
deaths were considered the events of interest for survival analysis. All other observed
survival times including 'not cancer related' deaths were right-censored. Overall survival
estimates were calculated by the Kaplan-Meier method. The log rank test was used to
analyse the difference between survival curves. Factors independently associated with
overall survival were identified by multivariate analysis using Cox´s proportional hazard
model. Step down regression methods were used to set up parsimonious statistical models
for relating prognostic factors to the overall survival of patients with complete data.
Results
153 patients were treated by liver resection without chemotherapy and 49 received
adjuvant 5-fluorouracil based chemotherapy. All potential prognostic factors were almost
evenly distributed among patients receiving chemotherapy and those treated by surgery
alone (Table 4). In a univariate survival analysis, 7 of 13 factors were found to
correlate with overall survival in these patients. These included age, surgical margin,
disease free interval, lymph node metastasis at the initial diagnosis, amount of blood
loss during operation and grading of liver metastasis. These factors were then entered
into a multivariate analysis, which identified five parameters significantly associated
with treatment failure (Table 1). These included a high grading of the liver metastasis, a
small resection margin (<1 cm), disease free interval <12 months, lymph node
metastasis of the primary tumour, and an intraoperative blood loss exceeding 2 liters. All
patients were then allocated to one of three risk groups. 113 patients were assigned to
the low risk group (0-2 risk factors), 47 patients had a medium risk (3 risk factors) and
21 patients were in the high risk group (4 or all risk factors). The validity of the risk
groups was confirmed by analysis of the overall survival in the defined risk groups.
Patients with a low risk profile had a significantly better 5-year survival than patients
with the medium or high risk profile 5-year survival was 35, 17 and 0 percent for low,
intermediate and high risk groups.
Table 1: Identification of independent risk factors for treatment
failure after potentially curative resection of colorectal liver metastases determined by
multivariate analysis.
| Risk factor |
p |
relative risk |
| grading
of liver metastases |
<0.01 |
1.92 |
| resection
margin < 1 cm |
<0.01 |
1.76 |
| synchronous
(<12 months) metastases |
<0.03 |
1.56 |
| primary
tumor lymph node positive |
<0.05 |
1.47 |
| blood
loss > 2 l |
0.1 |
1.44 |
We then examined the effect of adjuvant chemotherapy on survival in all patients and in
the different risk groups. Regarding all patients (n=202), survival was better for
patients treated by surgery alone (n=153) then in patients receiving adjuvant chemotherapy
(n=49), but the observed difference was not significant (p=0.095; Figure 1). When high and
intermediate risk groups were examined separately no significant difference in survival
was observed in patients treated by adjuvant chemotherapy in the risk groups. However, in
the group of patients with a comparatively low risk of treatment failure, patients treated
by surgery alone had a significantly better overall survival than those receiving adjuvant
chemotherapy (Figure 2).
Figure 1: Survival of all patients treated by liver resection alone
or liver resection plus adjuvant chemotherapy. (n=202; Kaplan-Meier estimate, log-rank
test).
Figure 2: Survival of patients in the low risk group treated by liver
resection alone or liver resection plus adjuvant chemotherapy. (n=113; Kaplan-Meier
estimate, log-rank test).
Discussion
Resection of liver metastasis from colorectal carcinoma has been established as the
treatment of choice, if size and location of metastases permit complete removal of the
tumor. Surgical treatment alone has resulted in five-year survival rates of about 25% in
eligible patients. Although this must be considered very successful in patients with
metastatic solid tumors, it still implies that the majority of patients are rendered
tumor-free only temporarily and eventually will die from their tumor.
Reports in the literature regarding the effect of 5-FU based adjuvant chemotherapy
after potentially curative surgical resection of liver metastases are not uniform. Hughes
and co-workers in a collective review of 155 patients observed a 5-year survival of 33%
after systemic application of 5-FU. However, others following the same approach found that
only 15% of patients survived for 5 or more years (6). Two groups used 5-FU
intraarterially and found 5-year survival rates of 30 and 35%, but only 10 patients were
treated in these studies (7,8). Finally, the largest study of adjuvant chemotherapy using
different modes of application was published in 1992 by Nordlinger. In 666 patients, the
5-year survival rate after adjuvant chemotherapy was 25% (9). This rate is almost
identical to results from a collective review of 3643 patients where 5-year survival was
reported as 26% after liver resection alone (1).
In this non-randomized study identification of risk factors allowed definition of
patient groups at different risk for tumor-related death. In 202 patients undergoing
complete resection of colorectal liver no evidence was obtained to indicate that adjuvant
5-FU based chemotherapy would improve long-term outcome of these patients. When patients
were stratified according to presumptive risk factors it rather appeared that chemotherapy
after resection of hepatic metastases was associated with a reduced long term survival in
the subgroup of patients with a relatively good prognosis (low risk group). No significant
differences were observed in the intermediate and high risk groups. These observations do
not support the use of adjuvant chemotherapy after potentially curative resection of
colorectal hepatic metastases outside clinical studies.
References
- [1]
- Lehnert T, Otto G, Herfarth C. Therapeutic Modalities and Prognostic
Factors in Primary and Secondary Liver Tumors. World J Surg 1995; 19: 252-263.
- [2]
- Dubé S, Heyen F, Jenicek M. Adjuvant chemotherapy in colorectal
carcinoma. Dis Colon Rectum 1997; 40: 35-41.
- [3]
- Nitti D, Civalleri D, Samori G, Bachi V, Dapian P, Labianca R, Pancera G,
Lise M, Panvini D, Tinazzi A, Torri V, Marsoni S. Retrospective study on adjuvant
chemotherapy after surgical resection of colorectal cancer metastatic to the liver. Eur J
Surg Oncol 1994; 20: 454-460.
- [4]
- Tsujitani S, Watanabe A, Kakeji Y, Maehara Y, Tomoda H, Furusawa M,
Sugimachi K. Hepatic recurrence not prevented with low dosage long-term intraportal 5-FU
infusion after resection of colorectal liver metastases. Eur J Surg Oncol 17: 526-529,
1991.
- [5]
- Porte H, De Moulins H, Gambiez L, Wurtz A, Quandalle P. A pilot study of
adjuvant hepatic arterial infusion chemotherapy, associating 5-fluorouracil and
leucovorin, after resection of colorectal cancer liver metastases. Surg Oncol 1995; 4:
317-322.
- [6]
- O'Connell MJ, Adson MA, Schutt AJ, Rubin J, Moertel CG, Ilstrup DM.
Clinical trial of adjuvant chemiotherapy after surgical resection of colorectal cancer
metastatic to the liver. Mayo Clin Proc 60: 517-520, 1985.
- [7]
- Patt YZ, McBride CM, Ames FC, Claghorn LJ, Cleary KR, Boddie AW,
Charnsangavej C, Mavgilit GM. Adjuvant perioperative hepatic arterial mitomycin C and
floxuridine combined with surgical resection of metastatic colorectal cancer in the liver.
Cancer 59: 867-873, 1987.
- [8]
- Wagman LD, Kemeny- MM, Leong L, Terz JJ, Hill R, Beatty JD, Kokal WA,
Riihimaki DU. A prospective, randomized evaluation of the treatment of colorectal cancer
metastatic to the liver. J Clin Oncol 8: 1885-1893, 1990.
- [9]
- Nordlinger B, Jaeck D, Guiget M, Vaillant JC, Balladur P, Schaal JC.
Surgical resection of hepatic metastases. Multicentric retrospective study by the French
Association of Surgery. In: Treatment of hepatic metastases of colorectal cancer
(Nordlinger B, Jaeck D, eds.) Springer-Verlag 1992, pp 129-161.
-
Address for correspondence:
Thomas Lehnert, MD, PhD, Section of Surgical Oncology, Department of Surgery, University
of Heidelberg, Im Neuenheimer Feld 110, D-69120 Heidelberg (Tel ++49-6221-566250, - FAX
++49-6221-565506 - e-mail thomas_lehnert@krz.mail.krz.uni-heidelberg.de). |
