诊断学理论与实践

诊断学理论与实践 ›› 2023, Vol. 22 ›› Issue (03): 277-282.doi: 10.16150/j.1671-2870.2023.03.11

• 论著 • 上一篇    下一篇

帕金森病早期诊断诺谟图模型的建立及验证

魏坚1, 孙杰2(), 崔诗爽3,4()   

  1. 1.上海交通大学医学院附属瑞金医院 检验科，上海 200025
    2.山东省聊城市聊城退役军人医院神经内二科，山东 聊城 252000
    3.上海交通大学医学院附属瑞金医院 老年病科，上海 200025
    4.上海交通大学医学院附属瑞金医院 神经内科，上海 200025
  • 收稿日期:2023-05-05 出版日期:2023-06-25 发布日期:2023-11-17
  • 通讯作者: 孙杰 E-Mail: 1239959965@qq.com;崔诗爽 E-Mail：cuis2007@163.com

Development of a Nomogram model for early diagnosis of Parkinson disease

WEI Jian1, SUN Jie2(), CUI Shishuang3,4()   

  1. 1. Department of Clinical Laboratory，Ruijin Hospital，Shanghai Jiao Tong University School of Medicine 200025
    2. Neurology Department Ⅱ, Liaocheng Veterans Hospital 252000
    3. Geriatrics Department, Ruijin Hospital，Shanghai Jiao Tong University School of Medicine 200025
    4. Neurology Department,Ruijin Hospital，Shanghai Jiao Tong University School of Medic 200025
  • Received:2023-05-05 Online:2023-06-25 Published:2023-11-17

RichHTML

6

PDF (PC)

208

摘要：

目的：构建帕金森病（Parkinson’s disease, PD）早期诊断诺谟图模型并验证。方法：本研究连续纳入2013年6月至2019年12月期间在上海交通大学附属瑞金医院住院的PD患者201例（PD组），以同期住院的201例神经系统慢性疾病患者非原发性帕金森病作为对照（非PD组）。在402例研究对象中，分层随机抽取300例（PD患者和非PD患者各150例）作为训练集，其余102例（PD患者51例，非PD患者51例）作为验证集。本研究采用具有完整数据处理、计算和制图能力的R软件(4.2.1版本)对训练集进行数据分析，采用单因素回归分析筛选PD的危险因素，并进一步进行多因素logistic回归分析及列线图模型构建。利用校准曲线受试者操作特征曲线分别对训练集和验证集进行内部及外部验证。结果：多因素logistic回归分析显示，高龄（>60岁）（OR=3.987，95%CI为2.126~7.477，P=0.131）、认知功能障碍蒙特利尔认知评估量表(Montreal Cognitive Assessment,MoCA)评分>26）（OR=3.094，95%CI为1.654~5.787，P<0.001）、便秘(OR=2.630，95%CI为1.430~4.835，P=0.002)、快速眼动睡眠行为障碍（rapid-eye-movement sleep behavior disorder, RBD）(OR=2.710，95%CI为1.449~5.068，P=0.002)、嗅觉减退（OR=2.117，95%CI为1.172~3.824，P=0.013）、铜兰蛋白（<20 mg/L）（ceruloplasmin，CER）水平降低（OR=3.356，95%CI为1.923~5.855，P<0.001）是PD的危险因素。基于上述危险因素建立诺谟图模型，内外部验证曲线下面积分别为0.729、0.714。结论：诺谟图诊断模型能有效辅助诊断PD，具有一定的临床应用价值。

关键词: 帕金森病, 铜蓝蛋白, 诊断模型, 诺谟图

Abstract:

Objective: To construct a Nomogram model for early diagnosis of Parkinson’s and validate it. Methods: This study consecutively enrolled 201 Parkinson’s patients who were hospitalized in Ruijin Hospital affiliated to Shanghai Jiaotong University between June 2013 and December 2019 as the Parkinson’s group; 201 Patients with chronic neurological diseases of non-primary Parkinson’s disease who were hospitalized during the same period served as the control group. Of the 402 cases, 300 cases (150 Parkinson’s patients and 150 non-Parkinson’s patients) were stratified and randomly selected as the training set. The remaining 102 cases (51 Parkinson’s patients and 51 non-Parkinson’s patients) were used as the validation set. In this study, R software (version 4.2.1) with complete data processing, computation, and graphing capabilities was used to analyze the data in the training set. Univariate logistic regression analysis was used to screen the risk factors for Parkinson’s disease, and multivariate logistic regression analysis and nomogram model construction were further performed. Calibration and ROC curves were used to perform internal and external validation on the training and validation sets respectively. Results: Multivariate Logistic regression analysis showed advanced age（>60） (OR=3.987 95%CI=2.126-7.477 P=0.131), cognitive dysfunction（MoCA score >26） (OR=3.094 95%CI=1.654-5.787 P<0.001), constipation (OR=2.630 95%CI=1.430-4.835 P=0.002), RBD (OR=2.710 95%CI=1.449-5.068 P=0.002), hyposmia (OR=2.117 95%CI=1.172-3.824 P=0.013), and reduced CER（<20 mg/L） (OR=3.356 95%CI=1.923-5.855 P<0.001) were risk factors for Parkinson’s. The Nomogram model was established based on the above risk factors, and the areas under the internal and external validation curves were 0.729 and 0.714, respectively. Conclusions: The Nomogram diagnostic model can effectively assist in diagnosing Parkinson’s disease and has certain clinical application value.

Key words: Parkinson’s disease, Copper blue protein, Diagnostic model, Nomogram.

中图分类号: