VCC-DSA: A novel vascular consistency constrained DSA imaging model for motion artifact suppression

Area of Science:

• Medical Imaging
• Biomedical Engineering

Background:

• Digital Subtraction Angiography (DSA) is crucial for diagnosing cerebrovascular diseases but suffers from motion artifacts.
• Existing methods like registration-based compensation and learning-based synthesis have limitations in accuracy and data requirements.

Purpose of the Study:

• To develop a novel Vascular Consistency Constrained DSA Imaging Model (VCC-DSA) for robust motion suppression and precise vascular imaging.
• To overcome the limitations of current DSA techniques, particularly concerning motion artifacts and image quality.

Main Methods:

• Introduced a Learning-based Subtraction Mapping Paradigm to enhance algorithmic stability.
• Utilized Residual Dense Blocks and detail-shortcuts for improved performance in complex anatomical structures.
• Developed an innovative Vascular Consistency Strategy to extract intrinsic consistency and suppress motion artifacts.
• Implemented a Mixup-based Data Self-evolution Strategy for dynamic data enhancement during training.

Main Results:

• The VCC-DSA model demonstrated robust motion suppression and precise vascular imaging.
• Achieved significant improvements in Peak Signal-to-Noise Ratio (PSNR) by 73.4% and Structural Similarity Index Measure (SSIM) by 8.56% compared to other methods.
• Validated through human clinical data and a general anesthesia animal experiment.

Conclusions:

• The VCC-DSA model offers a superior solution for motion artifact reduction in DSA.
• This technique enhances the visibility of blood vessels, leading to more accurate diagnoses of cerebrovascular diseases.
• The proposed strategies address key challenges in DSA, paving the way for improved clinical applications.