女童中枢性性早熟预测模型的建立及评价

Establishment and evaluation of prediction model of central precocious puberty in girls

目的 构建基于实验室指标、超声参数及临床因素的女童中枢性性早熟(CPP)预测模型。方法 选取2020年1月至2022年4月该院收治的103例性早熟女童作为研究对象,其中单纯乳房早发育(IPT)组62例,CPP组41例。比较两组基线资料、超声参数(子宫容积、卵巢容积、>4 mm卵泡数),以及血清促卵泡激素(FSH)、黄体生成素(LH)峰值、胰岛素样生长因子-1(IGF-1)、胰岛素生长因子结合蛋白(IGFBP-3)、25羟维生素D[25-(OH)D]水平。采用Logistic回归分析CPP的影响因素,构建Logistic回归模型,采用受试者工作特征(ROC)曲线评估其对CPP的预测价值。结果 CPP组激素类食品饮食、营养滋补品饮食比例、LH峰值、血清IGF-1水平高于IPT组,子宫容积、卵巢容积、>4 mm卵泡数大于IPT组,25-(OH)D水平低于IPT组,差异均有统计学意义(P<0.05)。Logistic回归分析显示,激素类食品饮食、营养滋补品饮食、子宫容积、卵巢容积、>4 mm卵泡数、LH峰值、血清IGF-1、25-(OH)D水平是CPP发生的影响因素(P<0.05)。构建的风险预测模型:Logit(P)=-12.857+激素类食品饮食×0.423+营养滋补品饮食×0.530+子宫容积×0.971+卵巢容积×0.885+>4 mm卵泡数×0.772+LH峰值×1.145+IGF-1×0.736-25-(OH)D×0.681。ROC曲线分析显示,Logit(P)>0.5时,模型预测CPP发生的曲线下面积(AUC)为0.859,诊断灵敏度为70.92%,特异度为84.16%。结论 CPP发生与子宫容积、卵巢容积、>4 mm卵泡数、LH峰值、血清IGF-1、25-(OH)D水平以及激素类食品、营养滋补品饮食情况有关,据此构建的Logistic回归模型对CPP的预测价值较高,可为临床决策提供参考。

Objective To construct a prediction model of central precocious puberty(CPP)in girls based on laboratory indicators,ultrasonic parameters and clinical factors.Methods A total of 103 girls with precocious puberty in this hospital from January 2020 to April 2022 were selected as research objects,including 62 cases of simple premature breast development(IPT group)and 41 cases of CPP(CPP group).The clinical baseline data,ultrasound parameters(uterine volume,ovarian volume,number of>4 mm follicles),follicle-stimulating hormone(FSH),luteinizing hormone(LH)peak,serum insulin-like growth factor-1(IGF-1),insulin growth factor binding protein(IGFBP-3)and 25-hydroxyvitamin D[25-(OH)D]were compared between the two groups.The influencing factors of CPP were analyzed by Logistic regression model,and the predictive value of CPP was evaluated by receiver operating characteristic(ROC)curve.Results The proportion of hormone food diet,nutritional supplements diet,LH peak and serum IGF-1 level in CPP group were higher than those in IPT group,the uterine volume,ovarian volume and number of>4 mm follicles were higher than those in IPT group,and the 25-(OH)D level was lower than those in IPT group,with statistical significance(P<0.05).Logistic regression analysis showed that hormone food diet,nutritional supplements diet,uterine volume,ovarian volume,number of>4 mm follicles,LH peak,serum IGF-1,25-(OH)D levels were the influencing factors for CPP(P<0.05).The risk prediction model constructed was as follows:Logit(P)=-12.857+hormone food diet×0.423+nutritional tonic diet×0.530+uterine volume×0.971+ovarian volume×0.885+number of>4 mm follicles×0.772+LH peak×1.145+IGF-1×0.736-25-(OH)D×0.681.ROC curve analysis showed that when Logit(P)>0.5,the AUC of the model predicted the occurrence of CPP was 0.859,the diagnostic sensitivity was 70.92%and the specificity was 84.16%.Conclusion The occurrence of CPP is related to uterine volume,ovarian volume,the number of>4 mm follicles,LH peak,serum IGF-1,25-(OH)D levels,hormone foods and nutritional supplements.The established Logistic regression model has high predictive value for CPP,and may provide reference for clinical decision-making.