Lymphatic vessel density as a prognostic indicator in Asian NSCLC patients: a meta-analysis

Background To determine the association of lymphatic vessel density (LVD) with the prognosis of Asian non-small cell lung cancer (NSCLC) patients via a meta-analysis. Methods Eligible studies were selected by searching PubMed and EMBASE from inception to July 25, 2017. The reference lists of the retrieved articles were also consulted. The information was independently screened by two authors. When heterogeneity was significant, a random-effects model was used to determine overall pooled risk estimates. Results A total of 15 studies with 1075 patients were finally included in the meta-analysis. LVD was positively associated with the prognosis of NSCLC in the overall analysis (hazard ratio (HR) 1.14, 95% confidence interval (95% CI): 1.02–1.27, p = 0.000, I2 = 73.2%). Subgroup analyses were performed on 5 VEGFR-3 groups (p = 0.709, I2 = 0.0%), 3 LYVE-1 groups (p = 0.01, I2 = 86.4%), 5 D2–40 groups (p = 0.019, I2 = 66.2%), and 2 podoplanin groups (p = 0.094, I2 = 64.5%). Sensitivity analysis indicated robust results. There was no publication bias. Conclusions LVD is an indicator of poor prognosis in Asian NSCLC patients.


Background
Lung cancer is a malignant disease associated with the highest mortality rate (18.2%) among all types of cancer worldwide [1,2]. Non-small cell lung cancer (NSCLC) represents the majority (~85%) of all lung cancer cases, with lung adenocarcinoma (ADC) and squamous cell carcinoma (SCC) being the most frequently diagnosed histological types [3]. Approximately half of all NSCLC patients have metastasis, and this type of cancer is usually diagnosed at advanced stages. Despite great progress in treatment modalities (such as surgical resection, chemotherapy, radiotherapy, targeted therapy, biotherapy, and cellular immunotherapy), the prognosis of NSCLC remains poor, and the long-term survival of NSCLC patients is still dismal [4]. Thus, it is important to find novel prognostic therapeutic targets and precise prognostic markers for this type of cancer.
Cancer relapse and metastasis lead to poor prognosis. The most common mode of metastasis is lymph node metastasis. During the early stages of tumor dissemination, malignant cells spread from primary sites to regional lymph nodes. Therefore, the lymphatic system plays an important role in cancer biology [5]. The formation of new lymphatic vessels (lymphangiogenesis) occurs through several steps, including the migration, proliferation and sprouting of lymphatic endothelial cells, which are triggered by vascular endothelial growth factor receptor (VEGFR)-3, VEGF-C or VEGF-D [6]. The lymphatic vessel density (LVD) is the parameter that is most frequently used to quantify tumor lymphangiogenesis, especially for melanoma [7], oral squamous cell carcinoma [8], thyroid carcinoma [9], colorectal cancer [10], breast cancer [11], and lung cancer [12].
Previous studies have identified novel molecular markers of the lymphatic endothelium that have been used to study tumor-associated lymphangiogenesis via immunochemistry. These markers include VEGFR-3, Lymphatic vessel endothelial hyaluronan receptor-1 (LYVE-1), D2-40, podoplanin, Prox-1 and desmoplakin, among others [13][14][15][16]. VEGFR-3, also known as Flt4, is a member of the fms-like tyrosine kinase family, and it specifically binds VEGF-C and VEGF-D. LYVE-1 is a homolog of the vascular endothelium-specific hyaluronan receptor CD44 [17]. The antibody against D2-40 has been shown to specifically recognize the M2A antigen and podoplanin [18,19]. Podoplanin is a glomerular podocyte membrane mucoprotein [20]. The transcription factor prox-1 is a homolog of the drosophila homeobox gene product that is involved in the regulation of early lymphatic development [21]. Desmoplakin, also known as desmosome-related transmembrane protein, is a desmosomal protein expressed at intercellular junctions. Some studies have shown that lymphatic endothelium markers can be used to predict poor prognoses in NSCLC patients [22][23][24], but other studies have refuted this view [25][26][27]. Therefore, whether LVD is a prognostic biomarker for the survival of NSCLC patients remains controversial. The aim of this meta-analysis was to examine whether LVD can predict the prognosis of Asian NSCLC patients.

Study selection
The inclusion criteria were as follows: 1) a cohort study; 2) an Asian study population; 3) diagnosis of NSCLC based on lung histology, with the most important histological types being ADC, SCC and large cell cancer (LCC); 4) evaluation of the association between LVD and the prognosis of NSCLC patients; 5) analysis of lymph microvessel markers by immunohistochemistry; and 6) the presence of sufficient data to calculate the adjusted hazard ratio (HR) or risk ratio (RR) and the corresponding 95% confidence intervals (CIs). Studies were excluded if they had non-human study subjects. If the data were duplicated or the same population was used in more than one study, we chose the most recent or complete study.

Data extraction
The eligible studies selected for our meta-analysis were independently evaluated by two reviewers (XXQ and XSL) based on the aforementioned selection criteria. The following information was extracted from the eligible studies: the name of the first author, publication year, study period, country, sample number, sex of patients, median follow-up period, mean age or age range of patients, histology, histological type, TNM stage, and lymphatic endothelium markers (in Table 1). In addition, HR and 95% CIs were evaluated. Two authors (TZW and XSY) summarized the extracted data. Any disagreements were resolved by discussion.

Statistical analyses
To compute a pooled HR with a 95% CI, the Q-test and the I 2 test were used to assess heterogeneity among the studies [28]. We also calculated P values for the Q-test, which represented heterogeneity; heterogeneity was present if the P value was less than 0.10. The random-effects model was applied when I 2 > 50% [29]; otherwise, the fixed-effects model was applied [30]. Subgroup analyses based on lymphatic endothelium markers were performed to further explore the source of heterogeneity. Additionally, Begg's rank correlation test and Egger's linear regression test were conducted to assess the extent of potential publication bias [31]. Finally, a sensitivity analysis was performed by sequentially omitting one study per cycle to evaluate the stability of the results [32]. The data analyses were conducted using the STATA statistical software version 12.0 (STATA Corp. LLC, College Station, TX, USA).
The characteristics of the 15 eligible studies are shown in Table 1. These studies included 1075 participants from Asia, including Japan and China;a total of 11 studies investigated NSCLC, 3 studies investigated ADC, and 1 study investigated lung cancer. All studies used immunohistochemistry to assess LVD using different lymphatic endothelium markers, including VEGFR-3 in 5 studies, LYVE-1 in 3 studies, D2-40 in 5 studies, podoplanin in 2 studies.

Subgroup meta-analysis
The results of subgroup analyses using the lymphatic endothelium markers that were selected to evaluate LVD via immunohistochemistry support our findings. A positive relationship was observed between the expression of lymphatic endothelium markers and the prognoses of NSCLC patients (p = 0.000, I 2 = 73.2%). No statistically significant heterogeneity was observed in the 5 VEGFR-3 group (p = 0.709, I 2 = 0.0%); however, there was considerable heterogeneity in the 3 LYVE-1 groups (p = 0.01, I 2 = 86.4%), NSCLC non-small cell lung carcinoma, SCLC small cell lung carcinoma, ADC adenocarcinoma, SCC squamous cell cancer, LCC large cell cancer, VEGFR-3 vascular endothelial growth factor receptor-3, LYVE-1 lymphatic vessel endothelial receptor 1, y year, ND no data the 5 D2-40 groups (p = 0.019, I 2 = 66.2%) and the 2 podoplanin groups (p = 0.094, I 2 = 64.5%) (Fig. 3). Nevertheless, the data were not sufficient to determine the prognostic value of LVD among Asian populations based on sex, median follow-up period, mean age or age range, histological type, or TNM stage.

Sensitivity analysis
To evaluate the robustness of our analysis, we conducted a sensitivity analysis by recalculating the pooled results from the primary analyses after excluding one study per iteration. None of the studies when excluded altered the overall combined results (Fig. 4).

Discussion
NSCLC is the most common subtype of lung cancer, with a high incidence, high mortality, low survival rate, low diagnosis rate and treatment rate. It has been challenging to improve survival rates due to the lack of precise prognostic markers. To overcome this problem, a comprehensive understanding of lymphatic endothelium markers is needed. It is important to examine whether LVD can be an indicator of the prognosis in Asian NSCLC patients. In our present meta-analysis, LVD was positively associated with the prognosis of NSCLC (HR: 1.14, 95% CI: 1.02-1.27), indicating that high LVD indeed predicts poor survival in Asian NSCLC populations. To date, only Wang and colleagues [42] have described the relationship between LVD and the prognoses of NSCLC patients worldwide. Nevertheless, there was considerable heterogeneity among the included studies, which may make the results unreliable. However, sensitivity analysis did not reveal the source of heterogeneity. Furthermore, subgroup analyses were conducted using lymphatic endothelium markers. Additionally, publication bias was detected. Our study only focused on Asian patients, and thus our results are applicable for Asian populations. Although heterogeneity was also observed, the findings  were stable and robust based on our sensitivity analysis. In addition, subgroup analyses were performed based on the four lymphatic endothelium markers to further explore the origin of heterogeneity. Except VEGFR-3, the other three markers gave rise to considerable heterogeneity. Our meta-analysis included five additional studies that were more recent than those included in the study by Wang and colleagues. Moreover, no publication bias was observed in our study. The study by Zheng and colleagues [43] showed that the VEGF family is important for tumorigenesis and metastasis and that high VEGF and/or VEGFR expression, especially VEGF-C/VEGFR-3 co-expression, is indicative of poor survival in patients with NSCLC. However, that study did not evaluate other lymphatic endothelium markers, which were included in subgroup analyses in our study.
The role of LVD as a prognostic predictor in NSCLC remains controversial. Kajita and colleagues were the first to report VEGFR-3 expression in lung cancer cells, but they did not evaluate its impact on the prognosis or   [44]. Later, many studies demonstrated that VEGF-C, VEGF-D, VEGFR-3 and other markers are independent markers of poor prognostic in patients with NSCLC. Thus, these markers may be ideal targets for diagnosis or therapy to improve the prognosis of NSCLC patients [34]. The study by Arinaga demonstrated that the combined expression of VEGF-C and VEGFR-3 has a negative impact on the prognosis of patients with NSCLC [27]. In addition, the study by Zhang and colleagues revealed that D2-40-positive peritumoral LVD may be an independent prognostic factor for lung adenocarcinoma. Thus, D2-40-positivity may be used to predict patient prognosis in lung adenocarcinoma. Moreover, the reduction of peritumoral lymphangiogenesis has been suggested to inhibit the metastasis of lung adenocarcinoma [22]. However, some studies have claimed that high LVD may be a marker for good prognosis. The study by Nunomiya and colleagues showed that lung cancer patients with lower LYVE-1 levels have poorer prognoses than patients with higher LYVE-1 levels [25]. Yang and his team demonstrated the role of the epigenetic regulation of desmoplakin in increasing the sensitivity of cancer cells to anticancer drug-induced apoptosis, implying the clinical value of desmoplakin for the treatment of patients with lung cancer [45]. Nevertheless, more studies are needed in the near future to verify whether LVD is indicative of good or bad prognosis in NSCLC patients.
VEGFR-3, D2-40, LYVE-1 and podoplanin are widely used and extremely valuable markers of lymphatic vessels. However, one study has reported that lymphatic  endothelium markers are not only expressed on lymphatic vessels but also expressed on blood vessels, tumor cells or in normal tissues [13]. One of the major drawbacks is the lack of specific markers for the lymphatic endothelium. One study [46] indicated that LYVE-1 and Prox-1 are molecular markers of lymphangiogenesis in NSCLC and that they can be used as important markers for the evaluation of lymphatic metastasis and prognoses in patients with NSCLC. Another study [43] showed that high VEGF and/or VEGFR expression is indicative of poor survival in patients with NSCLC and that VEGF-C/ VEGFR-3 co-expression is a better prognostic indicator than other markers. Therefore, the evaluation of co-expressed markers may be useful to determine LVD.
Irrespective of its strengths, the meta-analysis also has certain limitations. First, although we searched all retrospective studies for the association between LVD and the prognosis of NSCLC, the eligible studies were restricted to those published in English or Chinese. Because of linguistic barrier, some non-English or non-Chinese studies were excluded. In addition, we also missed some studies that may have been published in books or journals that were not available in the online databases. Additionally, studies with negative data may not have been submitted by investigators, or studies with nonsignificant results may have been rejected by journals. Nevertheless, there was no significant publication bias in our study, although we could not completely rule out publication bias. Second, few studies did not present clear or complete data, making data analysis difficult. When we could not obtain original data from the authors via email or other means, we had to exclude those studies. Third, because of the small number of eligible articles, our study was not the most comprehensive. Fourth, our results cannot be generalized to populations worldwide, especially non-Asian populations. Thus, more comprehensive and higher quality analyses are still required in the future.

Conclusions
In summary, this meta-analysis indicated that LVD is an indicator of the prognosis of Asian NSCLC patients. However, higher quality and more comprehensive analyses are still needed as more data are published in the future.