Article Summary
胡姝雯,高飞飞,余 佳,赵碧英,李 佳.生长激素缺乏症早期诊断列线图预测模型的建立与验证[J].现代生物医学进展英文版,2024,(19):3772-3774.
生长激素缺乏症早期诊断列线图预测模型的建立与验证
Establishment and Validation of a Predictive Model for Early Diagnosis of Growth Hormone Deficiency Using a Column Chart
Received:April 22, 2024  Revised:May 16, 2024
DOI:10.13241/j.cnki.pmb.2024.19.045
中文关键词: GHD  诊断  预测模型  列线图
英文关键词: Growth hormone deficiency  Diagnosis  Prediction model  Column chart
基金项目:陕西省科技研发计划项目(2022SF-047)
Author NameAffiliationE-mail
胡姝雯 西安市儿童医院内分泌遗传代谢科 陕西 西安 710003 hongshu0312@163.com 
高飞飞 西安市儿童医院内分泌遗传代谢科 陕西 西安 710003  
余 佳 西安市儿童医院内分泌遗传代谢科 陕西 西安 710003  
赵碧英 西安市儿童医院内分泌遗传代谢科 陕西 西安 710003  
李 佳 西安市儿童医院内分泌遗传代谢科 陕西 西安 710003  
Hits: 12
Download times: 6
中文摘要:
      摘要 目的:探讨生长激素缺乏症(GHD)儿童一般资料及血清学指标的差异,并建立GHD早期诊断列线图预测模型。方法:回顾性分析2021年1月至2023年10月期间因身高低到我院就诊的719例儿童的住院资料,按生长激素激发试验将719例儿童分为GHD组(474例)和非GHD组(245例),检测患者血清IGF-1的水平,运用LASSO回归分析筛选最佳预测因子,运用多因素Logistic回归分析建立列线图预测模型。ROC曲线验证该模式的性能。结果:与GHD组相比,非GHD组儿童的年龄、BMI、ALT、AST、ALKP、IGFBP、FFA与IGF-1数据差异具有统计学意义(P<0.05)。Logistic回归分析结果表明,年龄、ALT、AST、ALKP、IGFBP、FFA、IGF-1为GHD早期的预测因子。构建列线图预测模型内部验证的C-index较高。预测模型的AUC为0.845(95% CI:0.801-0.891),证明该模型具有较好的预测效能和判别能力。结论:年龄、ALT、AST、ALKP、IGFBP、FFA、IGF-1为GHD早期的预测因子,基于上述指标成功建立GHD早期诊断列线图预测模型,内部验证显示模型性能良好,具有一定的临床诊断价值。
英文摘要:
      ABSTRACT Objective: To explore the differences in general information and serological indicators in children with growth hormone deficiency (GHD), and to establish a predictive model for early diagnosis of GHD using a column chart. Methods: A retrospective analysis was conducted on the hospitalization data of 719 children who visited our hospital due to low height between July 2021 and October 2023. The 719 children were divided into GHD group (474 cases) and non GHD group (245 cases) according to the growth hormone stimulation test. The serum IGF-1 levels of the patients were measured, and LASSO regression analysis was used to screen for the best predictive factors. Multiple factor logistic regression analysis was used to establish a column chart prediction model. The ROC curve verifies the performance of this mode. Results: Compared with the GHD group, the age The differences in BMI, ALT, AST, ALKP, IGFBP, FFA, and IGF-1 data were statistically significant(P<0.05). The results of logistic regression analysis indicate that age ALT, AST, ALKP, IGFBP, FFA, and IGF-1 are early predictors of growth hormone deficiency. The C-index for internal validation of the column chart prediction model is relatively high. The AUC of the prediction model is 0.845 (95% CI: 0.801-0.891), which proves that the model has good predictive performance and discriminative ability. Conclusion: Age ALT, AST, ALKP, IGFBP, FFA, and IGF-1 are predictive factors for early growth hormone deficiency. Based on the above indicators, a column chart prediction model for early diagnosis of growth hormone deficiency has been successfully established. Internal validation shows that the model has good performance and certain clinical diagnostic value.
View Full Text   View/Add Comment  Download reader
Close