Robust low-dose dynamic cerebral perfusion CT image restoration via coupled dictionary learning scheme

Area of Science:

• Medical Imaging
• Radiology
• Computational Imaging

Background:

• Dynamic cerebral perfusion CT (PCT) is crucial for diagnosing cerebrovascular diseases but involves significant radiation dose.
• Reducing radiation exposure in PCT is essential for patient safety and frequent monitoring.

Purpose of the Study:

• To develop and validate a novel image restoration method for low-dose dynamic PCT.
• To enable accurate hemodynamic assessment with reduced radiation exposure.

Main Methods:

• A coupled dictionary learning (CDL) scheme was proposed, integrating 2D background and 3D enhancement information.
• Dictionaries were trained using low-dose unenhanced CT and normal-dose PCT data.
• Restoration of low-dose PCT images was achieved using a modified dictionary learning algorithm.

Main Results:

• The CDL method successfully restored clinically acceptable low-dose PCT images.
• Enhanced kinetic details and diagnostic hemodynamic parameter maps were achieved.
• The proposed method outperformed state-of-the-art techniques in accuracy.

Conclusions:

• Coupled dictionary learning offers a promising solution for low-dose dynamic PCT imaging.
• This approach can improve diagnostic capabilities for stroke and cerebrovascular diseases.
• The method effectively balances image quality, diagnostic accuracy, and radiation safety.