Abstract:

Objective　To study the clinical value of the risk of ovarian malignancy algorithm (ROMA) index in the diagnosis of ovarian cancer through the meta-analysis on the ROMA index. Methods　The literatures on the ROMA index in the diagnosis of ovarian cancer were searched from PubMed database and CNKI from January 2009 to December 2021. Two reviewers screened the literatures and extracted the data according to the established inclusion and exclusion criteria. After the quality evaluation of the included literatures was carried out according to the QuadAS-2 scale, the statistical software Stata 12.0 was used for the meta-analysis. Results　A total of 3 167 subjects were included in the 13 literatures, including 1 391 subjects in the ovarian cancer group and 1 776 subjects in the control group (including the patients with benign ovarian disease and healthy people who underwent physical examination). The meta-analysis showed that the literatures had heterogeneity (P≤0.1,I²>50%); the ROMA index in the diagnosis of ovarian cancer had a combined sensitivity of 0.91 (95%CI 0.88-0.94), a combined specificity of 0.93 (95%CI 0.83-0.97), a combined positive likelihood ratio of 12.0 (95%CI 5.1-30.4), a combined negative likelihood ratio of 0.09 (95%CI 0.07-0.13), and a combined DOR of 90.76 (95%CI 46.12-178.60); the probability of diagnosis was 26%, and the probability of missed diagnosis was 0. Conclusion　This meta-analysis confirms that the risk of ovarian malignancy algorithm has some value in diagnosis of ovarian cancer.

Key words:

Risk of ovarian malignancy algorithm, Carbohydrate antigen 125, Human epididymal protein 4, Ovarian cancer, Meta-analysis

摘要：

目的　通过对卵巢癌风险预测模型（ROMA）指数进行meta分析，研究ROMA指数在诊断卵巢癌中的临床价值。方法　计算机检索PubMed数据库、中国知网（CNKI）中关于ROMA指数诊断卵巢癌的文献，检索时间为2009年1月至2021年12月。两位评审员根据制定的纳入和排除标准对文献进行筛选并提取信息，使用QUADAS-2量表评估所纳入文献的质量后，采用统计软件Stata 15.1进行Meta分析。结果　最终13篇文献中共3 167例受试者被纳入研究，其中卵巢癌组共1 391例样本，对照组（包括卵巢良性疾病患者和健康体检人群）共1 776例。采用meta分析，文献的异质性分析结果显示存在异质性（P≤0.1，I²>50%），ROMA指数在卵巢癌诊断中的合并敏感度为0.91（95%CI 0.88～0.94），合并特异度为0.93（95%CI 0.83～0.97），合并阳性似然比为12.0（95%CI 5.1～30.4），合并阴性似然比为0.09（95%CI 0.07～0.13），合并诊断优势比（DOR）为90.76（95%CI 46.12～178.60），诊断概率为26%，漏诊概率为0。结论　ROMA指数对卵巢癌具有较高的诊断价值。

关键词:

卵巢癌风险预测模型, 糖类抗原125, 人附睾蛋白4, 卵巢癌, meta分析