摘要
目的:探寻血液系统恶性肿瘤(hematological malignancies,HMs)化疗后不良事件发生的危险因素,构建其风险预测模型。方法:采用Logistic回归分析筛选危险因素、建立风险预测模型,并用H-L卡方检验、ROC曲线对模型进行检验。结果:肝功能损害、感染、凝血功能异常、化疗强度、急性白血病、肥胖症是HMs化疗后不良事件发生的独立危险因素。H-L卡方检验(χ2=13.236,P=0.067),ROC曲线下面积为0.870,95%CI(0.844,0.895),说明模型具有较好的校准度及鉴别效度;验证组验证该模型的灵敏性、特异性、阳性预测值、阴性预测值、一致率分别是72.7%、87.4%、87.3%、73.0%、80.0%。结论:多因素Logistic回归模型能够较好地预测HMs化疗后不良事件发生的风险情况。
Objective To explore the risk factors of adverse events induced by chemotherapy for hematological malignancies （HMs）, and to establish a clinical risk prediction model. Methods We used univariate and multivariate logistic regression analysis to determine the independent risk factors and established a risk model. The calibration and discrimination of the model were assessed by the H-L test and the area under the ROC curve, respectively. Results Multivariate logistic regression showed that liver dysfunction, infection, coagulation abnormalities, intensity of chemotherapy, acute leukemia, and obesity were independent risk factors for HMs patients with adverse events induced by chemotherapy. The predictive model showed better discrimination （area under the ROC curve = 0.870, 95%CI 0.844 to 0.895）and calibration （H-L statistic X2 = 13.236, P = 0.067）. The sensitivity, specificity, positive predictive value, negative predictive value, and consistency rate in the prediction model group were 72.7%, 87.4%, 87.3%, 73.0%, and 80.0%, respectively. Conclusion Risk prediction model is accurate and sufficient to estimate the adverse events induced by chemotherapy for hematological malignancies.
出处
《实用医学杂志》
CAS
北大核心
2013年第7期1073-1075,共3页
The Journal of Practical Medicine
基金
国家自然科学基金项目(编号:81202358)
江苏省自然科学基金项目(编号:BK2008236)
镇江市社会发展项目(编号:SH2010030
SH2011021)
关键词
血液肿瘤
化疗不良事件
LOGISTIC模型
Hematologic neoplasms
Adverse events induced by chemotherapy
Logistic risk prediction model
