Clinical-Radiomics Nomogram Model Based on CT Angiography for Prediction of Intracranial Aneurysm Rupture: A...

Area of Science:

• Neurosurgery
• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Intracranial aneurysm rupture risk estimation is crucial for treatment decisions.
• Multicenter studies on clinical-radiomics models for aneurysm rupture prediction are scarce.

Purpose of the Study:

• To develop a clinical-radiomics model for predicting intracranial aneurysm rupture.
• To evaluate the added value of this model in discriminating aneurysm rupture.

Main Methods:

• Retrospective enrollment of 516 aneurysms (273 ruptured) from four hospitals.
• Construction of clinical, radiomics, and clinical-radiomics models using logistic regression.
• Performance evaluation via ROC curves, decision curves, calibration curves, and DeLong tests.

Main Results:

• The clinical-radiomics model achieved AUCs of 0.880 (training), 0.807 (internal), and 0.815 (external) datasets.
• This model demonstrated superior diagnostic performance compared to clinical-only and radiomics-only models (P < 0.05).
• Three clinical factors and fourteen radiomic features were identified as significant predictors.

Conclusions:

• The developed clinical-radiomics model shows significant promise for predicting intracranial aneurysm rupture.
• This integrated approach offers improved accuracy over existing methods.
• A comprehensive nomogram was developed for intuitive clinical application.