Purpose

To combine different independent endometrial markers to classify the presence of endometriosis.

Methods

Endometrial biopsies were obtained from 109 women with endometriosis as well as 110 control women. Nine candidate biomarkers independent of cycle phase were selected from the literature and NanoString was performed. We compared differentially expressed genes between groups and generated generalized linear models to find a classifier for the disease.

Results

Generalized linear models correctly detected 68% of women with endometriosis (combining deep infiltrating and ovarian endometriosis). However, we were not able to distinguish between individual types of endometriosis compared to controls. From the 9 tested genes, FOS, MMP7, and MMP11 seem to be important for disease classification, and FOS was the most over-expressed gene in endometriosis.

Conclusion(s)

Although generalized linear models may allow identification of endometriosis, we did not obtain perfect classification with the selected gene candidates.

中文翻译：

---

使用 NanoString 技术对子宫内膜异位症女性在位子宫内膜中假定的生物标志物进行外部验证

目的

结合不同的独立子宫内膜标志物对子宫内膜异位症的存在进行分类。

方法

子宫内膜活检取自 109 名子宫内膜异位症女性和 110 名对照女性。从文献中选择了九个独立于循环阶段的候选生物标志物，并进行了 NanoString。我们比较了组间差异表达的基因，并生成了广义线性模型以找到该疾病的分类器。

结果

广义线性模型正确检测到 68% 的子宫内膜异位症（结合深部浸润和卵巢子宫内膜异位症）。然而，与对照组相比，我们无法区分不同类型的子宫内膜异位症。在测试的 9 个基因中，FOS、MMP7和MMP11似乎对疾病分类很重要，而FOS是子宫内膜异位症中过表达最多的基因。

结论

尽管广义线性模型可以识别子宫内膜异位症，但我们没有对选定的候选基因进行完美的分类。