Cone-Beam Pair-Wise Data Consistency Conditions in Helical CT

Area of Science:

• Medical Imaging
• Physics
• Computer Science

Background:

• Data consistency conditions (DCC) are crucial for correcting X-ray projection inconsistencies in computed tomography (CT).
• Existing DCC methods often exclude projection pairs that cause singularities, limiting their applicability.
• Helical CT acquisitions with cylindrical detectors are standard in medical diagnostics.

Purpose of the Study:

• To investigate and adapt Data Consistency Conditions (DCC) for helical CT acquisitions using cylindrical detectors.
• To develop a method for handling singularities in DCC calculations caused by intersecting source position lines and the field-of-view (FOV).
• To evaluate the efficacy of DCC in detecting breathing motion and propose improvements.

Main Methods:

• Derived DCC directly in detector coordinates for helical cone-beam projections, avoiding rebinning.
• Accounted for singularities by including projection pairs where the source position line intersects the FOV.
• Identified non-truncated fan-beams to address axial projection truncation.
• Evaluated DCC's motion detection capability on simulated and real data, introducing a noise-aware metric.

Main Results:

• Developed DCC applicable to helical CT with cylindrical detectors, including previously excluded projection pairs.
• Demonstrated that DCC can detect breathing motion when the baseline and FOV do not intersect.
• Showed that motion-induced inconsistency is masked by noise and discretization errors when the baseline and FOV intersect.
• Proposed a noise-aware DCC metric, which, combined with a moving average, effectively detects breathing motion.

Conclusions:

• The adapted DCC are effective for helical CT, enhancing data consistency.
• A novel noise-aware DCC metric significantly improves breathing motion detection, especially in challenging acquisition geometries.
• These findings offer a more robust method for motion artifact correction in CT imaging.