摘要
目的长链非编码RNA(lncRNA)在卵巢癌中各种作用引起了越来越多的关注。研究确定基于lncRNA的特征,用于建立卵巢癌患者的生存预测。方法从基因表达数据库(GEO)下载GSE9891卵巢癌285例基因表达数据集及相应的临床数据。采用单因素Cox回归分析筛选与卵巢癌总生存率(OS)显著相关的lncRNA,并用最小绝对值收敛和选择算子(LASSO)回归进一步缩减变量。最终采用多因素Cox回归分析确定的lncRNA构建风险评分系统。所有患者根据风险评分的中位值分为高或低风险组。评分系统与临床变量合并后再用多因素Cox回归分析识别出独立的预后因素来建立列线图。采用区分度、校准图和临床实用性用于评价列线图的预测准确度。结果采用多因素Cox回归分析共获得15个与卵巢癌预后显著相关的lncRNA(NORAD、DHRS4-AS1、SNHG18、LINC00174、PCAT6、HOXA-AS2、PCAT19、DOCK9-DT、HDAC4-AS1、TENM3-AS1、DPP10-AS1、FAM201A、RBMS3-AS3、LINC00558、LINC01527),并用于建立一个风险评估系统。风险评分系统成功地将患者分为高风险组和低风险组。最终确定了紫杉醇治疗、国际妇产联合会(FIGO)分期和15个lncRNA特征作为卵巢癌的独立预后因子。结合这些预测因子的列线图显示良好的区分度及校准图,C-index为0.76(95%可信区间为0.717~0.803),优于FIGO分期(C-index为0.65;95%可信区间为0.600~0.700)。此外,决策分析曲线表明列线图提供了优于FIGO分期系统的优势,带来了更多的临床净收益。结论基于lncRNA的风险评分系统能明显区分高风险组和低风险组卵巢癌患者的生存率。新建立的列线图可以较FIGO分期系统更有效地预测卵巢癌患者的总生存率。该列线图较高的预测能力可以帮助临床医生制定合适的个体治疗并进行个性化的预后评估。
Objective To establish survival prediction of ovarian cancer(OC) patients based on characteristics of long non-coding RNA(lncRNA), in the consideration of effects of lncRNA in OC. Methods Gene expression data of 285 cases of GSE9891 OC and corresponding clinical data were downloaded from Gene Expression Omnibus(GEO). The univariate Cox regression analysis was chosen to screen lncRNA that were significantly associated with overall survival(OS) of OC, and least absolute value convergence and selection operator(LASSO) regression were used to reduce variables. Finally, lncRNA determined by multivariate Cox regression analysis was used to construct risk scoring system. All patients were divided into high risk group and low risk group based on median risk score. The scoring system was combined with clinical variables and then multivariate Cox regression analysis was used to identify independent prognostic factors and establish nomogram. The discrimination,calibration chart and clinical applicability were used to evaluate nomogram prediction accuracy. Results A total of 15 lncRNA(NORAD, DHRS4-AS1, SNHG18, LINC00174, PCAT6, HOXA-AS2, PCAT19, DOCK9-DT, HDAC4-AS1, TENM3-AS1, DPP10-AS1, FAM201 A, RBMS3-AS3, LINC00558, LINC01527) significantly related to prognosis of OC were obtained from multivariate Cox regression analysis, and used to establish risk assessment system. The risk scoring system successfully stratified patients and divided the patients into high risk group and low risk group. The paclitaxel treatment, International Federation of Obstetrics and Gynecology(FIGO) staging and 15 lncRNA characteristics were finally identified as independent prognostic factors for OC. The nomogram combined with predictive factors showed good discrimination and calibration chart.The C-index was 0.76(95 % confidence interval was 0.717-0.803), which was better than FIGO staging system(C-index was0.65;95 % confidence interval was 0.600-0.700). The decision analysis curve showed that nomogram was better than FIGO staging system, and with more clinical net benefits. Conclusion The lncRNA-based risk scoring system could clearly distinguish OS of OC patients with high risk group and low risk group. The novel proposed nomograms could predict OS of OC patients more effectively than FIGO staging system, which helps to make appropriate individual treatment plans and conduct personalized prognostic assessments.
作者
张雪梅
高琴
杨学
ZHANG Xue-mei;GAO Qin;YANG Xue(Department of Obstetrics and Gynecology,Puren Hospital,Wuhan University of Science and Technology,Wuhan 430081,Hubei,China)
出处
《生物医学工程与临床》
CAS
2020年第3期322-327,共6页
Biomedical Engineering and Clinical Medicine
