Application of Radiomics in Vesselness Analysis of CT Angiography Images of Stroke Patients

Area of Science:

• Medical Imaging
• Radiology
• Neurology

Background:

• Vascular collateral development in the lesion area is crucial for treatment selection and outcomes in ischemic stroke (IS) patients.
• CT angiography (CTA) is the primary method for visualizing vessel branching, but its analysis can be enhanced.
• Quantifying collateral development is key to improving patient management and therapeutic effectiveness.

Purpose of the Study:

• To apply radiomics methods to vesselness measurements derived from CTA images for improved analysis of vascular collaterals in IS.
• To develop and implement a novel filter for generating 3D vesselness probability maps from CTA data.
• To investigate differences in vesselness between lesion and contralateral areas in IS patients using radiomics features.

Main Methods:

• Vesselness measurement calculated using Hessian matrix analysis with practical modifications.
• Implementation of a filter generating 3D images indicating the probability of vessel-like structures.
• Radiomics feature extraction from vesselness maps using the PyRadiomics library for 30 IS patients.

Main Results:

• A novel radiomics-based algorithm was developed to quantify vesselness from CTA.
• Significant differences in vesselness distribution were observed between the affected and intact hemispheres in IS patients.
• The study demonstrated the potential of radiomics for characterizing vascular changes in ischemic stroke.

Conclusions:

• Radiomics analysis of CTA-derived vesselness shows promise for evaluating collateral development in ischemic stroke.
• This approach can provide quantitative insights into cerebrovascular changes, potentially aiding clinical decision-making.
• Further research with larger cohorts is warranted to validate these findings and explore clinical applications.