首页|基于决策树模型分析非预期卵巢低预后影响因素

基于决策树模型分析非预期卵巢低预后影响因素

扫码查看
目的 利用决策树模型分析不孕患者非预期卵巢低预后发生的影响因素。方法 回顾性分析北部战区总医院自2017 年1 月至2021 年12 月收治的行体外受精-胚胎移植助孕的1 112 例患者的临床资料。参照波塞冬标准将患者分为低预后组和正常反应组。比较两组患者的临床资料,包括年龄、体质量指数、不孕年限、基础窦卵泡数、基础卵泡刺激素、基础黄体生成素、抗缪勒管激素、基础雌二醇、促性腺激素起始量、促性腺激素总天数及促性腺激素总量等。再利用决策树模型分析发生非预期卵巢低预后的影响因素。结果 1 112 例患者中,正常反应组766 例,低预后组346 例。低预后组患者的黄体生成素、抗缪勒管激素、窦卵泡数、促性腺激素起始量均低于正常反应组,年龄、体质量指数、不孕年限、卵泡刺激素均高于正常反应组,差异有统计学意义(P<0。05)。决策树模型发现抗缪勒管激素是决策树的根节点,抗缪勒管激素水平是影响非预期卵巢低预后的最主要因素;次一级的影响因素为窦卵泡数、促性腺激素起始量、促性腺激素总量;最低一级影响因素为体质量指数。决策树模型准确率为81。65%。结论 决策树模型能够有效对患者进行分类,判断不同因素对患者的影响程度,具有较高的准确性,对临床决策具有一定的参考价值。
Influencing factors of unanticipated low ovarian prognosis by using decision tree model
Objective To analyze the influencing factors of unanticipated low ovarian prognosis in infertility patients by using decision tree model.Methods The clinical data of 1 112 patients admitted to General Hospital of Northern Theater Command from January 2017 to December 2021 who underwent in vitro fertilization-embryo transfer were retrospectively analyzed.Patients were divided into low prognosis group and normal response group according to Poseidon criteria.The clinical data of the two groups were compared,in-cluding age,body mass index,infertility years,basal sinus follicle number,basal follicle stimulating hormone,basal luteinizing hor-mone,anti-Mullerian hormone,basal estradiol,gonadotropin initiation,total gonadotropin days and total gonadotropin.Decision tree model was used to analyze the factors affecting the prognosis of unanticipated ovarian hypoplasia.Results Among 1 112 patients,766 were in the normal response group and 346 were in the low prognosis group.Luteinizing hormone,anti-Mullerian hormone,antral follicle number and gonadotropin initiation in low prognosis group were lower than those in normal response group,and age,body mass index,infertility years and follicle-stimulating hormone were higher than those in normal response group,with statistical significance(P<0.05).The decision tree model showed that anti-Mullerian hormone was the root node of the decision tree,and the level of anti-Mullerian hormone was the most important factor affecting the prognosis of unanticipated low ovary.The secondary influencing factors were the number of sinus follicles,initiation of gonadotropin and total gonadotropin.The lowest level of influence factor was body mass index.The accuracy of decision tree model was81.65% .Conclusion The decision tree model can effectively classify patients and determine the degree of impact of different factors on patients,with high accuracy and reference value for decision-making.

Decision tree modelPoseidon standardLow ovarian prognosisAssisted reproductive technologyAnti-Mulle-rian hormone

杜超、侯开波、关小川、王博伦、孙凯旋、于月新

展开 >

北部战区总医院 生殖医学科,辽宁 沈阳 110016

决策树模型 波塞冬标准 卵巢低预后 辅助生殖技术 抗缪勒管激素

辽宁省科技计划

2020JH2/10300118

2024

临床军医杂志
解放军沈阳军区卫生人员训练基地

临床军医杂志

CSTPCD
影响因子:0.465
ISSN:1671-3826
年,卷(期):2024.52(1)
  • 7