Anatomical background and generalized detectability in tomosynthesis and cone-beam CT

Area of Science:

• Medical Imaging Physics
• Radiological Sciences
• Systems Analysis

Background:

• Anatomical background noise significantly hinders image detectability in 2D radiography, tomosynthesis, and cone-beam CT (CBCT).
• Understanding and quantifying this anatomical noise is crucial for optimizing imaging system performance and diagnostic accuracy.

Purpose of the Study:

• To develop and apply a generalized cascaded systems analysis to quantify anatomical noise in 2D and 3D imaging.
• To derive generalized metrics of noise-equivalent quanta (NEQ) and detectability index as a function of source-detector orbital extent.

Main Methods:

• A fractal self-similar phantom was used to generate power-law spectral density mimicking anatomical sites.
• Background power spectra were computed for varying source-detector orbital extents (tomosynthesis and CBCT) under different acquisition schemes.
• Generalized NEQ and detectability index were calculated using 3D cascaded systems analysis.

Main Results:

• The phantom successfully replicated power-law spectra, demonstrating the relationship between clutter magnitude/correlation and tomosynthesis angle.
• The generalized NEQ framework effectively analyzed tradeoffs between anatomical, quantum, and electronic noise, dose, and orbital extent.
• Optimal system design for breast and chest tomosynthesis requires application-specific selection of orbital extent, projection number, and dose based on imaging tasks.

Conclusions:

• Generalized cascaded systems analysis provides a robust framework for understanding and optimizing imaging systems.
• This approach effectively describes complex tradeoffs among various noise sources in projection imaging, tomosynthesis, and CBCT.
• The findings offer valuable insights for designing and improving medical imaging systems, particularly for tomosynthesis and CBCT applications.