Kim Pil-Jong, Park Ji Y, Kim Hong-Gee, Cho Yong Mee, Go Heounjeong
Biomedical Knowledge Engineering Laboratory, Seoul National University School of Dentistry and Dental Research Institute, Seoul, Korea.
Department of Pathology, Catholic University of Daegu School of Medicine, Daegu, Korea.
BJU Int. 2017 Sep;120(3):343-350. doi: 10.1111/bju.13783. Epub 2017 Feb 10.
To identify new biomarkers for biochemical recurrence (BCR) of prostate adenocarcinoma.
Clinical information of 500 patients with prostate adenocarcinoma and their 152 RNA-sequencing and protein-array data from The Cancer Genome Atlas (TCGA) were separated into a discovery set and a validation set. Each dataset was analysed according to the Gleason grade groups reflecting BCR. The results obtained from the analysis using TCGA dataset were confirmed by immunohistochemistry analyses of a confirmation cohort composed of 395 patients with localised prostate adenocarcinoma.
TCGA discovery set was subgrouped into lower- and higher-risk groups for recurrence-free survival (RFS) (P < 0.001). Cyclin B1 (CCNB1), dishevelled segment polarity protein 3 (DVL3), paxillin (PXN), RAF1, transferrin, X-ray repair cross complementing 5 (XRCC5) and BIM had lower expression in the lower-risk group than that in the higher-risk group (all, P < 0.05). In TCGA validation set, CCNB1, DVL3, transferrin, XRCC5 and BIM were also differently expressed between the two groups. Immunohistochemically, DVL3 positivity was associated with high prostate-specific antigen (PSA) levels, resection margin involvement, and BCR (all, P < 0.05). A high Gleason score indicated a marginal relationship (P = 0.055). BIM positivity was related to high PSA levels, lymphovascular invasion, and BCR (all, P < 0.05). Both DVL3 positivity (P = 0.010) and BIM positivity (P = 0.024) were associated with shorter RFS, but statistical significance was lost when the multivariate Cox regression model included all patients. In the lower-risk group, the multivariate Cox model confirmed that DVL3 was an independent predictor for poor RFS (hazard ratio 1.80, P = 0.040), and the concordance index (C-index) was 0.805.
DVL3 and BIM were expressed in patients with a higher risk of BCR. DVL3 may be a novel and easily applicable recurrence predictor of localised prostate adenocarcinoma.
确定前列腺腺癌生化复发（BCR）的新生物标志物。
500例前列腺腺癌患者的临床信息以及来自癌症基因组图谱（TCGA）的152份RNA测序和蛋白质阵列数据被分为发现集和验证集。每个数据集根据反映BCR的Gleason分级组进行分析。使用TCGA数据集分析得到的结果通过对由395例局限性前列腺腺癌患者组成的验证队列进行免疫组织化学分析得以证实。
TCGA发现集被分为无复发生存期（RFS）的低风险组和高风险组（P<0.001）。细胞周期蛋白B1（CCNB1）、散乱段极性蛋白3（DVL3）、桩蛋白（PXN）、RAF1、转铁蛋白、X射线修复交叉互补蛋白5（XRCC5）和BIM在低风险组中的表达低于高风险组（均P<0.05）。在TCGA验证集中，CCNB1、DVL3、转铁蛋白、XRCC5和BIM在两组之间也存在差异表达。免疫组织化学分析显示，DVL3阳性与高前列腺特异性抗原（PSA）水平、手术切缘受累及BCR相关（均P<0.05）。高Gleason评分显示出边缘性关联（P=0.055）。BIM阳性与高PSA水平、淋巴管浸润及BCR相关（均P<0.05）。DVL3阳性（P=0.010）和BIM阳性（P=0.024）均与较短的RFS相关，但当多变量Cox回归模型纳入所有患者时，统计学意义消失。在低风险组中，多变量Cox模型证实DVL3是RFS不良的独立预测因子（风险比1.80，P=0.040），一致性指数（C指数）为0.805。
DVL3和BIM在BCR风险较高的患者中表达。DVL3可能是局限性前列腺腺癌一种新型且易于应用的复发预测指标。
