OUP user menu

Lymphangiogenesis in Esophageal Adenocarcinomas--Lymphatic Vessel Density as Prognostic Marker in Esophageal Adenocarcinoma

Reda S. Saad MD, PhD, Jennifer L. Lindner DO, Yulin Liu MD, PhD, Jan F. Silverman MD
DOI: http://dx.doi.org/10.1309/AJCPKWUQSIPVG90H 92-98 First published online: 1 January 2009


We studied tumor lymphatic vascular density (LVD) as a predictive marker for the risk of lymph node (LN) metastasis and its relationship to other prognostic parameters and survival in 75 patients with esophageal adenocarcinoma. Samples were immunostained for D2-40, CD31, and vascular endothelial growth factor (VEGF). Microvessels were counted in densely vascular/lymphatic foci (hot spots) at ×400 field (0.17 mm2). Intensity of staining for VEGF was scored on a 2-tiered scale. CD31 microvessel counts showed significant correlation with tumor stage and patient survival (P < .01). D2-40 LVD demonstrated a significant correlation with LN metastases, lymphovascular invasion, and tumor stage (r = 0.45, r = 0.47, and r = 0.37, respectively) and with shorter disease-free survival. D2-40 detected lymphovascular invasion in 29 of 75 cases, more than with CD31 (23/75) and H&E (18/75). VEGF was expressed in 48 (64%) of 75 cases and was significantly correlated with lymphovascular invasion, LN metastases, and overall survival. Our study showed that angiogenesis and lymphangiogenesis have important roles in the progression of esophageal adenocarcinoma.

Key Words:
  • D2-40
  • Lymphangiogenesis
  • Angiogenesis
  • Vascular endothelial growth factor
  • VEGF
  • Esophageal adenocarcinoma

Barrett esophagus, defined as incomplete intestinal metaplasia, is a precancerous condition for esophageal adenocarcinoma, with approximately 12,000 new cases of esophageal carcinoma diagnosed each year in the United States.1,2 In comparison with other gastrointestinal tract cancers, esophageal adenocarcinoma is characterized by early lymphatic metastasis.3 The presence of lymph node (LN) metastasis is a poor prognostic indicator for patients with esophageal adenocarcinoma.4 After surgery, the 5-year survival is approximately less than 20% with positive LNs,5 whereas it is greater than 50% in patients without LN metastases.6

Angiogenesis, the formation of new blood vessels within the tumor, is essential for tumor growth by providing nutrients and eliminating metabolic waste products.79 In addition, neoangiogenesis helps the process of metastasis because newly formed tumor vessels have less basement membrane material and fewer intercellular junctional complexes, which result in increased permeability and provide a route of exit of tumor cells into the circulation.10 Angiogenesis has been proposed as a prognostic marker in a variety of human neoplasms, including esophageal adenocarcinoma.1113

Clinical and pathologic data support that detection of tumor cells in lymphatic vessels and regional LNs is an important early event in the metastasis of carcinomas and a key factor in the staging of human tumors and subsequent therapy.1,13,14 However, little is known regarding the mechanisms through which tumor cells gain entry into the lymphatic system.15,16 Some investigators suggested that the lymphatic vessels have a passive role, with tumor cells entering existing lymphatics.17 Conversely, other reports have suggested that new lymphangiogenesis, the formation of new tumor-associated lymphatics, has an active role in metastatic spread.1820

Only a few studies have investigated the potential clinical significance of lymphatic vessel density (LVD) as a prognostic marker in esophageal carcinoma, and they have reported contradictory results.2123 However, previous studies have been limited by the lack of specific lymphatic endothelial markers that could be used to discriminate between lymphatics and blood vessels.21,24 Recently, the monoclonal antibody D2-40, an oncofetal membrane antigen M2A identified in ovarian carcinoma cell lines and germ cell tumors,25 was reported to be a specific marker for lymphatic endothelium in normal and neoplastic tissue.26

In this study, we investigated tumor lymphangiogenesis by using monoclonal antibody D2-40 as a predictive marker for the risk of LN metastasis. We then correlated the results of D2-40 LVD, CD31 microvessel density, and vascular endothelial growth factor (VEGF) expression with other standard prognostic parameters such as size, grade, depth of invasion (pT), lymphovascular invasion, disease stage, and disease-free and overall survival.

Materials and Methods

Case Selection

We retrieved 75 cases of resected esophageal adenocarcinomas from the computer database of Allegheny General Hospital, Pittsburgh, PA. Patients had been treated by radical surgical resection—transthoracic or transhiatal esophagectomy—between January 1996 and December 2000. Postoperative adjuvant chemotherapy was received in all cases. All cases were reviewed by 2 pathologists (R.S.S. and J.L.L.), and the following features were recorded: tumor differentiation, depth of invasion, lymphovascular invasion, LN metastasis, distant metastasis, and tumor stage at the time of diagnosis. Disease-free and overall patient survival data were recorded from patients’ medical files. Complete follow-up information was available for all patients. Computed tomography scanning was performed in all cases as a part of clinical staging. None of the patients had received neoadjuvant therapy. Cases with metastatic lesions were subjected to surgical biopsy and/or fine-needle aspiration cytology to confirm the diagnosis of metastases. Postoperative pathologic staging was performed according to the American Joint Committee on Cancer staging system.27 The institutional review board of Allegheny General Hospital approved the study.

Tissue samples from the specimens were fixed in 10% buffered formalin, paraffin-embedded, and stained with H&E. H&E-stained slides of all cases were reviewed to confirm the diagnosis and evaluated for the presence of lymphovascular invasion. One paraffin block with the maximum bulk of tumor and maximum depth of invasion was chosen from each case for immunohistochemical studies.

Immunohistochemical Analysis

For immunohistochemical studies, 4-μm-thick sections from the paraffin-embedded tissue blocks were deparaffinized in xylene and rehydrated in a descending ethanol series. Sections were incubated for 5 minutes in 3% hydrogen peroxide to quench endogenous tissue peroxidase. The sections were immunostained for a monoclonal antibody (clone D2-40, Signet Laboratories, Dedham, MA) at a 1:50 dilution for D2-40, a mouse monoclonal antibody (clone 1A10, Cell Marque, Hot Springs, AR) for CD31, and a rabbit polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at a 1:500 dilution for VEGF. Heat-induced epitope retrieval techniques were used for antigen retrieval as follows: citrate buffer (pH 6.0) and a steamer at 120°C for 5 minutes for D2-40 and CD31 and 95°C for VEGF. The Ventana Immunostainer and iView 3-3′-diaminobenzidine detection kit (Ventana Medical Systems, Tucson, AZ) were used. Appropriate positive and negative control experiments were run for each batch of slides.


The microvessel count was assessed by a modification of the method described by Weidner et al.28 After scanning the immunostained section at low magnification, 3 highly vascular areas of carcinoma (hot spots) were selected by 2 observers (R.S.S. and J.L.L.) at the same time. In the absence of apparent hot spots, 3 or more randomly selected areas were counted. For lymphatic density, 5 to 10 random fields were counted at the periphery and in the center of each tumor. Intratumoral lymphatic vessels were defined as vessels within the main tumor mass, surrounded by tumor cells. Vessels more than 1 high-power field (×250; field diameter) away from the invasive tumor front were not counted. Any immunostained cells or separate clusters of endothelial cells, with or without an identifiable lumen, were considered and counted as a single vessel.29 The 2 observers then independently evaluated the slides for microvessel count by using ×400 magnification (0.17 mm2 field) without knowledge of patient status and stain used. Discordant cases were recounted, and consensus resolved any discrepancy of more than 10% in the microvessel count. Maximum values were used for statistical analysis. For VEGF staining, intensity was recorded as negative (<20% staining) or positive (≥20% staining).

Statistical Analysis

Survival data for patients were obtained from the laboratory information system of Allegheny General Hospital. Disease-free survival was calculated from the date of surgery to the date of progression or the date of last follow-up. Death of esophageal cancer was the terminal event for survival calculations. Patients who died within 30 days from operation were excluded from the study. Postoperative survival curves were constructed using the Kaplan-Meier method. Correlation with other prognostic factors (ie, age, sex, tumor differentiation, tumor size, depth of invasion, lymphovascular invasion, and stage) was evaluated by using the Spearman correlation. Mean differences in microvessel counts were compared with the use of paired t tests. All statistical analyses were performed with SPSS (Statistical Package for Social Sciences, SPSS, Chicago, IL).


Table 1 shows the clinical data for patients with esophageal adenocarcinoma included in this study. There were 61 men and 14 women. At the end of the study, 31 patients remained alive without disease, 9 were alive with disease (recurrence or residual tumor), and 35 patients died of disease. Follow-up distant metastases were identified in 30 cases, including 12 in lung, 6 in liver, 5 in bone, 2 in peritoneum, and 1 in spine and 4 cases with widespread metastases. Survival decreased with increasing age, which was statistically significant (r = 0.42; P < .01). For clinical parameters, there was a significant correlation among depth of invasion and LN and distant metastases (r = 0.37 and r = 0.31, respectively; P < .05).

View this table:
Table 1

Clinical Data for 75 Patients With Esophageal Adenocarcinoma*

Age (y)
  Mean62.7 ± 11.3
Mean ± SD follow-up (mo)27.3 ± 10.2
Mean ± SD disease-free survival (mo)19 ± 9
Tumor grade
  Well-differentiated14 (19)
  Moderately differentiated51 (68)
  Poorly differentiated10 (13)
Distant metastases (during follow-up)30 (40)
Lymphovascular invasion
  H&E18 (24)
  CD3123 (31)
  D2-4029 (39)
Lymph node metastases44 (59)
  I12 (16)
  II28 (37)
  III30 (40)
  IV5 (7)
  • * Data are given as number (percentage) unless otherwise indicated.

CD31 stained intratumoral microvessels in 71 (95%) of 75 cases (negative in 4 cases) Image 1. D2-40 stained peritumoral lymphatic vessels in 73 (97%) of 75 cases and intratumoral lymphatic vessels in 33 (44%) of 75 cases Image 2. The mean ± SD number of microvessels identified with CD31 was significantly different from the mean number of peritumoral lymphatic vessels identified by D2-40 (19 ± 9 vs 11 ± 6/0.17 mm2 field, respectively; P < .05), without significant correlation. D2-40 stained only lymphatic endothelial cells, whereas endothelial cells of blood vessels were negative.

Image 1

CD31-stained microvessels in esophageal adenocarcinoma (×400).

Image 2

D2-40 expression in esophageal adenocarcinoma. Note the strong staining of D2-40 for lymphatics in peritumoral (A) and intratumoral (B) locations. The intratumoral lymphatic microvessels are linear and collapsed (×400).

Table 2 and Table 3 show our results by marker and stage and the statistical analysis of our findings. CD31 and D2-40 showed significant correlation with tumor stage (r = 0.44 and r = 0.37, respectively; P < .05). Only D2-40 lymphatic vessel counts correlated significantly with lymphovascular invasion and LN metastases (r = 0.47 and r = 0.45, respectively; P < .05). In multivariate analysis, D2-40 correlated significantly with lymphovascular invasion and LN metastases, independent of tumor stage. No significant correlation was identified between D2-40 lymphatic vessel counts and tumor size (r = 0.18; P > .05 for D2-40), but CD31 showed borderline correlation (r = 0.28; P = .07). The CD31 microvessel count demonstrated a significant correlation with depth of invasion (r = 0.36; P < .01), whereas the D2-40 microvessel count was insignificant (r = 0.21; P > .05). Neither of our markers (CD31 and D2-40) showed a significant correlation between microvessel counts and grading of the tumor (r = 0.02; P = .86 for CD31; r = 0.13; P = .33 for D2-40). D2-40 stained intratumoral lymphatics in 33 (44%) of 75 cases, all with an infiltrative growth pattern. The presence of intratumoral lymphatics was significantly correlated with the presence of lymphovascular invasion (r = 0.36; P < .01). D2-40 detected lymphovascular invasion in 29 (39%) of 75 cases Image 3, and CD31 detected lymphovascular invasion in 23 (31%) of 75 cases.

View this table:
Table 2

Spearman Correlation Between Mean ± SD D2-40 and CD31 Microvessel Densities and VEGF Expression With Other Prognostic Markers in Patients With Esophageal Adenocarcinoma*

D2-40 (Vessel Density, 11 ± 6/0.17 mm2)CD31 (Vessel Density, 19 ± 9/0.17 mm2)VEGF (+, 64%; –, 36%)
Tumor size0.18 (.15)0.28 (.07)0.20 (.12)
Tumor grade0.13 (.33)0.02 (.86)0.13 (.33)
Depth of invasion0.21 (.13)0.36 (<.01)0.27 (.09)
Lymph node metastases0.45 (<.01)0.30 (.06)0.39 (<.01)
Lymphovascular invasion0.47 (<.01)0.10 (.35)0.48 (<.01)
Distant metastases0.08 (NS)0.45 (<.01)0.48 (<.01)
Tumor stage0.37 (<.05)0.44 (<.01)0.47 (<.01)
  • VEGF, vascular endothelial growth factor.

  • * Data are given as r value (P value).

  • Significant.

  • Borderline correlation.

View this table:
Table 3

Lymphovascular Invasion, Microvessel Densities, and VEGF Expression in Different Tumor Stages*

Stage I (n = 12)Stage II (n = 28)Stage III (n = 30)Stage IV (n = 5)
Age (y)61 ± 1161 ± 766 ± 1159 ± 3
CD31 vessel count13 ± 625 ± 825 ± 727 ± 4
D2-40 lymphatic count8 ± 68 ± 512 ± 812 ± 5
VEGF expression4/12 (33)15/28 (54)24/30 (80)5/5 (100)
Follow-up (mo)21 ± 8.515 ± 716 ± 817 + 7
LVI (D2-40)3/12 (25)6/28 (21)16/30 (53)4/5 (80)
  • LVI, lymphovascular invasion; VEGF, vascular endothelial growth factor.

  • * Data are given as mean ± SD or number/total (percentage).

Image 3

D2-40 stains lymphatic vessels and predicts increased lymphovascular invasion. Note the blood vessels are negative for D2-40, indicating D2-40 specificity for lymphatic vessels (×400).

Kaplan-Meier survival analysis showed that the CD31 microvessel count had an impact on the survival curve, with low microvessel counts indicating better survival (P < .01; log rank) compared with cases with high microvessel counts. For D2-40, there was no significant correlation between D2-40 lymphatic vessel counts and overall survival. However, in the angiogenic group (with high microvessel staining by CD31), the D2-40 lymphatic vessel count showed significant correlation (P < .05; log rank). In addition, there was a significant correlation between high D2-40 lymphatic vessel counts and shorter disease-free survival Figure 1. Lymphovascular invasion detected by D2-40 showed a negative impact on overall and disease-free survival but was not statistically significant.

Figure 1

Kaplan-Meier survival curve for patients with esophageal adenocarcinoma. D2-40 lymphatic vessel counts show significant impact on disease-free survival in patients (using the median as cutoff). HPF, high-power fields.

High VEGF expression was identified in 48 (64%) of 75 cases and low or negative expression in 27 (36%) of 75 cases. VEGF showed a significant correlation with LN status (r = 0.39; P < .01), distant metastases (r = 0.48; P < .01), and tumor stage (r = 0.47; P < .01). No significant correlation was found between VEGF expression and grade or size of the tumor (r = 0.13 and r = 0.20, respectively; P > .05).

For stage I disease, VEGF expression showed significant correlation with disease-free and overall survival (P < .05), whereas CD31 microvessel and D2-40 lymphatic counts were not significant.


The lymphatic system is the primary pathway of metastasis for most human cancers, and the extent of LN involvement is a crucial prognostic factor for patient outcome. Recently, antibodies specific for lymphatic endothelium, such as D2-40, have become available, providing important new insights into the process of tumor-associated lymphangiogenesis and its possible clinical relevance.25,26,30 In this study, we investigated the D2-40 LVD as a potential marker for LN metastases in a series of esophageal adenocarcinoma cases.

Previous studies reported that solid tumors do not have intratumoral lymphatic vessels16 owing to the increased interstitial pressure created by the proliferating cancer cells, whereas lymphatics present at the tumor margin can facilitate spread of tumor cells.31,32 However, Clarijs et al33 reported newly formed lymphatic vessels within the solid tumors using double staining with CD31 and PAL-E. Others have recently observed intratumoral lymphatics in melanoma34 and breast,35 head and neck,36,37 colonic,30,38 and pancreatic39 carcinomas. One study demonstrated a significantly higher number of intratumoral lymphatics in metastatic vs nonmetastatic melanomas and a significant correlation with poor disease-free survival.34 Similar to previous studies, we found intratumoral lymphatic vessels in esophageal adenocarcinoma in 44% of cases using D2-40. In addition, most intratumoral lymphatics were small, flattened, and more difficult to see owing to high interstitial pressure in tumors,31,40 contrasting with the widely open peritumoral lymphatics. In the present study, the presence of intratumoral lymphatics was significantly correlated with the presence of lymphovascular invasion. It is believed that intratumoral lymphatic vessels are likely to have an important role in absorbing tissue fluid containing metabolites excreted by tumor cells and in maintaining an appropriate environment for tumor cells.24 In addition, they may have an active role in tumor metastases.

Few studies have investigated the role of lymphangiogenesis in esophageal carcinoma and lack clear consensus on the usefulness of LVD as a prognostic marker.2123,41,42 Only 2 of these studies investigated the role of lymphangiogenesis in esophageal adenocarcinoma. One study demonstrated VEGF-C expression in Barrett epithelium as it progressed through dysplasia to adenocarcinoma with a similar trend in VEGFR-3 expression in lymphatic vessels.23 However, LVD was not measured in the study. The other study used LYVE-1 and demonstrated no significant correlation with prognostic parameters and clinical outcome, whereas entry of tumor cells into preexisting peritumoral lymphatic vessels confers a significantly worse overall survival.22 Our results demonstrated that by using D2-40, peritumoral LVD in esophageal adenocarcinoma correlated significantly with the presence of lymphovascular invasion and LN metastases. Moreover, LVD by D2-40 showed a statistically significant correlation with short disease-free survival. In multivariate analysis, the D2-40 LVD correlated significantly with lymphovascular invasion and LN metastases, independent of tumor stage.

The contradicting results about the role of lymphangiogenesis in tumor progression may be due to differences in case selection, methods, and/or the types of tumors included in the analyses. Tumor lymphangiogenesis and lymphatic metastasis are complex mechanisms that can differ significantly in tumors of different types or anatomic locations.36 However, the main reason for the controversial results may be the previous lack of specific immunohistochemical markers for lymphatic vessels. Although it is claimed that LYVE-1 is a specific lymphatic marker, its value in identifying fully functional lymphatics has been questioned. Two recent studies failed to detect lymphatics within non–small cell lung carcinoma using LYVE-1,43,44 whereas they were identified using podoplanin.44 LYVE-1 and VEGFR-3 are not completely specific for lymphatic endothelium in the tumor because they also stain endothelial blood vessels.15,44

Studies have indicated that D2-40 can be useful in identifying the presence of lymphatic invasion in various malignant neoplasms.45 D2-40 outlined tumor emboli in lymphatics otherwise indiscernible by H&E and increased the detection rate of lymphatic invasion by 16% in melanoma patients.46 The same investigators observed a trend in the association between lymphatic invasion and other prognostic indicators. Our results showed that D2-40 immunostaining is a useful immunohistochemical marker for identifying the presence of lymphatic invasion. D2-40 strongly labeled lymphatic endothelial cells, highlighting the presence of lymphatic invasion in the tumors. In our study, D2-40 increased the detection rate of lymphovascular invasion by 15% over H&E. Similar to a previous study,46 we showed a significant correlation between lymphatic invasion detected by D2-40, tumor depth of invasion, and LN and distant metastases.

Previous studies have demonstrated a significant correlation of VEGF expression and other prognostic parameters and patient outcome in esophageal squamous cell carcinoma.4750 However, there are few available studies regarding VEGF expression in Barrett dysplasia and adenocarcinoma, and they failed to give prognostic correlation.51 In the present study, the expression of VEGF was found in 64% of cases and was significantly correlated with a higher microvessel count. Our study showed that VEGF is a useful prognostic marker that is significantly correlated with lymphovascular invasion, LN status, distant metastases, and overall survival.

Angiogenesis and lymphangiogenesis have important roles in the progression of esophageal adenocarcinoma. Peritumoral lymphangiogenesis in esophageal adenocarcinoma correlates with lymphovascular invasion and LN metastases. In addition, D2-40 increases the frequency of detection of lymphatic invasion relative to conventional H&E staining and the panendothelial marker CD31. Therefore, assessing LVD with D2-40 in esophageal adenocarcinoma could be a valuable parameter for predicting patients with an increased risk of metastatic disease.


  • Presented in part as poster at the Annual Meeting of the United States–Canadian Academy of Pathology; March 2007; San Diego, CA.


View Abstract