| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2004;78:1944-1949
© 2004 The Society of Thoracic Surgeons
a Department of Thoracic SurgeryRoyal Victoria Hospital, Belfast, Northern Ireland
b Department of Pathology, Royal Victoria Hospital, Belfast, Northern Ireland
Accepted for publication March 31, 2004.
* Address reprint requests to Dr Spence, Department of Thoracic Surgery, Royal Victoria Hospital, Belfast BT12 6BA, Northern Ireland
gary.spence{at}virgin.net
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
METHODS: Bone marrow aspirates from 49 patients with esophageal cancer were assessed and assigned to be positive or negative for micrometastases. Routine histologic assessment of the primary tumor was also undertaken. Circulating and tumor levels of the angiogenic cytokine vascular endothelial growth factor were determined in plasma and tumor homogenate. Intratumor microvessel density was evaluated by counting anti-CD34 positive neovessels.
RESULTS: Twenty-two patients were positive for bone marrow micrometastases (44.9%). The detection of micrometastases was associated with advanced T stage (T3/4 vs T1/2; p = 0.023), circumferential margin involvement (p = 0.002) and lymphovascular invasion (p = 0.024). Plasma vascular endothelial growth factor was significantly more elevated in micrometastatic-positive patients than in those without micrometastases (p = 0.018). No difference was noted in tumor vascular endothelial growth factor expression. For adenocarcinomas alone, intratumor microvessel density was significantly higher in micrometastatic positive cases (p = 0.03). This was not the case for squamous cell carcinomas.
CONCLUSIONS: The detection of bone marrow micrometastases is associated with esophageal tumors of advanced T stage and specifically for adenocarcinomas with tumor vascularity. Plasma vascular endothelial growth factor is elevated in micrometastatic positive cases and might be derived from sources other than the primary tumor.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Numerous animal and clinical human studies correlate microvascular density [3] and angiogenic growth factor expression [4] with the development of metastasis and poor patient outcome. Presently the most widely accepted marker of tumor angiogenesis is quantification of the number of microvessels per unit area on a representative histopathologic section [5]. This measure is known as the intratumor microvessel density (IMD).
Vascular endothelial growth factor (VEGF) is an angiogenic cytokine that has a pivotal role in regulating tumor angiogenesis [6]. Quantitative tissue expression of VEGF is related to IMD and prognosis in esophageal [7] and gastric cancers [8]. VEGF is also detectable in the circulation of patients with cancer and may be a reliable, repeatable surrogate marker of angiogenic activity and tumor progression [4].
The systemic dissemination of tumor cells is an essential prerequisite for metastasis formation. Subclinical metastatic disease, undetectable by conventional staging and imaging techniques, may be present at the time of diagnosis [9, 10]. The detection of cancer cells in bone marrow samples suggests the existence of otherwise undetectable systemic micrometastases [11] and might indicate patients at risk of developing clinically important metastatic disease [12, 13].
Cancers arising in the esophagus and at the esophagogastric junction are increasing in prominence in the United States and Europe. This is due to the rising incidence of the adenocarcinoma cell type [14, 15]. Surgery remains the mainstay of treatment, but the prognosis for patients with esophageal cancer treated by surgery alone is poor. The 5-year survival rate remains disappointingly low, averaging 20% in most series [16]. In addition to conventional chemotherapy and radiation therapy, novel therapies such as those that target the tumor vasculature are being developed [17].
Disease-free and overall survival has been shown to be significantly shorter in esophageal cancer patients who have bone marrow micrometastases (BMMs) [18, 19]. Although still being debated, patient staging that is based on the presence or absence of bone marrow micrometastases (BMM+/BMM–) may be a valuable early marker of poor outcome, against which novel tumor-derived molecular prognosticators may be assessed.
In two published clinical studies that investigated the relationship between angiogenesis and the detection of disseminated tumor cells in bone marrow—one in patients with breast cancer [20] and the other in gastric cancer [21]—an association was discovered between elevated IMD and the presence of bone marrow micrometastases. The significance of angiogenesis in micrometastasis development remains incompletely assessed for patients who have cancer of the esophagus. We therefore examined the bone marrow of such patients for the presence of cancer cells to see if a relationship existed between indicators of a tumor's angiogenic status and the presence of BMMs. The angiogenic markers measured included IMD, tumor VEGF levels (T-VEGF), and plasma VEGF (P-VEGF) levels.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Patients were excluded if they had a past history of malignant disease or of an active angiogenic-related disease process. These included inflammatory arthritis, psoriasis, active ulceration, and critical cardiovascular disease. Patients were also excluded if neoadjuvant chemotherapy had been administered before surgery.
Clinical Staging Protocol
All patients underwent clinical staging by physical examination, upper gastrointestinal endoscopy, chest roentgenogram, and contrast-enhanced computed tomography (CT) of the thorax and upper abdomen. Bronchoscopy, isotope bone scanning, and laparoscopy were used selectively, depending on history, clinical examination, and the results of CT.
Operative Intervention
Resection with curative intent was achieved in 42 patients (85.7%). Twenty-three underwent a total thoracic esophagectomy through a left thoracolaparotomy that incorporated a cervical anastomosis. A further 7 patients underwent an esophagogastrectomy through a left thoracolaparotomy only. Two patients had a three-stage esophagectomy, 8 had a two-stage esophagectomy and 2 had a transhiatal esophagectomy. No tissue was obtained from the 7 patients who underwent exploratory surgery only.
Pathologic Staging Protocol
The pathologic stage of disease was classified according to the Union Internationale Contre le Cancer tumor node metastasis (TNM) classification and current stage groupings [22].
Retrieval of Tissue Samples
PRIMARY TUMOR
Tissue blocks were randomly obtained from each tumor specimen. Half of each block was placed in 10% unbuffered formal saline for 24 hours before being embedded in paraffin wax (Surgipath Medical Industries Inc, Richmond, IL). The other half of each block was placed in polypropylene tubes (Nalge Nunc Laboratory Supplies & Instruments, Hereford, UK), snap-frozen and stored in liquid nitrogen.
Because of small tumor size, the material available from two patients was inadequate to facilitate bisection of tumor blocks. Consequently the complete block was processed solely for paraffin wax embedding.
The presence of tumor tissue in each block was confirmed by histologic assessment of a representative tissue section.
BONE MARROW
Bone marrow aspiration was performed with a single-use bone marrow aspiration needle (Allegiance Healthcare Corporation, McGaw, IL) under general anesthesia at the time of surgery. A single bone marrow aspirate was obtained from the left posterior iliac crest before tumor manipulation. Ten milliliters of bone marrow was aspirated into a syringe containing 2500 IU heparin and added to 10 mL phosphate buffered saline for preparation of cytospins.
PLASMA
Ten mL of peripheral venous blood was obtained from each patient the day before surgery and immediately added to 0.25 mL sodium ethylenediamine tetra-acetic acid inhibit clotting. Without delay, the blood sample was transported on ice directly to the laboratory where it underwent centrifugation at 13,600g for 3 minutes at 4°C to separate the plasma and cell fractions. The plasma was removed and stored at –80°C until analysis. The time between venesection and plasma storage was consistent and prompt for all patients.
Detection of Bone Marrow Micrometastases
PREPARATION OF BONE MARROW CYTOSPINS
Cytospins were prepared by density gradient centrifugation through Ficoll-Hypaque (Pharmacia, Freiburg Germany) [23]. Resuspended bone marrow samples were adjusted to a final concentration of 1 x 106 nucleated cells per 150 µL before being spun onto aminopropyltriethoxysilane coated slides and stored in 95% alcohol at room temperature until immunocytochemical assessment.
CYTOSPIN ASSESSMENT BY IMMUNOCYTOCHEMISTRY
Cytokeratins are proteins that comprise the epithelial cell cytoskeleton [24]. Anticytokeratin antibodies can be used in the detection of micrometastatic disease in breast, prostatic, and gastrointestinal malignancy [24].
Cytospins were stained with anticytokeratin antibodies, CAM 5.2 at 2.5 µg/mL (BDIS, San Jose, CA) and AE1/AE3 at 0.214 µg/mL (Dako, Carpenteria, CA) using alkaline phosphatase-antialkaline phosphatase methodology (Dako). Two cytospins (2 x 106 cells) were screened for each antibody. Cytokeratin-positive cells were identified by light microscopy and confirmed to have morphologic characteristics of malignant cells by an experienced cytopathologist (WGMcC).
Intratumor Microvessel Density
Sections from each primary tumor block were immunostained with anti-CD34 clone number QBEND-10 (Serotec, Oxford, UK), an endothelial cell marker, at 0.2 µg/mL using a standard streptavidin-biotin complex immunoperoxidase method for paraffin wax-embedded sections (Dako).
The tumor microvessel count was assessed using "hotspot" methodology after the technique described by Graham and colleagues [23]. IMD (expressed as microvessels per mm2) was then derived by dividing the microvessel count by the fixed field size (0.41 mm2).
T-VEGF
VEGF protein expression in esophageal tumor blocks was measured using the technique described by Crew and colleagues [25] optimized for use in our laboratory.
Briefly, snap-frozen nonfixed tumor blocks were homogenized. After ultracentrifugation, the resulting cell cytosolic fraction was quantitatively assayed for T-VEGF using a commercially available enzyme linked immunosorbent assay (ELISA) kit specific for Quantikine human VEGF (R & D Systems, Minneapolis, MN).
T-VEGF levels were standardized for total protein concentration to allow for the effect of varying sized tumor blocks. Total protein was quantified using a technique based on the Bradford colorimetric method for total protein quantification [26]. A commercial protein assay reagent was employed (Coomassie Plus Protein Assay Reagent, Pierce, IL). Results were expressed as pg VEGF/mg total protein.
P-VEGF
P-VEGF levels were measured as described in a previous report [27] by using the same commercial ELISA kit that was used to measure T-VEGF. P-VEGF concentrations are expressed as pg/mL.
Statistical Analysis
Statistical analysis was undertaken using SPSS version 9 for Windows (SPSS Inc, Chicago, IL). Graphs were constructed using GraphPad Prism version 3.0 for Windows (GraphPad Software, San Diego, CA). Data are expressed as mean (SEM) or median {IQR}, where appropriate. Categorical variables were compared with the 
2 test unless, where stated, by the Fisher exact test. Continuous variables that followed a parametric distribution were compared using the independent samples t test. Nonparametric data were compared using the Mann-Whitney U test. Correlation analysis was undertaken using Spearman rank correlation. Statistical significance was accepted if p was less than 0.05.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
BMM and Pathological Disease Stage
There was no difference in the size of the primary tumor between those patients who were BMM+ and those who were BMM–; the median tumor area was 9 {7–18} cm2 vs 9 {2.7–21} cm2, respectively ( z = –0.835, p = 0.404, Mann Whitney U test). The incidence of disseminated tumor cells in bone marrow and their relationship to primary tumor staging, histopathologic features, and margin involvement of the resected tumor specimen are shown in Tables 1, 2, and 3, respectively.
|
|
|
A study of lung cancer showed that adenocarcinomas have higher IMD than squamous cell cancers [23]. We have demonstrated similar findings in esophageal cancer comparing adenocarcinoma and squamous cell (413 (29) vs 193 (14) microvessels/mm2, respectively; t = 3.844, p < 0.001, independent samples t test with 40 df).
Consequently, we decided to analyze the influence of IMD on bone marrow status for each cell type in turn. High tumor vascularity is associated with detection of BMMs for adenocarcinomas (Fig 1). This is not seen for squamous cell carcinomas (Fig 1).
|
P-VEGF and BMM
Compared with BMM–patients, P-VEGF was significantly elevated in the 22 esophageal cancer patients (out of 49) who had micrometastases detectable in their bone marrow (Fig 2).
|
= –0.117, p = 0.473, Spearman rank test). |
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
The BMM detection rate was found to be increased in patients who had tumors of advanced T stage (Table 1) and in those who had circumferential margin involvement (Table 3). In contrast, BMM+ patients did not have tumors of larger mucosal surface area than BMM– patients, nor was there a relationship between lymph node positivity and detection of BMMs (Table 1).
Similar findings have been described previously in breast cancer [29]. Braun and colleagues found that the detection of BMMs was associated with advanced T stage but not lymph node involvement [29]. They postulated that alternative pathways exist for tumor cell dissemination that might result in distinct patterns of metastasis [29].
As such, it may be speculated that two distinct routes to metastasis also exist for esophageal cancer. Tumor cells may disseminate initially to the lymphatics or directly by way of the microcirculation. The collective importance of both pathways as a means to tumor cell dissemination is confirmed by the finding that BMMs are more common in patients with esophageal tumors that have lymphatic or vascular invasion noted on conventional histologic analysis. Indeed, no case in which lymphovascular invasion was absent had micrometastases detected in the bone marrow (Table 2).
With regard to the association between tumor angiogenic markers and detection of BMMs, the study findings are more equivocal. Although a correlation between IMD and BMMs could be demonstrated for esophageal adenocarcinomas (Fig 1), IMD disparity between cell types, insufficient numbers in the squamous cell subgroup, and biologic variation are all potential reasons why BMMs were not associated with IMD when all patients were collectively assessed. Another possibility is that metastasis formation is a complex multistage process and angiogenesis may not be the rate-limiting step to tumor cell dissemination in squamous cell cancers. Some other component of the metastatic process, such as loss of cell–cell adhesion or degradation of the extracellular matrix, may hold that position.
The positive correlation between high tumor vascularity and the presence of BMMs (as seen for esophageal adenocarcinomas) is postulated to be because of a larger endothelial surface area with which potentially metastatic cells can interact. Supporting this theory is the finding of a positive association between tumor microvessel density and the number of circulating tumor cells detected intraoperatively in patients undergoing breast cancer surgery [30].
No difference in T-VEGF between BMM+ and BMM– cases was found. Although perhaps unexpected, it should not be surprising when one considers that alternative angiogenic factors may be involved in mediating tumor angiogenesis for patients with esophageal carcinoma. It is possible that VEGF is not the only, or even the pre-eminent, angiogenic growth factor responsible for neovascularization of primary esophageal tumors. Other cytokines such as thymidine phosphorylase [31] and transforming growth factor-
[32] may be more critical.
In contrast, VEGF measured in the plasma of BMM+ esophageal cancer patients was significantly higher than that measured in the plasma of those who were BMM– (Fig 2). It is however, counter-intuitive that P-VEGF can influence detection of BMMs as a function of primary tumor angiogenesis when T-VEGF levels clearly do not.
Thus, it may be proposed that the association between P-VEGF and bone marrow positivity might simply relate to a common correlation with advancing tumor T stage. We have not however, found this to be the case. Although it has been demonstrated that patients with advanced primary tumors are more likely to be BMM+ (Table 1), advanced T stage tumors do not have higher levels of P-VEGF (T1/T2 vs T3/T4, 13 {0–23} vs 18 {0–27} pg/mL, respectively; z = –0.262, p = 0.793, Mann Whitney U test).
Another explanation might be that in addition to the primary tumor, VEGF in the circulation may be derived, not only from nontumor-derived circulating elements such as platelets [27] but also from alternative sources such as systemically dispersed micrometastatic tumor deposits. This might also explain why no correlation was found between tumor and circulating VEGF levels. Micrometastases remain in a dormant state, during which proliferation and apoptosis are in balance [33]. A balance of proangiogenic and antiangiogenic factors may maintain the dormancy of micrometastases [34]. VEGF may be one such angiogenic stimulator secreted by micrometastatic tumor cells [34]. Thus disseminated micrometastatic cells may have a differing angiogenic cytokine profile to the primary tumor and in combination with platelets, which are known to be potent source of circulating VEGF [27], may contribute to P-VEGF levels.
This study provides evidence that for esophageal adenocarcinomas, systemic tumor cell dissemination is positively associated with primary tumor neovascularity. Furthermore, we speculate that high circulating VEGF levels in the plasma of esophageal cancer patients may be linked to sources other than the primary tumor.
The detection of micrometastases in the bone marrow may in itself be a useful adjunct to presurgical and postsurgical staging. Although not discriminatory with regard to resectability in this study, the detection of BMMs at diagnosis may influence the surgeon to consider preoperative neoadjuvant therapy before surgical intervention. Proponents postulate that the detection of BMMs at the time of surgery indicates the presence of a metastatic phenotype. Others argue that such detection merely reflects transient shedding of cells from the primary tumor and does not indicate increased metastatic potential [35]. For patients in this study, long-term follow-up will be required to clarify if the detection of bone marrow micrometastases at the time of surgery has independent prognostic significance.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
