Profile of Plasma Angiogenic Factors Before and After Hepatectomy for Colorectal Cancer Liver Metastases
- First Online:
- Cite this article as:
- Yoon, S.S., Kim, S.H., Gonen, M. et al. Ann Surg Oncol (2006) 13: 353. doi:10.1245/ASO.2006.03.060
- 127 Views
Circulating angiogenic factors in patients with colorectal cancer liver metastases may promote tumor growth and contribute to liver regeneration after partial hepatectomy.
We analyzed blood samples from 26 patients with colorectal cancer liver metastases before and after liver resection and used samples from 20 healthy controls as a reference. Plasma levels of vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), epidermal growth factor (EGF), and hepatocyte growth factor (HGF) were measured, and levels were correlated with recurrence.
The median preoperative levels of all four factors were significantly higher and more variable in colorectal cancer liver metastasis patients than in controls. HGF and bFGF levels increased significantly 3 days and 1 month after hepatectomy, respectively, and returned to near preoperative levels at 3 months. Postoperative VEGF and EGF levels remained relatively stably increased over 3 months. After a median follow-up of 19 months, 10 patients (42%) experienced recurrence. Higher preoperative VEGF and HGF levels correlated with subsequent recurrence (P = .018 and .021, respectively), and a preoperative adjusted total value of all four factors accurately identified patients at low, moderate, and high risk of recurrence (P = .034). Patients who experienced disease recurrence also had relatively higher bFGF levels 3 months after operation (P = .035).
Plasma angiogenic factors are increased in patients with colorectal cancer liver metastases and remain increased at least 3 months after partial hepatectomy. Measurement of certain factors before and after hepatic resection can predict recurrence. Targeted biological agents may counteract the tumor-promoting effects of these circulating factors on subclinical disease.
KeywordsColorectal cancerLiver metastasesAngiogenesisVascular endothelial growth factorBasic fibroblast growth factorHepatocyte growth factor
Colorectal cancer in the United States is the second leading cause of cancer-related mortality,1 and death usually results from uncontrolled metastatic disease. The liver is the most common site of metastasis for colorectal cancer2: nearly 50% of patients present with liver metastases or develop liver metastases during the course of disease.3 Angiogenesis, or the formation of new blood vessels, is a critical requirement for both primary and metastatic tumors to grow beyond 1 to 2 mm3.4 Tumor angiogenesis is regulated by a balance between proangiogenic and antiangiogenic factors,5 and vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) are two of the most important factors that promote angiogenesis.6 VEGF is overexpressed in most solid tumors and acts specifically on endothelial cells to promote new vessel formation.7 Basic FGF was one of the first angiogenic factors to be discovered and induces endothelial cell proliferation, migration, and capillary tube formation.8 Circulating blood levels of both VEGF and bFGF are increased in patients with primary and metastatic colorectal cancer.9
Isolated liver metastases from colorectal cancer can be surgically resected by partial hepatectomy, resulting in a 5-year survival ranging from 24% to 40%.10,11 As much as 70% to 80% of the liver can be resected during partial hepatectomy because of the liver’s tremendous capacity to regenerate.12 Liver regeneration involves both hepatocyte proliferation within the architecture of the remnant liver and endothelial cell proliferation.13 Peak DNA synthesis in human livers after hepatectomy occurs at approximately 1 week,14 most liver regeneration occurs by 2 weeks, and complete regeneration to the preresection liver volume takes up to 1 year.15 Several growth factors are required for hepatic regeneration, including hepatocyte growth factor (HGF) and epidermal growth factor (EGF).16 HGF knockout mice die in utero with atrophic livers.17 EGF acts in concert with HGF to stimulate hepatocyte proliferation after partial hepatectomy.18
It is interesting to note that both HGF and EGF not only promote hepatocyte proliferation, but also promote angiogenesis and are thus proangiogenic factors.19–21 The HGF receptor c-met is found in blood vessels, and HGF is one of the most potent mitogens of endothelial cells.20 The EGF receptor is expressed by liver sinusoidal endothelial cells and is upregulated after partial hepatectomy.22
In this study, we sought to examine the circulating plasma levels of VEGF, bFGF, EGF, and HGF before and after hepatectomy in patients with colorectal cancer liver metastases. Preoperative increases were attributed to the presence of liver metastases, and these levels were correlated with subsequent recurrence-free survival. Levels of circulating factors were also measured 3 days, 1 month, and 3 months after hepatectomy to determine how these factors changed during the primary period of liver regeneration.
Blood Sample Collection and Enzyme-Linked Immunosorbent Assay
From September 2002 through August 2003, 26 patients with potentially resectable colorectal cancer liver metastases and 20 healthy controls were entered into an institutional review board–approved prospective study at a single institution (Memorial Sloan-Kettering Cancer Center). Informed consent was obtained from all patients and controls. Approximately 10 mL of blood was collected in EDTA-containing tubes from patients within 1 month before partial hepatectomy and 3 days, 1 month, and 3 months after partial hepatectomy. Blood samples were centrifuged at 1000 × g for 10 minutes, followed by collection of the plasma. Plasma samples were stored at −80°C until analysis.
One blood sample was collected from each of the 20 healthy controls at a single time point. For the 26 patients enrolled in the study, all patients had preoperative and postoperative day 3 blood samples drawn. Twenty-four of 26 samples were collected at 1 month after operation; 2 patients did not return for the 1-month follow-up visit. Three months after hepatectomy, 21 of 26 patients had blood samples collected; 1 patient died of disease, and 5 patients did not return for this follow-up visit. Overall, 97 of 104, or 93%, of patient blood samples were collected.
Plasma samples were measured for VEGF, bFGF, EGF, and HGF by using the following commercially available enzyme-linked immunosorbent assay (ELISA) kits: Quantikine Human VEGF ELISA Kit, Quantikine bFGF HS ELISA Kit, Human EGF Quantikine ELISA Kit, and Quantikine Human HGF Immunoassay (all from R&D Systems, Minneapolis, MN). The manufacturer’s protocols were followed, and samples were measured in duplicate. ELISA plates were read by using the Emax Precision Microplate Reader (Molecular Devices, Sunnyvale, CA), and sample values were determined against four parameter standard curves. The mean value of duplicate samples was used as the final concentration.
Demographic data for the healthy controls, including age and sex, were recorded at the time of blood collection. Clinical information regarding patients was obtained from the institution’s prospective hepatobiliary database and from medical records. Details of the primary colorectal cancer and treatment were recorded and included site, tumor-node-metastasis stage, and surgical treatment. The disease-free interval was calculated from the time of colorectal cancer resection to the time of diagnosis of liver metastases. Patients were fully staged before liver resection with at least a chest/abdomen/pelvis computed tomographic scan, and all cases were presented at a twice-weekly hepatobiliary conference. Details regarding extent of liver resection, pathologic findings, and adjuvant chemotherapy were recorded. The size, number, and location of liver metastases were determined by review of preoperative imaging and pathology reports.
This pilot study was originally designed for 50 patients and 20 healthy controls. The sample size was not chosen on the basis of power considerations but rather was driven by expected accrual. When enrollment was slower than anticipated, the study was terminated at 26 patients, and the data were analyzed. Levels of angiogenic factors between patients and controls were compared by using the Wilcoxon rank sum test. Within patients, postoperative levels were compared with baseline by using the Wilcoxon signed rank test. Recurrence-free survival was evaluated for 24 patients who underwent resection of their colorectal cancer liver metastases. Survival curves were estimated by the Kaplan-Meier method23 by using the statistical software package SPSS 11.0.1 (SPSS Inc., Chicago, IL) and were compared for various factors by using the log-rank test.
Characteristics of Patients and Primary Colorectal Cancer
Twenty-six patients with colorectal cancer liver metastases that were considered surgically resectable were enrolled into this study over a 1-year period. The median age of patients was 64 years (range, 32–86 years), and there were 18 men (69%) and 8 (31%) women. Twenty healthy controls were chosen with a median age of 60 years; 14 controls (70%) were men, and 6 (30%) were women.
The patients’ original primary colorectal cancers were located in the right colon (n = 6), left colon (n = 1), sigmoid colon (n = 9), and rectum (n = 9), and one patient had multifocal colonic tumors. Right, left, and sigmoid colon tumors were resected by standard operations. The rectal lesions were resected by low anterior resection in eight cases and transanal excision in one case. The patient with multifocal disease had a total abdominal colectomy. The tumor-node-metastasis stage of patients at the time of colorectal resection was stage II in 6 patients, stage III in 7 patients, stage IV in 10 patients, and unknown in 3 patients. Of the 10 patients with stage IV disease, 8 had synchronous liver metastases, and 2 had limited intraperitoneal metastases that were resected at the time of colorectal resection.
Characteristics of Liver Metastases and Hepatectomy
Liver metastases and hepatectomy (n = 26)
Number of metastases
Size of largest metastasis (cm)
Location of metastases
Liver resection involved less than a lobectomy in approximately one half of patients and a lobectomy or extended lobectomy in the other half (Table 1). Two of the 26 patients were found to have unresectable disease at the time of operation as a result of extrahepatic disease, and these 2 patients were excluded from subsequent analyses. The median length of hospital stay was 8 days and ranged from 3 to 23 days. There were five major complications in four patients (25%), including bleeding, pulmonary embolism, intra-abdominal collection, wound infection, and ileus.
Levels of Circulating Factors
Plasma angiogenic factor levels in patients and controls
Patients (n = 24)
Controls (n = 20)
28.9 ± 22.5
330.1 ± 49.8*
367.5 ± 56.3
288.9 ± 46.4
197.5 ± 30.9**
2.2 ± .4
17.7 ± 2.6*
17.5 ± 4.0
38.1 ± 7.6**
22.5 ± 3.8
23.4 ± 5.5
55.3 ± 15.8*
32.2 ± 9.8
54.8 ± 17.5
42.2 ± 8.3
125.7 ± 17.9
405.3 ± 59.6*
1183.5 ± 145.2**
352.0 ± 161.0
347.4 ± 67.4
An estimate of tumor burden in the liver was made by calculating the volume of metastases according to computed tomographic scans. There was no significant correlation between the estimated tumor burden and preoperative levels of circulating factors. We also further analyzed postoperative bFGF levels at 1 month and HGF levels at 3 days to determine whether the increases in these factors at these time points were related to the extent of liver resection. Posthepatectomy increases in bFGF and HGF tended to be higher in patients who underwent extended left or right hepatectomies compared with lesser operations, but these differences did not reach statistical significance. When patients who underwent an extended lobectomy were compared with patients who underwent a lobectomy or less, the median bFGF levels at 1 month were 69.8 ± 18.4 pg/mL (median ± SEM) and 31.6 ± 7.7 pg/mL, respectively (P > .05). The median HGF levels on postoperative day 3 were 1470.0 ± 480.4 pg/mL and 1105.5 ± 110.6 pg/mL in these two groups (P > .05).
Two patients were found to have unresectable disease at the time of exploration for liver resection: one for limited peritoneal disease and the other for a positive porta hepatis lymph node. These patients were excluded from subsequent analyses, but postoperative blood samples were still collected on these patients. One of these patients had the highest bFGF levels (ranging from 147.9 to 229.6 pg/mL) of all collected samples. The next highest recorded bFGF levels were less than one half of these values. The other patient with unresectable disease had the highest preoperative EGF level (236.7 pg/mL) and the second highest preoperative VEGF level (709 pg/mL) of all patients.
In this study of colorectal cancer liver metastases, patients with greater than median increases in preoperative VEGF and HGF levels were much more likely to experience disease recurrence than patients with lower levels, and there was a similar trend in patients with high bFGF and EGF levels. Several studies have examined circulating VEGF levels in patients with colorectal cancer.24–29 Some of these studies have demonstrated a correlation between circulating VEGF levels and tumor stage, tumor-doubling time, response to therapy, or survival. In one study that examined serum HGF levels in patients with colorectal cancer (both localized and metastatic), HGF levels correlated with tumor stage.30 When the preoperative levels of all four factors were examined in combination, only one of the eight patients with a low combined level had disease recurrence, whereas six of the eight patients with high combined levels had disease recurrence. Thus, this study demonstrates the prognostic utility of measuring plasma angiogenic factors before resection of colorectal cancer liver metastases, and these preliminary results should be validated in a larger cohort of patients.
Prior analysis from our institution of more than 1000 patients who underwent hepatectomy for CRC liver metastases identified the most important prognostic factors for recurrence to be node-positive primary colorectal cancer, a disease-free interval <12 months, more than one liver metastasis, largest liver metastasis >5 cm, and carcinoembryonic antigen (CEA) >200 ng/mL.31 In examining these factors in this small cohort of patients, only size >5 cm was significantly prognostic for recurrence. The other prognostic factors likely did not attain significance because of the sample size. Of note, no patient in this study had a CEA level >200 ng/mL. The median preoperative CEA level in this study’s patients was 13.5 ng/mL. Similar to our analysis with angiogenic factors, we correlated patients with low and high CEA levels (relative to the median level), and this was not predictive of recurrence (P > .05). Thus, analysis of angiogenic factors gave important prognostic information even compared with other more widely used prognostic variables.
The resection of healthy liver along with metastatic tumor adds some complexity in the analysis of circulating factors in the postoperative period. One source or stimulus for these factors, the tumor itself, is removed. Fujisaki et al.26 found that circulating levels of VEGF decreased after resection of primary colorectal cancers. However, liver resection and subsequent regeneration provides another stimulus for upregulation of angiogenic and growth factors. Circulating levels of HGF are known to be highly upregulated after partial hepatectomy.32 After liver resection for metastases, levels of circulating angiogenic factors will reflect a balance between the loss of the tumor stimulus and the addition of the liver-regeneration stimulus.
To our knowledge, this is the first study to examine a panel of circulating plasma angiogenic factors in human subjects at several time points after hepatectomy. HGF is produced by a wide variety of nonparenchymal liver cells, including Kupffer cells, Ito cells, endothelial cells, fibroblasts, and tumor cells.33 HGF may also enhance tumor cell growth, given that the HGF receptor is upregulated in many tumors, including colorectal cancers.30 In this study, plasma HGF increased dramatically in the first few days after hepatectomy and likely promotes the proliferation of hepatocytes and endothelial cells. Plasma bFGF did not peak until approximately 1 month after liver resection and stimulates continued new blood vessel formation, as well as stroma formation. Postoperative EGF and VEGF levels remained increased to near preoperative levels, except for a decrease in VEGF at 3 months.
Levels of bFGF and HGF at 3 months after hepatectomy were higher in patients who ultimately had disease recurrence compared with patients who did not. Only one patient with low bFGF and one patient with low HGF at 90 days had disease recurrence within the study period. There are at least two possible interpretations for this finding. Patients with higher bFGF and HGF levels may have had subclinical disease that caused increases in these factors. Alternatively, patients with higher levels of these factors may have an environment within the liver or in other metastatic sites that is conducive to the growth of micrometastatic disease.
Even 3 months after liver resection, levels of all four factors remained significantly higher than levels found in healthy controls. Approximately two thirds of patients with colorectal cancer liver metastases experience recurrence after potentially curative liver resection, and the most common site of recurrence after liver resection is the remnant liver.34 One could speculate that increases in angiogenic and growth factors after hepatectomy may promote the growth of residual micrometastatic disease35 and that adjuvant biological therapies in combination with chemotherapy may counteract the tumor-promoting milieu found in the posthepatectomy period. Oxaliplatin-based chemotherapy combined with bevacizumab (Avostin, San Francisco, CA), an anti-VEGF antibody that binds and neutralizes VEGF, is currently an effective strategy for patients with unresectable metastatic colorectal cancer.36,37 Cetuximab (Erbitux) is a monoclonal antibody that binds the EGF receptor and blocks its activation, and this agent has been found to be an effective therapy when used with irinotecan for irinotecan-refractory metastatic colorectal cancer.38,39 The combination of bevacizumab and cetuximab is currently being investigated for metastatic colorectal cancer.40 These and other targeted agents may counteract increases in angiogenic and growth factor levels in the posthepatectomy period and may be effective adjuvant treatments for patients who undergo resection of colorectal cancer liver metastases.
There are several analyses that were not performed in this study that could be the subject of future studies. First, the 20 control patients used in this study were healthy patients with no history of cancer. Analysis of other possible controls groups would add important information. For example, measurement of angiogenic factor levels in patients with benign liver lesions before and after liver resection would provide a nice comparison to the levels seen in our group of patients with metastatic colorectal cancer. Second, measurement of angiogenic factor levels beyond 3 months after hepatectomy in patients with colorectal cancer liver metastases could determine exactly how long it takes for these factors to return to normal in patients who do not experience recurrent disease. Finally, measurement of levels at the time of recurrence may determine whether these angiogenic factors could be used similarly to the CEA level in monitoring for recurrence.
In summary, this study examines a panel of four circulating angiogenic factors in patients with potentially resectable colorectal liver metastases. Before operation, levels of VEGF, bFGF, EGF, and HGF are all increased in these patients compared with controls, and higher preoperative levels of these factors predict subsequent recurrence. These factors remain increased for at least 3 months after partial hepatectomy: HGF levels peak at 3 days, and bFGF levels peak at 1 month after surgery. Higher postoperative increases of bFGF and HGF at 3 months also correlated with recurrence. The prolonged period of elevation in these and other factors after partial hepatectomy may promote the growth of subclinical disease. Thus, biological therapies that target these angiogenic and growth pathways should be investigated in the adjuvant setting for resectable colorectal cancer liver metastases.
Supported by National Institutes of Health grant K12CA87723-02 (S.S.Y.) and CA94503 (R.P.D.). The authors thank Rachel Jacobi, Beata Korytowski, and Brenda Guilfoyle for assistance with blood specimen collection and Sandra Ryeom, PhD for reviewing the manuscript.