摘要
       血管生成在促进肿瘤进展、侵袭性和转移方面起着至关重要的作用。虽然很少有研究表明一些血管生成相关基因（ARGs）可以作为前列腺癌症（PCa）的预后相关生物标志物，但ARGs的综合作用尚未得到系统研究。前列腺癌（PRAD）的RNA测序数据和临床信息作为发现数据集从癌症基因组图谱（TCGA）下载。通过单变量Cox回归分析，总共有23个ARG与PRAD的预后相关，并且通过具有十倍交叉验证的最小绝对收缩和选择算子（LASSO）Cox回归，进一步发展了19-ARG特征，该特征与PRAD无病生存率（DFS）显著相关。特征将PRAD患者分为高ARGs和低ARGs特征评分组，而那些ARGs高特征评分组的预后显著较差（中位DFS：62.71个月vs未达到，p < 0.0001）。随后在GSE40272和PRAD_MSKCC两个独立的队列中验证了ARGs信号的预测能力。值得注意的是，19-ARG特征在预测PRAD的DFS方面优于典型的临床特征或每种特征都涉及ARG。此外，用三个独立的预后因素构建了预后列线图，包括ARGs特征、T分期和Gleason评分。列线图的预测结果（C指数 = 0.799，95%置信区间 = 0.744–0.854）与观察到的结果很好地匹配，这通过校准曲线进行了验证。诺模图在1年、3年和5年时DFS的受试者工作特征曲线下面积（AUC）值分别为0.82、0.83和0.83，表明诺模图模型的性能具有相当高的准确性和稳健性。此外，功能富集分析证明了E2F靶标、G2M检查点途径和细胞周期途径抑制PRAD进展的潜在靶标。值得注意的是，高危PRAD患者对免疫疗法更敏感，但Treg可能会阻碍免疫疗法的益处。此外，这一已建立的工具还可以预测对新辅助雄激素剥夺疗法（ADT）和一些化疗药物的反应，如顺铂、紫杉醇和多西他赛等。新的ARGs信号具有预后意义，可以进一步促进靶向治疗在不同分层前列腺癌患者中的应用。
Abstract
Angiogenesis plays the critical roles in promoting tumor progression, aggressiveness, and metastasis. Although few studies have revealed some angiogenesis-related genes (ARGs) could serve as prognosis-related biomarkers for the prostate cancer (PCa), the integrated role of ARGs has not been systematically studied. The RNA-sequencing data and clinical information of prostate adenocarcinoma (PRAD) were downloaded from The Cancer Genome Atlas (TCGA) as discovery dataset. Twenty-three ARGs in total were identified to be correlated with prognosis of PRAD by the univariate Cox regression analysis, and a 19-ARG signature was further developed with significant correlation with the disease-free survival (DFS) of PRAD by the least absolute shrinkage and selection operator (LASSO) Cox regression with tenfold cross-validation. The signature stratified PRAD patients into high- and low-ARGs signature score groups, and those with high ARGs signature score were associated with significantly poorer outcomes (median DFS: 62.71 months vs unreached, p < 0.0001). The predicting ability of ARGs signature was subsequently validated in two independent cohorts of GSE40272 & PRAD_MSKCC. Notably, the 19-ARG signature outperformed the typical clinical features or each involved ARG in predicting the DFS of PRAD. Furthermore, a prognostic nomogram was constructed with three independent prognostic factors, including the ARGs signature, T stage and Gleason score. The predicted results from the nomogram (C-index = 0.799, 95%CI = 0.744–0.854) matched well with the observed outcomes, which was verified by the calibration curves. The values of area under receiver operating characteristic curve (AUC) for DFS at 1-, 3-, 5-year for the nomogram were 0.82, 0.83, and 0.83, respectively, indicating the performance of nomogram model is of reasonably high accuracy and robustness. Moreover, functional enrichment analysis demonstrated the potential targets of E2F targets, G2M checkpoint pathways, and cell cycle pathways to suppress the PRAD progression. Of note, the high-risk PRAD patients were more sensitive to immune therapies, but Treg might hinder benefits from immunotherapies. Additionally, this established tool also could predict response to neoadjuvant androgen deprivation therapy (ADT) and some chemotherapy drugs, such as cisplatin, paclitaxel, and docetaxel, etc. The novel ARGs signature, with prognostic significance, can further promote the application of targeted therapies in different stratifications of PCa patients.

https://link.springer.com/article/10.1186/s10020-022-00504-6