通过机器学习使用临床和放射变量来预测内动脉瘤破裂状态
Mark D Johnson1, Pradyumna Elavarthi2, Seth Street1
1Department of Neurosurgery, University of Cincinnati, Cincinnati, Ohio, United States.
Surgical neurology international
|August 21, 2025
概括
随机森林模型使用3D形状特征准确预测内动脉瘤破裂. 这些人工智能技术在理解动脉瘤行为和患者风险评估方面提供了有前途的进展.
科学领域:
- 神经外科
- 医学成像
- 人工智能
背景情况:
- 人工智能 (AI) 和机器学习 (ML) 越来越多地用于分析内动脉瘤 (IA) 破裂的临床和放射数据.
- 之前的研究已经探索了各种ML技术,以将患者特征与IA破裂状态相关联.
研究的目的:
- 应用和比较多个ML模型的性能,包括随机森林 (RF),XGBoost (XGB),支持矢量机 (SVM) 和多层感知器 (MLP),用于预测IA破裂状态.
- 确定对IA断裂预测有最重要的临床和放射性特征.
主要方法:
- 分析了178个具有53个特征的IA数据集,删除了高度相关的特征以减少冗余.
- 使用网格搜索进行超参数调整,并通过5次交叉验证对5次代进行评估.
- 计算了性能指标,包括准确性,精度,回忆,F1得分和曲线下的面积 (AUC). 威尔科克森签名等级测试比较了AUC得分.
主要成果:
- 与XGBoost (0.76),SVM (0.69) 和MLP (0.65) 模型相比,随机森林 (RF) 的准确度最高 (85%) 和AUC (0.85) 更高 (P<0.05).
- 在所有模型中, 分形尺寸被认为是最关键的特征.
- 三维 (3D) 形状特征构成了15个最重要的特征中的8个.
结论:
- 射频模型表现出高精度和平衡的精度/回忆力,用于预测IA断裂.
- 三维几何特征是IA破裂状态的关键预测因素,强调它们在临床评估和AI驱动分析中的重要性.
相关概念视频
Aneurysm II: Clinical Manifestations and Diagnostic Studies
22
Thoracic, aortic arch and abdominal aneurysms are significant vascular conditions that can present with various clinical manifestations and lead to serious complications. Understanding these manifestations and the appropriate diagnostic studies is essential for effective management and treatment.Thoracic Aortic AneurysmsThoracic aortic aneurysms often remain asymptomatic until they reach a size that impinges on adjacent structures. They typically cause deep, diffuse chest pain that radiates to...
22
Aneurysm III: Interprofessional Care
24
Aneurysm management involves either conservative medical therapy or surgical intervention, depending on the size and symptoms of the aneurysm. Conservative management is generally reserved for smaller, asymptomatic aneurysms, while larger or symptomatic aneurysms often necessitate surgical repair.Conservative Medical TherapyFor small, asymptomatic aneurysms, particularly abdominal aortic aneurysms (AAA) less than 5.5 centimeters in diameter, conservative medical therapy is recommended. This...
24


