Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Lymphangiogenic Markers and Their Impact on Nodal Metastasis and Survival in Non-Small Cell Lung Cancer - A Structured Review with Meta-Analysis

  • Thomas K. Kilvaer ,

    kilvaer@gmail.com

    Affiliations Department of Oncology, University Hospital of North Norway, Tromso, Norway, Institute of Clinical Medicine, UiT The Arctic University of Norway, Tromso, Norway

  • Erna-Elise Paulsen,

    Affiliations Department of Oncology, University Hospital of North Norway, Tromso, Norway, Institute of Clinical Medicine, UiT The Arctic University of Norway, Tromso, Norway

  • Sigurd M. Hald,

    Affiliation Institute of Clinical Medicine, UiT The Arctic University of Norway, Tromso, Norway

  • Tom Wilsgaard,

    Affiliation Department of Community Medicine, UiT The Arctic University of Norway, Tromso, Norway

  • Roy M. Bremnes,

    Affiliations Department of Oncology, University Hospital of North Norway, Tromso, Norway, Institute of Clinical Medicine, UiT The Arctic University of Norway, Tromso, Norway

  • Lill-Tove Busund,

    Affiliations Department of Clinical Pathology, University Hospital of North Norway, Tromso, Norway, Institute of Medical Biology, UiT The Arctic University of Norway, Tromso, Norway

  • Tom Donnem

    Affiliations Department of Oncology, University Hospital of North Norway, Tromso, Norway, Institute of Clinical Medicine, UiT The Arctic University of Norway, Tromso, Norway

Lymphangiogenic Markers and Their Impact on Nodal Metastasis and Survival in Non-Small Cell Lung Cancer - A Structured Review with Meta-Analysis

  • Thomas K. Kilvaer, 
  • Erna-Elise Paulsen, 
  • Sigurd M. Hald, 
  • Tom Wilsgaard, 
  • Roy M. Bremnes, 
  • Lill-Tove Busund, 
  • Tom Donnem
PLOS
x

Abstract

Background

In non-small cell lung cancer (NSCLC), nodal metastasis is an adverse prognostic factor. Several mediating factors have been implied in the development of nodal metastases and investigated for predictive and prognostic properties in NSCLC. However, study results differ. In this structured review and meta-analysis we explore the published literature on commonly recognized pathways for molecular regulation of lymphatic metastasis in NSCLC.

Methods

A structured PubMed search was conducted for papers reporting on the expression of known markers of lymhangiogenesis in NSCLC patients. Papers of sufficient quality, presenting survival and/or correlation data were included.

Results

High levels of vascular endothelial growth factor C (VEGF-C, HR 1.57 95% CI 1.34–1.84) and high lymphatic vascular density (LVD, HR 1.84 95% CI 1.18–2.87) were significant prognostic markers of poor survival and high expression of VEGF-C, vascular endothelial growth factor receptor 3 (VEGFR3) and LVD was associated with lymph node metastasis in NSCLC.

Conclusion

Lymphangiogenic markers are prognosticators of survival and correlate with lymph node metastasis in NSCLC. Their exact role and clinical implications should be further elucidated.

Introduction

Lung cancer is estimated to have the second highest incidence of all cancers in US women and men [1]. With a dismal prognosis, and over 163 000 expected deaths in 2014, lung cancer is the number one killer amongst cancers [1]. Lung cancer is staged according to the system advocated by the American Joint Committee on Cancer (AJCC) and the Union Internationale Contre le Cancer (UICC) which as of 2010 is in its 7th edition [2,3]. The system is built around tumor size and localization (T), extent of nodal involvement (N) and presence of distant metastasis (M) [3]. In short, stage I comprises small localized tumors, while stages II-IIIA represent larger tumors with or without nodal metastasis [3]. The five-year-survival for stage IA, IB, IIA, IIB and IIIA lung cancer is 65–81%, 54–72%, 46–59%, 46–47% and 33–38% respectively [46]. Stage IIIB and IV lung cancers are not considered for surgery and have an abysmal prognosis [46]. Non-small cell lung cancer (NSCLC) represent >80% of all lung cancer cases [7].

Nodal metastasis represents a major shift in NSCLC biology, from a localized to an invasive phenotype. Patients presenting with nodal metastases have a more advanced stage and a worse prognosis compared to patients without nodal involvement and the same tumor size [3]. Further, the prognosis of patients radically operated for stage I NSCLC differs widely, with 20–35% of patients developing recurrent disease, often in localized lymph nodes [46,8]. Vascular endothelial growth factors (VEGF) -C and -D, and their corresponding receptor vascular endothelial growth factor receptor 3 (VEGFR3, also known as Flt4), are by many considered the main players in the development of tumor associated lymphatic vessels [9]. These work by recruiting endothelial cells (ECs) and other stromal cells to develop and maintain a crude lymphatic network in the tumor micro-environment [9,10]. In tumor models of NSCLC, the presence of VEGF-C and VEGFR3 leads to proliferation, invasiveness and nodal metastases [11]. There is also evidence supporting that tumor derived VEGF-C induces the development of lymphatic vasculature in premetastatic lymph nodes; thus preparing them for the arrival and hosting of cancer cells [12,13]. The inhibition of VEGFR3 in a xenograft model of NSCLC indicated that abrogation of lymphangiogenesis could prevent lymphatic metastasis [14]. Interestingly, anti-VEGFR3 treatment had to be initiated before the development of lymphatic vasculature for its effect to take place and it did not abrogate lymphatic vessel co-option [14].

Lymphatic vessel density (LVD) represent the density of lymphatic vessels in the tumor micro-environment. For many cancers LVD is an established marker of adverse prognosis, but no real consensus regarding its evaluation exist [1518]. A connection between tumor VEGF-C expression and LVD in sentinel lymph nodes has been suggested [19]. In NSCLC, the expression of VEGF-C [2028], VEGF-D [21,22,29] and VEGFR3 [30] as well as lymphatic vessel density (LVD) [21,23,28,3133] have been correlated to nodal metastasis and linked to patient survival. Intriguingly, other studies fail to show these relationships.

Herein we present a structured review and meta-analysis of lymphangiogenesis and its relationship with lymph node metastasis and survival in NSCLC with emphasis on VEGF-C, VEGF-D, VEGFR3 and LVD.

Materials and Methods

Search strategy and study selection

The electronic database MEDLINE was searched for studies reported up to a date limit set to Sept. 22, 2014 with no lower date limit applied. The search string used for lymphangiogenesis was ((lymphatic vessel density) OR (LVD) OR (lymphatic micro vessel density) OR (LMVD) OR (D2-40) OR (VEGFR3) OR (flt4) OR (VEGF-C) OR (CD34)) AND ((non-small cell lung cancer) OR (non small lung cell cancer) OR (non-small lung cell cancer) OR (NSCLC) OR (adenocarcinoma of the lung) OR (lung adenocarcinoma) OR (squamous cell carcinoma of the lung) OR (lung squamous cell carcinoma) OR (lung SCC) OR (large-cell carcinoma of the lung) OR (lung large-cell carcinoma) OR (lung LCC)) AND (Humans[Mesh] AND English[lang]).

The searches were restricted to human species and English language. Publications selected for full reading were explored to complete the searches.

Inclusion criteria for the meta-analyses were as follows: (1) measure of VEGF-C, VEGF-D or VEGFR3 in primary NSCLC tissue using immunohistochemistry (IHC) or enzyme linked immunosorbent assay (ELISA)/reverse transcription-polymerase chain reaction (RT-PCR) or a measure of LVD; (2) providing survival information in the form of HR with 95% CI OR numbers with high/low expression and survival curves OR correlation with lymphatic metastasis and sufficient data to calculate RR; (3) follow-up exceeding two years; (4) only the most recent OR most adequate publication was used in the case of the same author reporting on the same population. Two reviewers (T. K and T. D) independently determined the study eligibility of publications selected for full reading, with disagreement resolved by consensus.

Data extraction and quality assessment

The eligible studies were assessed by two independent reviewers (T. K and T. D). The data retrieved from each study included author(s), test method, cut-off and data on survival and/or lymphatic metastasis. In the case of missing data the author of the primary study was contacted using the email address supplied in the original publication, missing data not supplied by the authors were extracted from the summary statistics where possible.

To assess the overall quality of studies included in the meta-analysis, these were scored according to a modified quality scale for biological prognostic factors for lung cancer developed by Steels et al. for the European Lung Cancer Working Group [34]. The quality scale incorporates four dimensions of method assessment 1) Scientific design (5 questions); 2) Laboratory methodology (7 questions); 3) Generalizability (6 questions); 4) Results analysis (4 questions). All questions were scored using an ordinal scale (values: 2, 1, 0), with results given as percentages of achievable points within each category [34].

Statistical analyses

All analyses were conducted in R-studio version 0.98.1087 using R-kernel 3.1.1.

For the meta-analyses of markers in relation to survival endpoints, univariate hazard ratios (HR) with corresponding 95% confidence intervals (CI) were combined to give the aggregated effect estimates. For cases in which these values were not reported, or supplied from the author(s) upon request, the data were extrapolated from available numerical data and survival curves according to methods described by Parmar et. al [35]. For cases in which survival curves were used, these were first digitalized using Engauge digitizer [36]. For the analysis of survival curves constant censoring during the follow-up of survivors were assumed. A 3-month interval between censoring calculations was deemed adequate. For papers with overall survival as end-point censoring was assumed to start at the time of minimum follow-up. For papers based on disease-specific survival censoring was assumed to start at patient inclusion. Maximal follow-up was set to the difference between minimum follow-up and the date of last follow-up. For papers where the date of the last follow-up was not given these dates were extrapolated from the survival curves.

For the meta-analyses of markers in relation to nodal metastasis, data was extracted in the form of 2x2 tables from which the effect estimates were calculated.

Aggregation of data was conducted using the R-package “metafor” [37]. A certain level of heterogeneity was expected and because of this, a random effects model was used to estimate the HRs and corresponding 95% CIs. Overall heterogeneity was examined using Q, tau 2, I2 and H2 statistics [38]. To explore the heterogeneity introduced by each study included in the meta-analyses, a function leaving out one article at a time (leave1out in the “metaphor” package) was used. Subgroup analyses, according to stage and histology, (adenocarcinoma or squamous cell carcinoma) were conducted for markers where sufficient information was provided.

Egger's test was used to evaluate publication bias [39]. Contour enhanced funnel plot's were used to help interpret, and to further explore publication bias in the case of funnel asymmetry [40]. The trim and fill method was used to visualize and to adjust for missing studies in the case of publication bias [41].

Results

Study selection and characteristics

Fig 1 summarizes the search strategy and inclusion processes. Three-hundred-and-forty-seven studies were identified in the initial search with an additional eight studies identified from reading bibliographies. After initial screening 42 studies were selected for full review [2033,4270]. The articles selected for full review and subsequently included in the meta-analyses are summarized in Table 1. As expected several papers reported on more than one lymphangiogenic marker.

thumbnail
Fig 1. Illustration of the literature search and inclusion processes.

Abbreviations: VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor; LVD, lymphatic vascular density.

https://doi.org/10.1371/journal.pone.0132481.g001

Thirty studies reported survival data for VEGF-C (23), VEGF-D(4), VEGFR3(9) and LVD(13) [23,24,28,29,31,32,4251,5458,6166,6870]. Of these 17, four, five and eight studies included sufficient survival data and were included in the aggregated survival-analyses of VEGF-C [24,29,31,43,44,47,48,50,5456,58,61,62,65,69,70] VEGF-D [29,44,56,70], VEGFR3 [43,44,55,56,70] and LVD [23,28,49,51,54,57,63,66] respectively, reporting on a total of 2185, 567, 825 and 849 patients.

Thirty studies reported correlations between nodal metastasis and VEGF-C (16), VEGF-D(4), VEGFR3(5) or LVD (14) [2033,45,4952,5464,66,67]. Of these 16, three, three and six included sufficient data in the form of 2x2 tables to be included in the aggregated analyses of VEGF-C [2024,27,30,33,50,55,58,6164,66], VEGF-D [21,22,30], VEGFR3 [30,55,64] and LVD [21,23,33,51,54,57] respectively, reporting on a total of 1889, 507, 556 and 677 patients.

Of the 30 studies included in the analyses, 29 used IHC for marker evaluation, while one study used RT-PCR. The percentage of patients with positive markers and the cut-offs used for marker evaluation varied extensively between studies (Table 1) and markers. For the evaluation of IHC, some studies used a straight forward percentage of positive cells, while other studies used a complex score (CS) consisting of several traits, including, but not limited to, percentage of positive cells and staining intensity. For LVD the cut-offs ranged from ≥6 to ≥38 vessels per high power field.

None of the authors prompted for supplementary data responded to the request.

Quality assessment

The scores from the quality assessment (QA) are given in Table 1 for the individual studies and summarized in Table 2 for the overall scores for each marker. The overall quality scores (QS) of papers reporting survival tended to be superior to papers reporting correlations with LNM. This difference was largest for the dimension "results analysis" and probably due to the fact that only one paper reporting associations with LNM explored their results in multivariable models.

thumbnail
Table 2. Results of the quality assessment of the included studies by marker.

Maximal score for any given category is 1.

https://doi.org/10.1371/journal.pone.0132481.t002

Meta-analysis

The results of the meta-analyses are summarized in Figs 2 and 3 and Table 3, while the individual univariate HRs, calculated or as reported, are given in Table 1.

thumbnail
Table 3. Summarizing the aggregated analyses and their corresponding heterogeneity tests.

Values for unadjusted, leave1out (test conducted without the one study introducing the most heterogeneity) and adjusted (test conducted with the inclusion of studies assumed missing in case of funnel asymmetry).

https://doi.org/10.1371/journal.pone.0132481.t003

thumbnail
Fig 2. Forest plots of the survival meta-analyses for; A) VEGF-C, B) VEGF-D, C) VEGFR3, D) LVD, Funnel plots showing the relationship between the observed HR and the standard deviation in the survival meta-analyses for; E) VEGFR3, F) LVD.

Abbreviations: VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor; LVD, lymphatic vascular density; N, number: HR, hazard ratio.

https://doi.org/10.1371/journal.pone.0132481.g002

thumbnail
Fig 3.

Forest plots of the lymphatic node metastasis meta-analyses for; A) VEGF-C, B) VEGF-D, C) VEGFR3, D) LVD, Funnel plots showing the relationship between the observed RR and the standard deviation in the lymphatic node metastasis meta-analyses for E) VEGF-C, F) LVD Abbreviations: VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor; LVD, lymphatic vascular density; N, number; RR, relative risk.

https://doi.org/10.1371/journal.pone.0132481.g003

VEGF-C: The overall HR for survival in patients expressing high tumor cell VEGF-C was 1.57 (95% CI: 1.34–1.84) across 18 studies using a random effects model and including 2107 patients (Fig 2A, Table 3). When using the trim and fill method to adjust for missing studies the HR was 1.46 (95% CI: 1.23–1.73, Table 3). Subgroup analyses reporting on VEGF-C and survival in 4 studies of adenocarinomas (n = 317) and 5 studies of stage I NSCLC (n = 542) revealed HRs of 1.45 (95% CI: 1.04–2.03) and 1.56 (95% CI: 1.09–2.24), respectively (Fig 2A).

The overall RR for the association between VEGF-C and nodal metastasis was 1.66 (1.28–2.15) across 15 studies using a random effects model and including 1889 patients (Fig 3A, Table 3). When using the trim and fill method to adjust for possible missing studies the RR was 1.26 (0.91–1.73).

VEGF-D: In this meta-analysis VEGD-D was neither found to be associated with survival nor nodal metastasis in NSCLC patients.

VEGFR3: The overall HR for survival in patients expressing high tumor cell VEGFR3 was 1.48 (95% CI: 0.77–2.86) across 5 studies using a random effects model and including 747 patients (Fig 2C, Table 3). When using the trim and fill method to adjust for missing studies the HR was 1.66 (95% CI: 0.88–3.15, Table 3). Another approach, leaving out the one study introducing the most heterogeneity, revealed an overall HR for survival of 1.98 (95% CI: 1.46–2.68, Table 3).

The overall RR for association between VEGFR3 and nodal metastasis was 1.71 (95% CI: 1.34–2.18) across 3 studies and including 556 patients (Fig 3C, Table 3). When using the trim and fill method to adjust for possible missing studies the RR was 1.56 (95% CI: 1.27–1.92).

LVD: The overall HR for survival in patients expressing high levels of LVD in tumor was 1.84 (95% CI: 1.18–2.87) across 8 studies using a random effects model and including 849 patients (Fig 2D, Table 3). The trim and fill method did not suggest any missing studies and an adjusted HR was therefore not calculated.

The overall RR for the association between LVD and nodal metastasis was 2.24 (95% CI: 1.13–4.56) across 6 studies and including 667 patients (Fig 3D, Table 3). The trim and fill method did not suggest any missing studies and a adjusted RR was therefore not calculated.

Publication bias

Egger's test for survival yielded p-values as follows: VEGF-C 0.400, VEGF-D 0.146, VEGFR3 0.020, VEGF-C in adenocarcinomas 0.073, VEGF-C in stage I patients 0.892 and LVD 0.006. This suggests significant publication bias in studies reporting on VEGFR3 and LVD and in the subgroup-analyses of VEGF-C. The contour enhanced Funnel plot for VEGFR3 (Fig 2E) suggests that the missing studies may be in an area of high statistical significance, but due to the low number of studies included, the Funnel plot has to interpreted with diligence. The contour enhanced Funnel plot for LVD (Fig 2F) suggests that the missing articles are those with low standard error of the effect estimate.

Egger's test for nodal metastasis yielded p-values as follows: VEGF-C <0.001, VEGF-D0.169, VEGFR3 0.375 and LVD 0.004. This suggests significant publication bias in the studies reporting on VEGF-C and LVD. The contour enhanced Funnel plot of VEGF-C (Fig 3E) indicates a strong positive publication bias. The contour enhanced Funnel plot of LVD (Fig 3F) indicates the lack of larger studies and a possible positive publication bias.

Statistical considerations

The results of the aggregated analyses were obtained using the results reported in univariate analyses only. Some information can potentially be lost from studies reporting only multivariate results with inadequate data to reconstruct the univariate analysis. Multivariate data are only valid in their own multivariate system. Including multivariate data in the aggregated analysis provides increased numbers of patients at the cost of increased heterogeneity. This may be tolerated if all included studies are adequately large and includes the same, or approximately the same, variables in their multivariate models. In this meta-analysis the number of patients in the included studies ranged from 48 to 335 and the variables included in the different multivariate models varied (data not shown). Thus making the inclusion of data from multivariate analyses in the meta-analysis a dubious venture.

Heterogeneity is difficult to avoid when conducting meta-analyses based on marker data, as there exists no consensus on how such studies should be conducted. Among the included studies (Table 1), all but one were retrospective and all but one used IHC to detect the protein markers. Moreover, they differ largely with respect to selected antibodies, choice of cut-offs for the different markers and in the percentage of positive cases. The European Lung Cancer Working Party has proposed a quality scale for biological prognostic factors for lung cancer, but few studies adhere to this scale [34]. We utilized this scale to rate the included studies (Table 2). Mixed model approaches (data not shown) utilizing histology, stage, method, percentage of positives cases, cut-offs or quality scores as static modifiers were attempted, but none of these contributed to the overall model design and thus were rejected.

The present meta-analysis was restricted to articles published in the English language. This could introduce a positive selection bias as there is a tendency for positive studies to be published in English while negative studies more often are published in the authors native language [71]. Indeed, testing in the current meta-analyses suggested significant publication bias for studies reporting survival based on VEGFR3 and LVD and for LNM based on VEGF-C and LVD. Based on this the results of this meta-analysis have to be interpreted carefully and should be confirmed in larger trials.

Obtaining individual patient data for meta-analysis would theoretically help to define the role of lymphangiogenic markers by adjusting for the same confounders before data aggregation across all included studies [72]. However, this does not appear to be feasible as all authors were prompted by e-mail for additional information, but none of them replied.

Discussion

Summary of the meta-analyses

This is, to our knowledge, the most comprehensive structured review with meta-analysis of VEGF-C, VEGF-D and LVD in NSCLC. One meta-analysis has previously reported on VEGF-C and VEGFR3 and their association with survival in NSCLC [73]. Zhan et al. published the aggregated results of 8 studies reporting on VEGF-C and 4 studies allegedly reporting on VEGFR3 (The authors considered Flt1 to represent VEGFR3 –Flt1 is in fact equal to VEGFR1, while Flt4 is equal to VEGFR3) in 2009. They found the aggregated results of the studies reporting on VEGF-C and VEGFR3 to be non-significant in NSCLC [73]. Further, Wang et al. reviewed the aggregated results of 10 studies reporting on LVD and survival in NSCLC patients in 2012. They included multivariable adjusted results and in addition one study reporting VEGFR3 in tumor cell cytoplasm [16]. Since 2009 several studies on VEGF-C, VEGF-D, VEGFR3 and LVD have been published. This current meta-analysis correlates VEGF-C, VEGFR3 and high levels of LVD with nodal metastasis in NSCLC patients and identifies high VEGF-C (HR 1.57 95% CI: 1.34–1.84) and high levels of LVD (HR 1.84 95% CI: 1.18–2.87) as significant prognostic markers of poor survival.

The VEGF-C/-D/R3 axis in relation to survival and lymphatic metastasis in NSCLC

VEGF-C, VEGF-D and their corresponding receptor VEGFR3 are well-known and strong lymphangiogenic markers [9,10,74]. In an extensive mapping of the VEGF-C/VEGFR3 axis in lung adenocarcinoma tumors, cell-lines and animal models Su. et al. found high expression of both VEGF-C and VEGFR3 to be correlated with nodal metastasis whereas expression was low for early-stage disease [11]. Further, NSCLC cell-lines over-expressing the VEGF-C/VEGFR3 axis showed increased migration, and, when introduced into xenograft models, more frequently formed lung metastases, compared to NSCLC cell-lines where VEGF-C/VEGFR3 signaling had been abrogated [11]. He et al. investigated the interaction between the lung cancer cell-line LNM35, which expresses high levels of VEGF-C, and lymphatic endothelial cells (LECs). They found VEGF-C to induce lymphatic vessel destabilization and enlargement of collecting lymphatic vessels, which further lead to passage of tumor clusters to sentinel lymph nodes [14]. In squamous cell carcinoma of the skin, Hirakawa et al. showed that VEGF-C over-expressing tumors maintained their lymphangiogenic profile after arrival in the sentinel lymph nodes [13]. Building on this, Liersch et al. found melanoma xenografts over-expressing VEGF-C to instigate lymphatic vascularization in the sentinel lymph nodes before the presence of metastatic cells could be detected [12]. These latter results are supported by a study on human surgical specimen (oral squamous cell carcinoma) demonstrating an increased number of high endothelial venules and lymphatic vessels without detecting metastatic cells [19]. However, in a recent study Nwogu et al. investigated the presence of nodal micro-metastases in pStage I and II NSCLC patients using conventional H&E, IHC and RT-PCR on resected lymph nodes. They found that 35/40, 33/40 and 16/40, respectively, were N0, indicating that conventional H&E and IHC might not be adequately sensitive to detect metastatic cells. They also demonstrated a strong correlation between the presence of nodal micro-metastases and the expression of VEGF-A, VEGF-C, VEGF-D and VEGFR3 in the lymph nodes [75]. Niki et al. isolated total RNA from 60 surgically resected lung adenocarcinomas of which 27 had lymph node metastasis and found only weak correlations between lymph node metastasis and VEGF-C. However, a high ratio VEGF-A,-B or -C to VEGF-D was associated with lymph node metastasis and the authors proposed that VEGF-D may have a regulatory role in tumor lymphangiogenesis [60]. Clearly, RNA expression does not necessarily translate to protein expression and it can be argued that a high ratio of VEGF-A, B or -C RNA to VEGF-D RNA simply represents over-expression of these molecules rather than VEGF-D having a regulatory role. The results of this meta-analysis indicate that VEGF-C is a prognostic factor for lymph node metastasis (HR 1.66 95% CI: 1.28–2.15) and survival (HR 1.57 95% CI: 1.34–1.84) in NSCLC patients, and further, that high expression of VEGFR3 is a prognostic factor for lymph node metastasis (HR 1.71 95% CI: 1.34–2.18)

LVD in relation to survival and lymphatic metastasis in NSCLC

LVD describes the density of lymphatic vessels in the tumor micro environment [76]. In lymphovascular invasion (LVI) is an established adverse prognostic factor in NSCLC describing the presence of cancer cells within the lymphovascular space. LVI was recently reviewed by Mollberg et al. for stage I NSCLC [15]. There is controversy regarding the clinical implications of LVD in NSCLC. As of today, no consensus for the evaluation of LVD exists. This results in LVD being evaluated using several antibodies including D2-40, podoplanin, and VEGFR3 and by several techniques including the single most positive high-powered field (hot-spot), the invasive tumor front, the tumor center and combinations of these. Our systematic approach identified 12 studies reporting survival and 14 studies reporting associations with LNM in NSCLC patients. The studies varied in methodological approach with some evaluating the presence of LVD in hot-spots and some in random areas, either within the central tumor, in the invasive tumor front, a mix of different areas or location not given. This lead to a considerable difference in cut-offs used for high/low vessel count (summarized in Table 1). Unfortunately the number of studies utilizing the same methodology was too small to warrant sub-group analyses. Nevertheless, our meta-analysis identified high LVD-levels as a marker of poor prognosis (HR 1.84 95% CI 1.18–2.87) and LNM (HR 2.24 95% CIT 1.13–4.46) in NSCLC patients. Interestingly, Sun et al. found peritumoral LVD to be linked to survival while intratumoral LVD was not, suggesting this to be of interest for future studies [66]. This has an intuitive ring to it, as it seems logical that tumor cell migration to lymph nodes only can occur in functional lymphatic vessels mostly found in the tumor periphery and not in intratumoral lymphatic vessels that tend to be collapsed and non-functional. However, there is evidence suggesting that tumor cells may utilize the latter approach when entering the lymphatic system [76]. Renyi-Vamos et al. stratified patients based on low and high angiogenic activity in tumors, and found peritumoral LVD to significantly worsen the prognosis of those in the high angiogenic group [63]. In oral squamous carcinoma, LVD in sentinel nodes was elevated regardless of the presence of metastatic cancer cells, and correlated to tumor VEGF-C expression, suggesting an interplay between tumor and lymph node prior to the arrival of cancer cells [19]. In a recent meta-analysis, including 1044 breast cancer patients, LVD was correlated to LNM, but not to other known breast cancer traits [77]. Results of this meta-analysis coincide with the published literature of other cancer entities and suggests LVD as a prognostic marker for survival and LNM in NSCLC patients.

Targeting lymphatic metastasis in the treatment of NSCLC

With the availability of small-molecule inhibitors and antibodies that could potentially target tumor lymphangiogenesis it seems appropriate that this treatment should be offered patients to whom it may prove beneficial. The challenge will be to select the correct patients. As pointed out in the xenograft model from He et al. timing seems to be of importance. In their model, no benefit of anti-VEGFR3 antibodies was observed when administered after the development of lymphangiogenic networks [14]. Evidently, results from murine models should be carefully evaluated before applied to actual human patients, but bad timing appears to be a feasible explanation for part of the moderate or lacking effects reported after a number of anti-angiogenic and also anti-lymphangiogenic approaches [78]. Interestingly, Tamura et al. have investigated serum VEGF-C levels of NSCLC patients with and without LNM [67]. They found high levels of VEGF-C to be associated with LNM with positive and negative predictive values of 70 and 77.3 respectively and that the addition of serum VEGF-A, serum matrix metalloproteinase 9 or CT-images further improved the diagnostic properties of the test [25,67,79]. These results were supported by the findings of Daly et al. in 2014 [80].

The results of this meta-analysis indicate that targeting lymphangiogenesis could prove beneficial for selected sub-groups of NSCLC patients. With timing being essential, theoretically, the best use of lymphangiogenic inhibitors could be in the adjuvant setting and for patients with low grade tumors in whom lymphatic node metastases have yet to develop. Obviously, the use of any drug in the adjuvant setting must be considered carefully as a proportion of patients have been cured of their cancer and will not benefit from further treatment. The studies linking serum levels of VEGF-C to nodal metastasis are especially interesting. These results indicate that NSCLC patients with no acknowledgeable LNM and high serum VEGF-C might be at risk and may be the patients who will benefit from anti-lymphaniogenic. According to the recent study by Daly et al. their group is setting up animal models with a plan to further elucidate their findings on serum-VEGF-C levels and LNM, hopefully culminating in a clinical trial; the results of these subsequent studies, will be very interesting [80].

During the past decade, cancer immunology has experienced a renaissance through the understanding of the immune check-points and the development of immune check-point inhibitors [8183]. This important clinical research effort which initially succeeded in melanoma is now benefiting a wide host of other tumor groups [8183]. Evidence indicates lymphangiogenic markers to mediate a profound impact on the immune system. Tumors expressing lymprhangiogenic factors are believed to alter the micro-environment in the lymph nodes, hence preparing them for the arrival and harboring of metastatic cells [12,13,19]. In fact, active lymphangiogenic processes may down-regulate immune responses and thus be partly responsible for the poor effect the immune system shows on most malignant tumors [84]. However, once the tumor associated lymphatic network is established, little is known of its role regarding the tumor immune responses. Tumor antigens must reach antigen presenting cells in order to induce a T-cell mediated immune response, and in this process the presence of, or interaction with, tumor-induced lymphatics may be of importance [84].

Conclusions and implications for research

Our results indicate a connection between lymphangiogenic markers, LVD, LNM and survival in NSCLC. Nevertheless, it is apparent that lymphangiogenic factors cannot explain the full extent of LNM. Alternative pathways exist for cancer cells to recruit and invade lymphovascular structures, such as vessel co-option and vascular mimicry [85].

Further studies are warranted to evaluate the results of this meta-analysis and it would be highly interesting to see studies combining expressions of lymphangiogenic markers in primary tumor tissues and metastatic lymph nodes. In addition, we believe that studies investigating this intricate relationship between tumor-lymphangiogenesis and immunology will prove beneficial, not only for our understanding of these principles, but also for patients who could benefit from combined therapeutic approaches.

Author Contributions

Analyzed the data: TKK TW EEP SMH. Wrote the paper: TD EEP SMH RMB LTB.

References

  1. 1. Siegel R, Ma J, Zou Z, Jemal A. Cancer statistics, 2014. CA Cancer J Clin. 2014;64: 9–29. pmid:24399786
  2. 2. Detterbeck FC, Postmus PE, Tanoue LT. The stage classification of lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143: e191S–210S. pmid:23649438
  3. 3. Sobin LH, Compton CC. TNM seventh edition: what’s new, what's changed: communication from the International Union Against Cancer and the American Joint Committee on Cancer. Cancer. 2010;116: 5336–9. pmid:20665503
  4. 4. Strand T-E, Rostad H, Wentzel-Larsen T, von Plessen C. A population-based evaluation of the seventh edition of the TNM system for lung cancer. Eur Respir J. 2010;36: 401–7. pmid:20075046
  5. 5. Wang J, Wu N, Zheng Q, Feng Y, Yan S, Lv C, et al. Evaluation of the 7th edition of the TNM classification for lung cancer at a single institution. J Cancer Res Clin Oncol. 2014;140: 1189–95. pmid:24676426
  6. 6. Yoshida Y, Murayama T, Sato Y, Suzuki Y, Saito H, Tanaka N. Validation of 7th TNM staging system for lung cancer, based on surgical outcomes. Asian Cardiovasc Thorac Ann. 2013;21: 693–9. pmid:24569328
  7. 7. Youlden DR, Cramb SM, Baade PD. The International Epidemiology of Lung Cancer: geographical distribution and secular trends. J Thorac Oncol. 2008;3: 819–31. pmid:18670299
  8. 8. Guerrera F, Errico L, Evangelista A, Filosso PL, Ruffini E, Lisi E, et al. Exploring Stage I non-small-cell lung cancer: development of a prognostic model predicting 5-year survival after surgical resection†. Eur J Cardiothorac Surg. 2014; 1–7.
  9. 9. Alitalo a, Detmar M. Interaction of tumor cells and lymphatic vessels in cancer progression. Oncogene. Nature Publishing Group; 2012;31: 4499–508.
  10. 10. Adams RH, Alitalo K. Molecular regulation of angiogenesis and lymphangiogenesis. Nat Rev Mol Cell Biol. 2007;8: 464–78. pmid:17522591
  11. 11. Su J-L, Yang P-C, Shih J-Y, Yang C, Wei L, Hsieh C-Y, et al. The VEGF-C/Flt-4 axis promotes invasion and metastasis of cancer cells. Cancer Cell. 2006;9: 209–23. pmid:16530705
  12. 12. Liersch R, Hirakawa S, Berdel WE, Mesters RM, Detmar M. Induced lymphatic sinus hyperplasia in sentinel lymph nodes by VEGF-C as the earliest premetastatic indicator. Int J Oncol. 2012;41: 2073–8. pmid:23076721
  13. 13. Hirakawa S, Brown LF, Kodama S, Paavonen K, Alitalo K, Detmar M. VEGF-C-induced lymphangiogenesis in sentinel lymph nodes promotes tumor metastasis to distant sites. Blood. 2007;109: 1010–7. pmid:17032920
  14. 14. He Y, Rajantie I, Pajusola K, Jeltsch M, Holopainen T, Yla-Herttuala S, et al. Vascular endothelial cell growth factor receptor 3-mediated activation of lymphatic endothelium is crucial for tumor cell entry and spread via lymphatic vessels. Cancer Res. 2005;65: 4739–46. pmid:15930292
  15. 15. Mollberg NM, Bennette C, Howell E, Backhus L, Devine B, Ferguson MK. Lymphovascular invasion as a prognostic indicator in stage I non-small cell lung cancer: a systematic review and meta-analysis. Ann Thorac Surg. Elsevier Inc; 2014;97: 965–71.
  16. 16. Wang J, Li K, Wang B, Bi J. Lymphatic microvessel density as a prognostic factor in non-small cell lung carcinoma: a meta-analysis of the literature. Mol Biol Rep. 2012;39: 5331–8. pmid:22167333
  17. 17. Bolenz C, Fernández MI, Tilki D, Herrmann E, Heinzelbecker J, Ergün S, et al. The role of lymphangiogenesis in lymphatic tumour spread of urological cancers. BJU Int. 2009;104: 592–7. pmid:19594740
  18. 18. Yu M, Liu L, Liang C, Li P, Ma X, Zhang Q, et al. Intratumoral vessel density as prognostic factors in head and neck squamous cell carcinoma: a meta-analysis of literature. Head Neck. 2014;36: 596–602. pmid:23780885
  19. 19. Chung MK, Do I-G, Jung E, Son Y-I, Jeong H-S, Baek C-H. Lymphatic vessels and high endothelial venules are increased in the sentinel lymph nodes of patients with oral squamous cell carcinoma before the arrival of tumor cells. Ann Surg Oncol. 2012;19: 1595–601. pmid:22124758
  20. 20. Li M, Chen H, Diao L, Zhang Y, Xia C, Yang F. Caveolin-1 and VEGF-C promote lymph node metastasis in the absence of intratumoral lymphangiogenesis in non-small cell lung cancer. Tumori. 2010;96: 734–43. Available: http://www.ncbi.nlm.nih.gov/pubmed/21302621. pmid:21302621
  21. 21. Adachi Y, Nakamura H, Kitamura Y, Taniguchi Y, Araki K, Shomori K, et al. Lymphatic vessel density in pulmonary adenocarcinoma immunohistochemically evaluated with anti-podoplanin or anti-D2-40 antibody is correlated with lymphatic invasion or lymph node metastases. Pathol Int. 2007;57: 171–7. pmid:17316411
  22. 22. Feng Y, Wang W, Hu J, Ma J, Zhang Y, Zhang J. Expression of VEGF-C and VEGF-D as significant markers for assessment of lymphangiogenesis and lymph node metastasis in non-small cell lung cancer. Anat Rec (Hoboken). 2010;293: 802–12.
  23. 23. Guo X, Chen Y, Xu Z, Qian Y, Yu X. Prognostic significance of VEGF-C expression in correlation with COX-2, lymphatic microvessel density, and clinicopathologic characteristics in human non-small cell lung cancer. Acta Biochim Biophys Sin (Shanghai). 2009;41: 217–222.
  24. 24. Kajita T, Ohta Y, Kimura K, Tamura M, Tanaka Y, Tsunezuka Y, et al. The expression of vascular endothelial growth factor C and its receptors in non-small cell lung cancer. Br J Cancer. 2001;85: 255–60. pmid:11461086
  25. 25. Tamura M, Oda M, Matsumoto I, Tsunezuka Y, Kawakami K, Ohta Y, et al. The combination assay with circulating vascular endothelial growth factor (VEGF)-C, matrix metalloproteinase-9, and VEGF for diagnosing lymph node metastasis in patients with non-small cell lung cancer. Ann Surg Oncol. 2004;11: 928–33. pmid:15383417
  26. 26. Takizawa H, Kondo K, Fujino H, Kenzaki K, Miyoshi T, Sakiyama S, et al. The balance of VEGF-C and VEGFR-3 mRNA is a predictor of lymph node metastasis in non-small cell lung cancer. Br J Cancer. 2006;95: 75–9. pmid:16755294
  27. 27. Zuo S, Ji Y, Wang J, Guo J. Expression and clinical implication of HIF-1alpha and VEGF-C in non-small cell lung cancer. J Huazhong Univ Sci Technolog Med Sci. 2008;28: 674–6. pmid:19107364
  28. 28. Zhang BC, Gao J, Wang J, Rao ZG, Wang BC, Gao JF. Tumor-associated macrophages infiltration is associated with peritumoral lymphangiogenesis and poor prognosis in lung adenocarcinoma. Med Oncol. 2011;28: 1447–52. pmid:20676804
  29. 29. Ko YH, Jung C-K, Lee M-A, Byun JH, Kang JH, Lee KY, et al. Clinical significance of vascular endothelial growth factors (VEGF)-C and -D in resected non-small cell lung cancer. Cancer Res Treat. 2008;40: 133–40. pmid:19688120
  30. 30. Donnem T, Al-Shibli K, Al-Saad S, Delghandi MP, Busund L-T, Bremnes RM. VEGF-A and VEGFR-3 correlate with nodal status in operable non-small cell lung cancer: inverse correlation between expression in tumor and stromal cells. Lung Cancer. 2009;63: 277–83. pmid:18599153
  31. 31. Li Q, Dong X, Gu W, Qiu X, Wang E. Clinical significance of co-expression of VEGF-C and VEGFR-3 in non-small cell lung cancer. Chin Med J (Engl). 2003;116: 727–30. Available: http://www.ncbi.nlm.nih.gov/pubmed/12875690.
  32. 32. Saijo T, Ishii G, Ochiai A, Hasebe T, Yoshida J, Nishimura M, et al. Evaluation of extratumoral lymphatic permeation in non-small cell lung cancer as a means of predicting outcome. Lung Cancer. 2007;55: 61–6. pmid:17134789
  33. 33. Takanami I. Lymphatic microvessel density using D2-40 is associated with nodal metastasis in non-small cell lung cancer. Oncol Rep. 2006;15: 437–42. Available: http://www.ncbi.nlm.nih.gov/pubmed/16391866. pmid:16391866
  34. 34. Steels E, Paesmans M, Berghmans T, Branle F, Lemaitre F, Mascaux C, et al. Role of p53 as a prognostic factor for survival in lung cancer: a systematic review of the literature with a meta-analysis. Eur Respir J Off J Eur Soc Clin Respir Physiol. 2001;18: 705–719. Available: http://www.ncbi.nlm.nih.gov/pubmed/11716177.
  35. 35. Parmar MK, Torri V, Stewart L. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med. 1998;17: 2815–34. Available: http://www.ncbi.nlm.nih.gov/pubmed/9921604. pmid:9921604
  36. 36. Enguage Digitizer [Internet]. Available: http://digitizer.sourceforge.net/.
  37. 37. Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw. 2010;36. Available: http://raptor1.bizlab.mtsu.edu/s-drive/TEFF/Rlib/library/metafor/doc/viechtbauer2010.pdf.
  38. 38. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21: 1539–1558. pmid:12111919
  39. 39. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315: 629–34. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2665616&tool=pmcentrez&rendertype=abstract. pmid:9310563
  40. 40. Palmer T, Peters J, Sutton A, Moreno S. Contour-enhanced funnel plots for meta-analysis. Stata J. 2008; Available: http://repec.org/usug2009/palmer_presentation.pdf.
  41. 41. Duval S, Tweedie R. A nonparametric “trim and fill” method of accounting for publication bias in meta-analysis. J Am Stat. 2000;95: 89–98. Available: http://amstat.tandfonline.com/doi/abs/10.1080/01621459.2000.10473905.
  42. 42. Anagnostou VK, Tiniakos DG, Fotinou M, Achimastos A, Syrigos KN. Multiplexed analysis of angiogenesis and lymphangiogenesis factors predicts outcome for non-small cell lung cancer patients. Virchows Arch. 2011;458: 331–40. pmid:21161269
  43. 43. Arinaga M, Noguchi T, Takeno S, Chujo M, Miura T, Uchida Y. Clinical significance of vascular endothelial growth factor C and vascular endothelial growth factor receptor 3 in patients with nonsmall cell lung carcinoma. Cancer. 2003;97: 457–64. pmid:12518370
  44. 44. Carrillo de Santa Pau E, Arias FC, Caso Peláez E, Muñoz Molina GM, Sánchez Hernández I, Muguruza Trueba I, et al. Prognostic significance of the expression of vascular endothelial growth factors A, B, C, and D and their receptors R1, R2, and R3 in patients with nonsmall cell lung cancer. Cancer. 2009;115: 1701–12. pmid:19197998
  45. 45. Chen F, Takenaka K, Ogawa E, Yanagihara K, Otake Y, Wada H, et al. Flt-4-positive endothelial cell density and its clinical significance in non-small cell lung cancer. Clin Cancer Res. 2004;10: 8548–53. pmid:15623638
  46. 46. Chen G, Liu X-Y, Wang Z, Liu F-Y. Vascular endothelial growth factor C: the predicator of early recurrence in patients with N2 non-small-cell lung cancer. Eur J Cardiothorac Surg. 2010;37: 546–51. pmid:19758816
  47. 47. Chen SC, Shih CM, Tseng GC, Cheng WE, Chiou J, Hsiao M, et al. Vascular endothelial growth factor C as a predictor of early recurrence and poor prognosis of resected stage I non-small cell lung cancer. Ann Acad Med Singapore. 2011;40: 319–24. Available: http://www.ncbi.nlm.nih.gov/pubmed/21870023. pmid:21870023
  48. 48. Enatsu S, Iwasaki A, Shirakusa T, Hamasaki M, Nabeshima K, Iwasaki H, et al. Expression of hypoxia-inducible factor-1 alpha and its prognostic significance in small-sized adenocarcinomas of the lung. Eur J Cardiothorac Surg. 2006;29: 891–5. pmid:16675263
  49. 49. Faoro L, Hutto JY, Salgia R, El-Zayaty SA, Ferguson MK, Cheney RT, et al. Lymphatic vessel density is not associated with lymph node metastasis in non-small cell lung carcinoma. Arch Pathol Lab Med. 2008;132: 1882–8. pmid:19061284
  50. 50. Huang C, Liu D, Masuya D, Nakashima T, Kameyama K, Ishikawa S, et al. Clinical application of biological markers for treatments of resectable non-small-cell lung cancers. Br J Cancer. 2005;92: 1231–9. pmid:15785747
  51. 51. Iwakiri S, Nagai S, Katakura H, Takenaka K, Date H, Wada H, et al. D2-40-positive lymphatic vessel density is a poor prognostic factor in squamous cell carcinoma of the lung. Ann Surg Oncol. 2009;16: 1678–85. pmid:19330380
  52. 52. Jin S, Zhu W, Shi Q, Zhang Z, Guo R. Clinicopathological significance of lymphatic vessel density marked by D2-40 and E-cadherin expression in non-small-cell lung cancer. Med Oncol. 2012;29: 3157–61. pmid:22528515
  53. 53. Kadota K, Huang C-L, Liu D, Nakashima N, Yokomise H, Ueno M, et al. The clinical significance of the tumor cell D2-40 immunoreactivity in non-small cell lung cancer. Lung Cancer. Elsevier Ireland Ltd; 2010;70: 88–93.
  54. 54. Kadota K, Huang C-L, Liu D, Ueno M, Kushida Y, Haba R, et al. The clinical significance of lymphangiogenesis and angiogenesis in non-small cell lung cancer patients. Eur J Cancer. 2008;44: 1057–67. pmid:18396396
  55. 55. Kojima H, Shijubo N, Yamada G, Ichimiya S, Abe S, Satoh M, et al. Clinical significance of vascular endothelial growth factor-C and vascular endothelial growth factor receptor 3 in patients with T1 lung adenocarcinoma. Cancer. 2005;104: 1668–77. pmid:16116610
  56. 56. Maekawa S, Iwasaki A, Shirakusa T, Enatsu S, Kawakami T, Kuroki M, et al. Correlation between lymph node metastasis and the expression of VEGF-C, VEGF-D and VEGFR-3 in T1 lung adenocarcinoma. Anticancer Res. 2007;27: 3735–41. Available: http://www.ncbi.nlm.nih.gov/pubmed/17970036. pmid:17970036
  57. 57. Min KH, Park SJ, Lee KS, Hwang SH, Kim SR, Moon H, et al. Clinical usefulness of D2-40 in non-small cell lung cancer. Lung. 2011;189: 57–63. pmid:21107593
  58. 58. Nakashima T, Huang C, Liu D, Kameyama K, Masuya D, Ueno M, et al. Expression of vascular endothelial growth factor-A and vascular endothelial growth factor-C as prognostic factors for non-small cell lung cancer. Med Sci Monit. 2004;10: BR157–65. Available: http://www.ncbi.nlm.nih.gov/pubmed/15173661. pmid:15173661
  59. 59. Niki T, Iba S, Yamada T, Matsuno Y, Enholm B, Hirohashi S. Expression of vascular endothelial growth factor receptor 3 in blood and lymphatic vessels of lung adenocarcinoma. J Pathol. 2001;193: 450–7. pmid:11276003
  60. 60. Niki T, Iba S, Tokunou M, Yamada T, Matsuno Y, Hirohashi S. Expression of vascular endothelial growth factors A, B, C, and D and their relationships to lymph node status in lung adenocarcinoma. Clin Cancer Res. 2000;6: 2431–9. Available: http://clincancerres.aacrjournals.org/content/6/6/2431.short. pmid:10873096
  61. 61. Ogawa E, Takenaka K, Yanagihara K, Kurozumi M, Manabe T, Wada H, et al. Clinical significance of VEGF-C status in tumour cells and stromal macrophages in non-small cell lung cancer patients. Br J Cancer. 2004;91: 498–503. pmid:15226767
  62. 62. Ohta Y, Nozawa H, Tanaka Y, Oda M, Watanabe Y. Increased vascular endothelial growth factor and vascular endothelial growth factor-c and decreased nm23 expression associated with microdissemination in the lymph nodes in stage I non-small cell lung cancer. J Thorac Cardiovasc Surg. 2000;119: 804–13. Available: http://www.ncbi.nlm.nih.gov/pubmed/10733773. pmid:10733773
  63. 63. Renyi-Vamos F, Tovari J, Fillinger J, Timar J, Paku S, Kenessey I, et al. Lymphangiogenesis correlates with lymph node metastasis, prognosis, and angiogenic phenotype in human non-small cell lung cancer. Clin Cancer Res. 2005;11: 7344–53. pmid:16243806
  64. 64. Saintigny P, Kambouchner M, Ly M, Gomes N, Sainte-Catherine O, Vassy R, et al. Vascular endothelial growth factor-C and its receptor VEGFR-3 in non-small-cell lung cancer: concurrent expression in cancer cells from primary tumour and metastatic lymph node. Lung Cancer. 2007;58: 205–13. pmid:17686546
  65. 65. Su JL, Shih JY, Yen ML, Jeng YM, Chang CC, Hsieh CY, et al. Cyclooxygenase-2 induces EP1- and HER-2/Neu-dependent vascular endothelial growth factor-C up-regulation: a novel mechanism of lymphangiogenesis in lung adenocarcinoma. Cancer Res. 2004;64: 554–64. pmid:14744769
  66. 66. Sun J, Wang Y, Chen Z, Zhuo W, Zhu B, Liao R, et al. Detection of lymphangiogenesis in non-small cell lung cancer and its prognostic value. J Exp Clin Cancer Res. 2009;28: 21. pmid:19216806
  67. 67. Tamura M, Ohta Y. Serum vascular endothelial growth factor-C level in patients with primary nonsmall cell lung carcinoma: a possible diagnostic tool for lymph node metastasis. Cancer. 2003;98: 1217–22. pmid:12973845
  68. 68. Wang J, Wang B, Bi J, Zhang C. Cytoplasmic HuR expression correlates with angiogenesis, lymphangiogenesis, and poor outcome in lung cancer. Med Oncol. 2011;28 Suppl 1: S577–85. pmid:21046284
  69. 69. Yamashita T, Uramoto H, Onitsuka T, Ono K, Baba T, So T, et al. Association between lymphangiogenesis-/micrometastasis- and adhesion-related molecules in resected stage I NSCLC. Lung Cancer. Elsevier Ireland Ltd; 2010;70: 320–8.
  70. 70. Donnem T, Al-Saad S, Al-Shibli K, Delghandi MP, Persson M, Nilsen MN, et al. Inverse prognostic impact of angiogenic marker expression in tumor cells versus stromal cells in non small cell lung cancer. Clin Cancer Res. 2007;13: 6649–57. pmid:18006765
  71. 71. Egger M, Zellweger-Zähner T, Schneider M, Junker C, Lengeler C, Antes G. Language bias in randomised controlled trials published in English and German. Lancet. 1997;350: 326–9. pmid:9251637
  72. 72. Stewart LA, Parmar MK. Meta-analysis of the literature or of individual patient data: is there a difference? Lancet. 1993;341: 418–22. Available: http://www.sciencedirect.com/science/article/pii/014067369393004K. pmid:8094183
  73. 73. Zhan P, Wang J, Lv X, Wang Q, Qiu L, Lin X, et al. Prognostic value of vascular endothelial growth factor expression in patients with lung cancer: a systematic review with meta-analysis. J Thorac Oncol. 2009;4: 1094–103. pmid:19687765
  74. 74. Achen MG, Stacker SA. Molecular control of lymphatic metastasis. Ann N Y Acad Sci. Blackwell Publishing Ltd; 2008;1131: 225–34.
  75. 75. Nwogu CE, Yendamuri S, Tan W, Kannisto E, Bogner P, Morrison C, et al. Lung cancer lymph node micrometastasis detection using real-time polymerase chain reaction: correlation with vascular endothelial growth factor expression. J Thorac Cardiovasc Surg. 2013;145: 702–7; discussion 707–8. pmid:23414988
  76. 76. Ji R-C. Lymphatic endothelial cells, tumor lymphangiogenesis and metastasis: New insights into intratumoral and peritumoral lymphatics. Cancer Metastasis Rev. 2006;25: 677–94. pmid:17160713
  77. 77. Chen Y, Yan J, Yuan Z, Yu S, Yang C, Wang Z, et al. A meta-analysis of the relationship between lymphatic microvessel density and clinicopathological parameters in breast cancer. Bull Cancer. 2013;100: 1–10. pmid:23501839
  78. 78. Jain RK. Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science. 2005;307: 58–62. pmid:15637262
  79. 79. Tamura M, Oda M, Tsunezuka Y, Matsumoto I, Kawakami K, Ohta Y, et al. Chest CT and serum vascular endothelial growth factor-C level to diagnose lymph node metastasis in patients with primary non-small cell lung cancer. Chest. 2004;126: 342–6. pmid:15302715
  80. 80. Daly S, Kubasiak JC, Rinewalt D, Pithadia R, Basu S, Ms CF, et al. Circulating Angiogenesis Biomarkers Are Associated With Disease Progression in Lung Adenocarcinoma. Ann Thorac Surg. Elsevier Inc; 2014;98: 1968–1975.
  81. 81. Antonia SJ, Larkin J, Ascierto P a. Immuno-oncology Combinations: A Review of Clinical Experience and Future Prospects. Clin Cancer Res. 2014;20: 6258–6269. pmid:25341541
  82. 82. Galon J, Angell HK, Bedognetti D, Marincola FM. The continuum of cancer immunosurveillance: prognostic, predictive, and mechanistic signatures. Immunity. 2013;39: 11–26. pmid:23890060
  83. 83. Galon J, Mlecnik B, Bindea G, Angell HK, Berger A, Lagorce C, et al. Towards the introduction of the “Immunoscore” in the classification of malignant tumours. J Pathol. 2014;232: 199–209. pmid:24122236
  84. 84. Card CM, Yu SS, Swartz MA. Emerging roles of lymphatic endothelium in regulating adaptive immunity. J Clin Invest. 2014;124: 943–52. pmid:24590280
  85. 85. Donnem T, Hu J, Ferguson M, Adighibe O, Snell C, Harris AL, et al. Vessel co-option in primary human tumors and metastases: an obstacle to effective anti-angiogenic treatment? Cancer Med. 2013;2: 427–36. pmid:24156015