基于细胞因子水平和机器学习预测免疫性不孕患者治疗效果  

Prediction of Prognosis in Women with Infrtility due to Immune Causes Based on Cytokine Levels and Machine Learning

作　　者：杨婧 仲子星 倪万茂[3] YANG Jing;ZHONG Zi-xing;NI Wan-mao(Laboratory Medicine Center,Department of Clinical Laboratory,Zhejiang Provincial People's Hospital(Affiliated People's Hospital,Hangzhou Medical College),Hangzhou 310014,China;Center for Reproductive Medicine,Department of Obstetrics,Zhejiang Provincial People's Hospital(Affiliated People's Hospital,Hangzhou Medical College),Hangzhou 310014,China;Cancer Center,Department of Pathology,Zhejiang Provincial People's Hospital(Affiliated People's Hospital,Hangzhou Medical College),Hangzhou 310014,China)

机构地区：[1]浙江省人民医院检验中心,杭州310014 [2]浙江省人民医院妇产科,杭州310014 [3]浙江省人民医院病理科,杭州310014

出　　处：《中国生物工程杂志》2023年第12期169-176,共8页China Biotechnology

摘　　要：目的:探讨免疫性不孕患者与健康女性之间细胞因子的差异,以及细胞因子水平与免疫性不孕患者治疗效果的关系。方法:96例女性免疫性不孕患者和57例健康女性分别作为实验组和对照组纳入研究,通过流式细胞术检测了7种细胞因子的水平,并进行了为期2年的生育状况追踪。结果:实验组和对照组血清中的IL-2、IL-4、IL-6、IL-10、TNF-α和IFN-γ等6种细胞因子水平存在显著差异(P<0.05),实验组中治疗效果不佳的非妊娠组IL-2、IL-6、TNF-α、IFN-γ、IL-17A水平均高于妊娠组(P<0.05)。基于细胞因子数据和患者的生育结局,以PyCaret库成功构建了一种新的免疫性不孕治疗效果预测模型。经评估,装袋二次判别分析(Bagging QDA)模型在训练集上表现最佳,测试集上的灵敏度为72.73%,特异度为81.25%,准确度为72.41%。结论:免疫性不孕及其治疗效果均与Th1/Th2细胞因子异常相关,而Bagging QDA模型在预测免疫性不孕的预后方面具有较高准确性。In order to investigate the difference of cytokines between immune infertility patients and healthy women,and the relationship between cytokine levels and treatment outcomes in immune infertility patients,96 women with female immune infertility and 57 healthy women were enrolled as the experimental and control groups,respectively.The levels of 7 cytokines were measured by flow cytometry,and fertility status was followed for 2 years.The results showed that there were significant differences in the serum levels of 6 cytokines including IL⁃2,IL⁃4,IL⁃6,IL⁃10,TNF⁃αand IFN⁃γ,between the experimental group and the control group(P<0.05).The levels of IL⁃2,IL⁃6,TNF⁃α,IFN⁃γ,and IL⁃17A in the non⁃pregnant group with poor treatment effect in the experimental group were significantly higher than those in the pregnant group(P<0.05).Based on cytokine data and reproductive outcomes of patients,a new prognostic prediction model for immune infertility was successfully constructed with the PyCaret library.The bagging quadratic discriminant analysis(Bagging QDA)model performed best on the training set,with a sensitivity of 72.73%,specificity of 81.25%,and accuracy of 72.41%on the test set.In summary,immune infertility is associated with Th1/Th2 cytokine abnormalities,while the Bagging QDA model has high accuracy in predicting the prognosis of immune infertility.

关 键 词：生育 免疫 细胞因子 机器学习 

分 类 号：Q132[生物学—普通生物学]