Radiomics Nomogram Model Based on TOF-MRA Images: A New Effective Method for Predicting Microaneurysms

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Intracranial microaneurysms pose diagnostic challenges.
• Accurate preoperative differentiation between true and pseudo microaneurysms is crucial for effective treatment.

Purpose of the Study:

• To develop a radiomics nomogram model using TOF-MRA for preoperative prediction of true microaneurysms.
• To assess the model's diagnostic performance compared to clinical factors and radiologists.

Main Methods:

• 118 patients with intracranial aneurysms were analyzed.
• A radiomics signature was built using LASSO regression on reproducible features.
• A nomogram model integrated clinical factors and the radiomics signature.

Main Results:

• The radiomics model achieved an AUC of 0.875, outperforming the clinical model (AUC=0.75).
• The radiomics nomogram model showed high diagnostic performance (AUC=0.913).
• Decision curve analysis confirmed the superior net benefit of the radiomics nomogram model.

Conclusions:

• Radiomics features from TOF-MRA reliably predict true microaneurysms.
• The developed radiomics nomogram model provides an objective basis for clinical treatment decisions.