|
| 关于自发性脑出血血肿扩大预测模型的构建及验证 |
| Construction and Verification of Prediction Model for Hematoma Expansion in Spontaneous Intracerebral Hemorrhage |
| 投稿时间：2024-04-16 修订日期：2024-04-16 |
| DOI： |
| 中文关键词: 自发性脑出血 血肿扩大 GCS 列线图 |
| 英文关键词: Spontaneous Intracerebral Hemorrhage Hematoma Expansion GCS Nomogram Model |
| 基金项目: |
|
| 摘要点击次数: 50 |
| 全文下载次数: 0 |
| 中文摘要: |
| 目的：探究自发性脑出血(Spontaneous Intracerebral hemorrhage,SICH)血肿扩大(Hematoma expansion,HE)独立危险因素构建预测模型并进行验证。方法：选择十堰市人民医院2021年1月至2023年1月神经重症收治的自发性脑出血患者316例为训练组,收集十堰市太和医院100例为外部验证组,以Bootstrap法抽取训练组数据为内部验证组,在发病6h内进行影像学及血液检查,发病24h内完善头颅CT复查,按照是否出现血肿扩大进行分组,以Mann-Whitney U检验、卡方检验进行两队列单因素比较分析,并结合Lasso回归分析,将筛选出的因素纳入多因素Logistic回归方程,并构建列线图(Nomogram model),以受试者工作特征曲线(Receiver operating characteristic curve,ROC)评估预测价值,行内部验证(Bootstrap法重复抽样1000次),外部验证,同时绘制校准曲线及临床决策曲线评价模型。结果：单因素分析结合Lasso回归得出性别、糖尿病、入院时收缩压、舒张压、入院格拉斯哥昏迷评分量表(Glasgow coma scale,GCS)、D-二聚体(D-Dimer,DD)、淋巴细胞、NHISS评分、出血部位、发病至基线CT时间、出血部位、血肿形状12个变量因素,将其纳入多因素Logistic回归方程,得出：性别、糖尿病、舒张压、入院GCS评分、发病至基线CT时间、出血部位是自发性脑出血血肿扩大的独立危险因素；ROC曲线中,联合预测HE价值较高,该预测模型的曲线下面积AUC为0.896(95%CI：0.860～0.933),敏感度和特异度分别是76.3%,86.0%；内部验证(Bootstrap法)提示该模型C-index为0.902,外部验证为0.915；内部与外部验证校准曲线P>0.05,拟合度基本吻合；内部验证决策曲线分析提示当阈值概率在3%～78%时,外部验证阈值概率范围大于3%,列线图模型具有较好的临床适用性。结论：性别、糖尿病、舒张压、发病至基线CT时间、入院GCS评分、出血部位是自发性脑出血血肿扩大的独立危险因素；基于此构建的风险预测模型可有效预测血肿扩大风险。 |
| 英文摘要: |
| Objective：To explore the independent risk factors and to predict and verify model of hematoma expansion (HE) after admission in patients with spontaneous intracerebral hemorrhage (SICH). Methods：A total of 316 patients with spontaneous cerebral hemorrhage treated in Shiyan people"s Hospital from January 2021 to January 2023 were selected as the training group, taking samples as internal verification group by Bootstrap method and 100 patients from Taihe hospitals in Shiyan City were collected as the external verification group. Imaging and blood examination were performed within 6 hours after the onset of the disease. The head CT was reexamined within 24 hours after the onset. The patients were divided into groups according to whether the hematoma was enlarged or not. Mann-Whitney U test and chi-square test were used for univariate analysis. Combined with Lasso regression analysis, the selected factors were included in the multi-factor Logistic regression equation, and the line diagram was constructed. The predictive value was evaluated by the receiver working characteristic curve, internal verification (Bootstrap repeated sampling 1000 times), external verification, calibration curve and clinical decision curve evaluation model at the same time. Results：Univariate analysis combined with Lasso regression showed gender, diabetes, systolic blood pressure on admission, diastolic blood pressure, admission Glasgow coma scale, D-dimer, lymphocytes, NHISS score, bleeding site, time from onset to baseline CT, bleeding site, hematoma shape 12 variables were included in the multivariate Logistic regression equation. It is concluded that sex, diabetes, diastolic blood pressure, GCS score on admission, time from onset to baseline CT and location of hemorrhage are independent risk factors for hematoma enlargement in spontaneous intracerebral hemorrhage. In the ROC curve,the value of joint prediction HE is higher. The area under the curve of the prediction model AUC is 0.896 and 95% confidence interval is 0.860～0.933,and the sensitivity and specificity are 76.3% and 86.0% respectively. The calibration curve of internal and external verification is P > 0.05, and the degree of fit is basically the same; the analysis of the decision curve of internal verification shows that when the threshold probability is 3% to 78%, the probability range of external verification threshold is greater than 3%,The line chart model has good clinical applicability. Conclusion：gender、diabetes、DBP、time from onset to baseline CT、admission GCS score and hematoma location are independent risk factors for hematoma expansion in patients with spontaneous intracerebral hemorrhage,and the risk prediction model based on this can effectively predict the risk of hematoma expansion. |
|
View Fulltext
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|