| 郑永胜,杨道明,郑武洪,王均祎.决策树模型在自发性脑出血患者继发癫痫风险预测中的应用.[J].中南医学科学杂志.,2025,(1):58-61. |
| 决策树模型在自发性脑出血患者继发癫痫风险预测中的应用 |
| Application of decision tree model in predicting the risk of secondary epilepsy in patients with spontaneous intracerebral hemorrhage |
| 投稿时间：2024-07-03 修订日期：2024-11-09 |
| DOI：10.15972/j.cnki.43-1509/r.2025.01.013 |
| 中文关键词: 自发性脑出血 继发癫痫 影响因素 决策树模型 风险预测 [ |
| 英文关键词:spontaneous cerebral hemorrhage secondary epilepsy influencing factors decision tree model risk prediction |
| 基金项目:福建省自然科学基金项目(2023J011205) |
|
| 摘要点击次数: 0 |
| 全文下载次数: 0 |
| 中文摘要: |
| 目的探讨自发性脑出血(SICH)患者继发癫痫的影响因素,并建立决策树模型进行风险预测。 方法选取146例SICH患者,并统计患者住院28天内继发癫痫发作情况。记录患者基线资料、影像学资料及实验室指标。采用单因素及多因素Logistic回归分析检验SICH患者继发癫痫发作的影响因素,应用决策树模型卡方自动交互检测算法建立风险预测模型并进行评价。 结果146例SICH患者中继发癫痫发生率为16.44%(24/146)。继发癫痫患者皮层受累占比、血肿体积、美国国立卫生院卒中量表(NIHSS评分)、中线移位、血清同型半胱氨酸(Hcy)水平高于未继发癫痫患者(P＜0.05)。经Logistic回归分析显示,NIHSS评分、皮层受累、血肿体积、中线移位及Hcy水平是影响SICH患者继发癫痫发作的独立影响因素(P＜0.05)。决策树模型显示,皮层受累是SICH患者继发癫痫发作最重要的影响因素；ROC曲线显示,模型预测SICH患者继发癫痫发作的AUC为0.912(P＜0.001),特异度为0.631,灵敏度为0.917,约登指数为0.548。 结论NIHSS评分、皮层受累、血肿体积、中线移位及Hcy水平是SICH患者继发癫痫发作的独立影响因素,根据以上因素建立的决策树模型能有效预测SICH患者继发癫痫风险,并反映各因素的相互关系。 |
| 英文摘要: |
| AimTo explore the influencing factors of secondary seizures in patients with spontaneous intracerebral hemorrhage (SICH), and to establish a decision tree model for risk prediction. Methods146 SICH patients were selected, and the occurrence of secondary epileptic seizures in patients within 28 days of hospitalization were counted. Univariate and multivariate Logistic regression analysis were used to test the influencing factors of secondary seizures in SICH patients. The decision tree model Chi-Square Aautomatic Interaction Detection algorithm was used to establish and evaluate the risk prediction model. ResultsThe incidence of recurrent epilepsy in 146 SICH patients was 16.44% (24/146). The proportion of cortical involvement, the hematoma volume, national institutes of health stroke scale (NIHSS) score, midline shift, and serum homocysteine (Hcy) level in patients with secondary epilepsy were higher than those in patients without secondary epilepsy (P＜0.05). Logistic regression analysis showed that NIHSS score, cortical involvement, hematoma volume, midline shift and Hcy level were independent influencing factors of secondary seizures in SICH patients (P＜0.05). The decision tree model shows that cortical involvement is the most important influencing factor for secondary epileptic seizures in SICH patients. The ROC curve showed that the AUC of the model for predicting secondary seizures in SICH patients was 0.912 (P＜0.001), the specificity was 0.631, the sensitivity was 0.917, and the Youden index was 0.548. ConclusionNIHSS score, cortical involvement, hematoma volume, midline shift and Hcy level are independent influencing factors of secondary epilepsy in SICH patients. The decision tree model established based on the above factors can effectively predict the risk of secondary epilepsy in SICH patients and reflect the relationship between various factors. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|