Joint Temporal and Spectral Processing for Improved Digital Subtraction Angiography Using Photon-Counting Detectors.

Area of Science:

• Medical Imaging
• Radiology
• Image Processing

Background:

• Digital subtraction angiography (DSA) is the standard for medical imaging and interventional guidance.
• Spectral imaging in DSA has been hindered by noise from material decomposition.
• Previous spectral DSA techniques faced challenges with noise amplification, limiting clinical use.

Purpose of the Study:

• To introduce a novel joint processing strategy for spectral DSA.
• To leverage both temporal and spectral information for improved material decomposition.
• To address noise amplification issues in spectral imaging for DSA.

Main Methods:

• Developed a model-based material decomposition approach using pre- and post-contrast images.
• Evaluated performance on a phantom with a photon-counting detector.
• Compared joint processing against temporal subtraction and other spectral DSA methods.

Main Results:

• The novel method achieved over an order of magnitude variance reduction compared to prior spectral DSA techniques.
• Demonstrated significant noise and bias reduction in spectral results, enabling three-material decomposition.
• Achieved a mean variance reduction of 23.9% (simulation) and 10.8% (experimental) over temporal subtraction.

Conclusions:

• The proposed joint processing strategy significantly enhances image quality in spectral DSA.
• This method overcomes key limitations of previous spectral imaging approaches.
• Improvements facilitate the clinical translation and adoption of spectral angiography.