Introduction

Osteosarcoma is the most frequent malignant osteogenic tumor, mostly occurring in children and young adults [1]. Over the past decade, the clinic appliance of neoadjuvant reduced the size of the localized tumor and delayed the progression, significantly improving the 5-year survival rate of patients with low-grade osteosarcoma [2]. However, metastasis has been reported to be present in approximately 25% of newly diagnosed osteosarcoma patients, and the mortality rate in these patients remains extremely high at approximately 20% [2,3,4]. There is currently an absence of viable methods for the early diagnosis and treatment of osteosarcoma. Given this, further investigation of prognostic molecular biomarkers is critical for a better understanding of osteosarcoma's pathophysiology and the development of more effective treatment modalities.

Angiogenesis plays a vital role in tumor development as the growth of tumors relies on the perfusion of neovascular [5]. Vascular endothelial growth factor (VEGF) is a potent pro-angiogenic factor that regulates vascular endothelial cell proliferation, differentiation and migration [6]. Overexpression of VEGF has been reported to be attributed to the invasion and metastasis of a wide range of solid tumors [7,8,9]. To clarify the mechanism of VEGF in the advancement of osteosarcoma, the association between VEGF and prognosis features of osteosarcoma has been assessed. However, the prognostic and clinicopathological value of VEGF remains controversial [10, 11]. Previous relevant meta-analyses have been performed to define the clinical significance of VEGF expression in osteosarcoma. Nevertheless, these analyses were inconclusive as inconsistent results, limited involved studies, and the absence of a thorough evaluation of study quality and pooled results [12,13,14]. Therefore, this current study aimed to comprehensively and systematically assess the prognostic value of VEGF in 22 studies involving 1144 osteosarcoma patients.

Materials and methods

This study was conducted entirely in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [15].

Search strategies

A comprehensive electronic literature search was performed in four databases: Web of Science, PubMed, Cochrane Library and Medline with no restrictions on language or publication date. The last search was conducted on September 12, 2021. The following terms were used to conduct the literature search: ("osteosarcoma" or "osteogenic sarcoma") and ("vascular endothelial growth factor" or VEGF). We additionally manually screened the references of identified articles to collect more studies.

Selection criteria

The eligible articles were selected in accordance with the following criteria: (1) patients were diagnosed with osteosarcoma pathologically; (2) the relationship between VEGF expression and clinicopathological characteristics or prognosis were investigated; (3) the expression VEGF was determined on samples of tumor tissue. Articles were excluded according to the following criteria: (1) studies were published in the form of conference abstracts, letters, case reports, expert opinions, reviews, or sequence data; (2) focused on tumor cell lines or animal experiments; (3) patients did not confirm the diagnosis of osteosarcoma; (4) when study comprised overlapping patient cohort. Two independent authors determined whether studies were eligible. Any discrepancies were settled by consensus following a discussion.

Data extraction and quality assessment

Two independent investigators carefully reviewed all eligible publications to extract interested data. The following data were collected, including (1) first author, publication year, patient source; (2) number of patients, age, gender, VEGF assay method, antibody type, source and dilution of immunohistochemistry (IHC), and cutoff value; (3) tumor stage at diagnosis, metastasis, local recurrence, tumor location, chemotherapy response, microvessel intensity (MVD), hazard ratio (HR) of VEGF expression and corresponding 95% CI.

If a study stated both univariate and multivariate survival results, the HRs from the multivariate analyses were used. When the survival results were not given explicitly while a Kaplan–Meier curve was present, the HRs with 95% CIs were retrieved using Engauge Digitizer 11.0 software and Tierney’s reported method [16].

Each involved study’s quality was assessed using the Newcastle–Ottawa Scale (NOS) by two independent reviewers [17]. The scale judges the quality of studies from three main aspects: the selection of the groups, comparability, and exposure, with a maximum of nine points. Articles with a NOS score of more than six were considered to be of high quality.

Statistical analysis

The statistical analysis in this study was performed by STATA 14.0 (Stata Corporation, College Station, TX, USA). We estimate the pooled HRs for survival results, the pooled odds ratio (OR) for the clinicopathological characteristics (age, gender, stage, metastasis, local recurrence, response to chemotherapy). The continuous variables are described as standard mean difference (SMD). The statistical between-study heterogeneity was assessed by the Chi-squared test and the Higgins I2 statistic. Significant heterogeneity was defined as a p > 0.10 or I2 > 50%. A fixed-effects model was utilized when there was no significant heterogeneity. Otherwise, a random-effects model was utilized. The potential publication bias was estimated by using Begg’s funnel plot and Egger’s test. Additionally, we performed sensitivity analyses to assess the stability of the pooled outcomes. p < 0.05 was considered statistically significant.

Results

Search results

A total of 2075 articles were identified from four online databases. After removing 640 duplicates, the remaining 1435 records were systematically evaluated by the titles and abstracts. Among these articles, 165 articles were excluded for irrelevant studies, 245 articles involved non-human experiments, 492 articles were conference abstracts, case reports, letters, and reviews, and 501 articles were not related to VEGF or osteosarcoma. After assessing the entire text of the remaining 32 studies, 10 articles were excluded for insufficient data. Finally, 22 studies with a total of 1144 osteosarcoma patients were included in this study [10, 11, 18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37]. The detailed flowchart of the study filtrating process is shown in Fig. 1.

Fig. 1
figure 1

The flowchart of the study selection in this meta-analysis

Full size image

Study characteristics and quality assessment

The summarized characteristics of the included study are shown in Table 1. Among them, 11 studies focused on the prognostic significance, 22 studies analyzed the correlation between VEGF expression and clinicopathological characteristics. All of the eligible research was published between 1999 and 2020, and it was written in English, with a patient population ranging from 25 to 153. Additionally, immunochemical staining (IHC) was the most often employed technique to measure VEGF expression (21/22, 95.5%), with 95.2% of the studies using IHC having defined the cutoff value of VEGF expression. Each article used the tissue as the sample. In terms of study quality, all of the eligible studies were high quality with a NOS score greater than 6 points. Other information about the involved studies is shown in Table 1.

Table 1 Main characteristics of the studies included in this meta-analysis
Full size table

VEGF expression and prognostic significance

The survival data, including overall survival (OS) or disease-free survival (DFS), were analyzed in 11 studies among eligible studies. Due to the lack of evident heterogeneity detected (I2 = 0.00%, p = 0.894), a fixed-effects model was utilized. The result showed that the elevated VEGF expression was associated with poor overall survival (HR, 2.42; 95% CI, 1.87–3.11, p < 0.001). To further find out the potential sources of heterogeneity, we undertook a subgroup analysis stratified by ethnicity, publication date, testing isoform, antibody type, positive rate, HR resource, sample size, and NOS score. As shown in Table 2, each subgroup presented a significant association with overall survival. Besides, disease-free survival was also extracted in studies. The fixed-effect model was employed to calculate the pooled HR (I2 = 0.00%, p = 0.485). Results reveal that the elevated VEGF expression predicted poor disease-free survival (HR, 2.604; 95% CI, 1.698–3.995, p < 0.001).

Table 2 The subgroups analysis for VEGF and overall survival in patients with osteosarcoma
Full size table

VEGF expression and clinicopathological features

The correlation between VEGF expression and clinicopathological values, including age, gender, metastasis, local recurrence, tumor stage, response to chemotherapy, and MVD, was investigated. A fixed-effects or a random-effects model was employed based on the heterogeneity results of each parameter. The detailed information is shown in Table 3. Under the fixed-effects model, overexpression of VEGF was significantly related to a higher rate of osteosarcoma metastasis (OR, 4.39; 95% CI, 2.77–6.95; p < 0.001). The random-effects model showed that the overexpression of VEGF was significantly related to a higher clinical stage (OR, 0.73; 95% CI, 0.62–0.87; p < 0.001). Besides, VEGF expression showed a significant correlation with microvessel density (MVD) according to the results of the random-effects model (SMD, 3.33, 95% CI,1.57–5.10, p < 0.001). However, we failed to find a significant relationship between overexpression of VEGF and gender, tumor location, local recurrence, age, and response to chemotherapy (Fig. 2).

Fig. 2
figure 2

Forest plots of the association between VEGF overexpression and clinicopathological features. A Distant Metastasis. B Clinical stage. C Tumor location D MVD. E Gender. F Local recurrence. G Age. H Chemotherapy response. OR, odds ratio; CI, confidence intervals; MVD, microvessel density

Full size image
Table 3 Pooled odds ratios of VEGF on clinicopathologic features in osteosarcoma
Full size table

Publication bias and sensitivity analysis

Publication bias was measured by using Begg's funnel plot and Egger's tests. As shown in Figs. 3B and 4B, there was no publication bias for overall survival (Begg's test, p = 0.436; and Egger's test, p = 0.745) and disease-free survival (Begg's test, p = 0.089; and Egger's test, p = 0.198).

Fig. 3
figure 3

Pooled analysis for the association between VEGF overexpression and Overall survival. A Forest plots. B Funnel plots. C Sensitive analysis. OS, overall survival; OR, odds ratio; CI, confidence intervals; s.e., standard error

Full size image
Fig. 4
figure 4

Pooled analysis for the association between VEGF overexpression and disease-free survival. A Forest plots. B Funnel plots. C Sensitive analysis. OS, overall survival; OR, odds ratio; CI, confidence intervals; s.e., standard error

Full size image

We performed a sensitivity analysis of overall survival and disease-free survival to investigate the influence of each study on the pooled HR. As Figs. 3C and 4C show, we did not find any significant alteration in the pooled HR when omitting any single study sequentially, demonstrating that the analyses were stable and credible.

Discussion

As the most frequent primary osteogenic tumor, osteosarcoma is characterized as aggressive cancer with a high risk of distant metastasis. Although immune checkpoint inhibitors have recently revolutionized the treatment of a wide range of solid malignancies, they have demonstrated limited efficacy in osteosarcoma [38,39,40]. Therefore, the identification of other biomarkers related to the prognosis of osteosarcoma is crucially essential to the development of new potential therapeutic targets.

Angiogenesis is essential for the proliferation and metastasis of tumor cells [5]. In the past decades, VEGF has been the most studied biomarker of tumor neovascularization for its crucial significance in angiogenesis and vasculogenesis [41]. Through binding to tyrosine kinases receptors, the VEGF signaling pathways play an important role in a variety of physiological and pathological processes. In the process of tumorigenesis, the transcription of several hypoxia-related genes induces the expression of VEGF, mainly via VEGFR-2, to activate angiogenesis [42]. High levels of VEGF expression are linked to endothelial barrier disruption in pathological tumor conditions, promoting cancer distant metastasis [43, 44]. Furthermore, VEGF is involved in regulating the immune response of tumors. A variety of innate immune cells have been reported to secrete VEGF in the tumor microenvironment to reduce the immune response of immune cells to tumor tissue [45,46,47]. The upregulated VEGF expression was also reported to actively participate in tumor escape from immune surveillance by suppressing the proliferation of T-cells and increasing the exhaustion of T-cells [48, 49].

Recently, it has been implicated that high expression of VEGF mediates metastasis and progression in many malignancies [7,8,9]. Several meta-analyses have previously assessed the clinical significance of VEGF expression in patients with osteosarcoma [12,13,14]. Nevertheless, Han et al. focused on the part of the clinicopathological characteristics of VEGF [12]. Researches on the prognostic effect of VEGF expression had inconsistent results and did not pay attention to the quality evaluation, heterogeneity, and sensitivity analysis [13, 14]. Moreover, these researches were published 5 years ago. Limited to the relatively small number of studies, the conclusion was not robust, and some crucial clinicopathological features were not evaluated. Here, we conducted a comprehensive literature search to combine all relevant studies related to VEGF expression's prognostic and clinicopathological value.

In the present meta-analysis study, we pooled 22 studies on VEGF expression in the prognosis or clinicopathology of osteosarcoma patients. In terms of survival data, our findings revealed that overexpression of VEGF was associated with poor overall survival and disease-free survival. The analyses did not find significant heterogeneity or obvious publication bias, and sensitivity analysis showed our results were robust and reliable. Therefore, we supported the hypothesis that elevated VEGF expression predicted poor DFS. In terms of clinicopathological characteristics, similar to previous reports, VEGF overexpression was related to a higher tumor grade and rate of metastasis but not associated with gender, age, tumor location, local recurrence, clinical stage and response to chemotherapy [12]. The results indicated that high levels of VEGF expression predict metastasis and an advanced stage of osteosarcoma. Additionally, previous meta-analyses had not assessed the association between VEGF and MVD. In our study, VEGF overexpression had a marked effect on promoting vascularization in osteosarcoma. However, the results should be interpreted cautiously as only limited studies were included in the analyses, and more related research is needed.

This meta-analysis has some limitations. Firstly, the methods for identifying and evaluating VEGF expression varied among the eligible studies. Although most of these studies applied IHC, the varied antibodies and dilutions utilized may have contributed to heterogeneity. In addition, there were discrepancies in the definition of VEGF positive. The staining methods, the details of the IHC scoring criteria, and cutoff values varied across the included studies. Secondly, the correlation between VEGF expression and some clinicopathological characteristics of osteosarcoma, such as tumor size, were not analyzed in our study due to the insufficient studies using the same criteria of tumor size. Furthermore, when the results of the multivariate survival analysis were reported, the survival data were extracted directly. When not stated in the original articles, the HRs with their corresponding 95% CIs were calculated through the reconstruction of survival curves, which may affect the robustness of the pooled overall survival and disease-free survival. In order to eliminate bias, more precise data extraction methods or better study quality were needed. Lastly, although this study comprised more than 1000 osteosarcoma patients, future studies with larger sample sizes are necessary to further elucidate the association between VEGF and prognosis and clinicopathological characteristics.

Conclusion

This meta-analysis indicated that elevated VEGF expression was correlated with adverse osteosarcoma clinicopathological features and poor prognosis. Our results suggest that VEGF is a predictive biomarker in patients with osteosarcoma. However, further large-scale, prospective research is required to validate our results.