Glioblastoma (GBM) which is the most common malignant brain tumor with poor outcomes accounts for approximately 51 % of intracranial malignancies [1]. And glioblastoma has a 5-year overall survival of merely 5 % [1,2]. Glioblastoma is treated with surgery, temozolomide chemotherapy, radiation therapy administered six weeks after surgery, and customized immunotherapy based on tumor markers [3]. Despite advancements in surgery and adjuvant methods, patients with standard first-line therapy remain approximately 14.6 months of median survival time [3,4]. And the 5-year survival of GBM patients is below 10 % [5]. Its aggressive character accounts for most of its bad prognosis. This is a result of glioblastoma's high degree of vascularization, which allows the tumor to receive enough blood flow to eliminate waste and absorb nutrients [6,7]. Consequently, resistance of glioblastoma to current anti-vascular therapy is a serious issue. It's critical to identify targets that effectively inhibit glioblastoma angiogenesis.

The concept of vasculogenic mimicry (VM), which is described as a kind of novel vascular network structure for the treatment of malignant melanoma, was introduced by Maniotis et al. [8]. Highly aggressive tumor cells derive energy from various growth factors and signaling pathways by activating their receptors, thus contributing to the formation and development of vascular-like structures [9,10]. This alternative angiogenic pathway is also known as VM. The significance of the angiogenic pathway in fostering angiogenesis within GBM is notable [11,12]. VM has been identified in various primary malignancies. VM has been associated with reduced overall survival, heightened mortality rates, and enhanced tumor metastasis [10]. Consequently, recent anti-angiogenic therapy focuses more on suppressing VM [10]. However, antiangiogenic therapy has shown limited efficacy [13]. While inhibiting VM by blocking individual signaling pathways has limitations, GBM can reactivate angiogenic mimicry through alternative signaling channels [10]. Currently, there is limited knowledge regarding the clinical significance and application of VM-related genes. Recent research has identified Rho-associated coiled-coil protein kinase 1 (ROCK1) and tissue factor pathway inhibitor (TFPI) as significant genes associated with vascular mimicry (VM), facilitating malignant invasion in cancer [14,15]. Therefore, it is imperative to investigate the involvement of ROCK1 and TFPI in glioblastoma VM.

To investigate the prognosis of GBM, our study collected data from The Cancer Genome Atlas Program (TCGA) and the Chinese Glioma Genome Atlas (CGGA) databases. We used Kaplan-Meier survival analysis combined with univariate and multivariate COX regression analysis to identify meaningful VRGs. The VM index model and risk score model were constructed using these analyses. We discovered that ROCK1 and TFPI increase GBM metastasis through vasculogenic mimicry. Our findings emphasize the importance of ROCK1 and TFPI in VM. These findings suggest that ROCK1 and TFPI represent novel targets for GBM treatment and provide innovative strategies for limiting GBM progression. Moreover, these results facilitate the translation of our findings into clinical practice.