Wang Zihang, Yan Chang, Yuan Wenqing, Jiang Shuangyan, Jiang Yongxiang, Chen Ting
Department of Radiology, The Second Affiliated Hospital of Chongqing Medical University, Chongqing, China.
The Second Clinical College, Chongqing Medical University, Chongqing, China.
Front Neurol. 2025 Jan 7;15:1507082. doi: 10.3389/fneur.2024.1507082. eCollection 2024.
This study aimed to develop and validate a multivariate logistic regression model for predicting intracranial aneurysm (IA) rupture by integrating clinical data, aneurysm morphology, and parent artery characteristics using high-resolution vessel wall imaging (HR-VWI).
A retrospective analysis was conducted on 298 patients with 386 aneurysms. Patients were randomly divided into training ( = 308) and validation ( = 78) sets. Key predictors, including aneurysm size, shape, aneurysm wall and parent artery wall enhancement, were identified through univariate analysis and then used to build the prediction model using multivariate logistic regression. The model was visualized as a nomogram and compared to PHASES and ELAPSS scores.
The logistic regression model demonstrated superior predictive performance with an area under the curve of 0.814, which was significantly higher than PHASES and ELAPSS scores ( < 0.05). The model revealed strong calibration and good agreement between predicted and observed rupture probabilities.
The multivariate model based on HR-VWI, which incorporates aneurysm and parent artery features, provides a more accurate prediction of IA rupture risk than conventional scoring systems, offering a valuable tool for clinical decision-making.
本研究旨在通过整合临床数据、动脉瘤形态和使用高分辨率血管壁成像（HR-VWI）的母动脉特征，开发并验证一种用于预测颅内动脉瘤（IA）破裂的多因素逻辑回归模型。
对298例患有386个动脉瘤的患者进行回顾性分析。患者被随机分为训练集（n = 308）和验证集（n = 78）。通过单因素分析确定关键预测因素，包括动脉瘤大小、形状、动脉瘤壁和母动脉壁强化，然后使用多因素逻辑回归建立预测模型。该模型以列线图形式呈现，并与PHASES和ELAPSS评分进行比较。
逻辑回归模型显示出卓越的预测性能，曲线下面积为0.814，显著高于PHASES和ELAPSS评分（P < 0.05）。该模型显示出很强的校准度，预测破裂概率与观察到的破裂概率之间具有良好的一致性。
基于HR-VWI的多因素模型结合了动脉瘤和母动脉特征，比传统评分系统能更准确地预测IA破裂风险，为临床决策提供了有价值的工具。
