Non-Isocentric Geometry for Next-Generation Tomosynthesis With Super-Resolution.

Area of Science:

• Medical Imaging
• Radiological Physics
• Biomedical Engineering

Background:

• Digital breast tomosynthesis (DBT) is an advancement over mammography.
• Traditional DBT systems often use isocentric geometry.
• Novel geometric designs are being explored to improve DBT image quality.

Purpose of the Study:

• To evaluate the impact of a novel non-isocentric geometry (superior-to-inferior detector motion) on DBT image quality.
• To analyze key image quality metrics including aliasing, point spread function (PSF), power spectra, and signal-to-noise ratio (SNR).

Main Methods:

• Theoretical and experimental analysis of aliasing using r-factor, bar patterns, and star patterns.
• Modeling of the point spread function (PSF) to assess out-of-plane blurring.
• Analysis of power spectra in an anthropomorphic phantom.
• Evaluation of signal-to-noise ratio (SNR) for calcification detection using the CIRS Model 020 BR3D Breast Imaging Phantom.

Main Results:

• The non-isocentric geometry demonstrated suppressed aliasing artifacts in reconstructions.
• Reduced point spread function (PSF) width was observed at most positions, indicating less out-of-plane blurring.
• Anatomic noise was reduced, and signal-to-noise ratio (SNR) for calcification detection was improved.
• Despite a potential SNR trade-off with smaller pixel reconstructions, overall SNR was enhanced by the detector-motion acquisition.

Conclusions:

• The investigated non-isocentric geometry offers significant improvements in digital breast tomosynthesis image quality.
• This novel design effectively reduces artifacts and noise while enhancing the detectability of microcalcifications.
• The superior-to-inferior detector motion represents a promising advancement for next-generation DBT systems.