Assessing cardiac function from total-variation-regularized 4D C-arm CT in the presence of angular undersampling.

Area of Science:

• Medical Imaging
• Radiology
• Cardiovascular Imaging

Background:

• Time-resolved tomographic cardiac imaging with C-arm devices aids minimally invasive therapy by analyzing heart function in the catheter lab.
• Clinically feasible protocols face reconstruction challenges due to sparse angular sampling.
• Compressed sensing offers image recovery from undersampled data via sparsity-based regularization.

Purpose of the Study:

• To investigate the impact of reduced projection views on 4D cardiac C-arm CT reconstruction quality.
• To determine the minimum data required for accurate cardiac function analysis.
• To evaluate the performance of convex optimization with primal-dual splitting for undersampled cardiac CT.

Main Methods:

• Applied a convex optimization approach using primal-dual splitting to 4D cardiac C-arm CT.
• Examined reconstruction quality degradation with as few as [Formula: see text] projection views per heart phase.
• Determined regularization weights using a numerical phantom study and performed task-based evaluation in eight patients.

Main Results:

• Spatially and temporally total-variation-regularized reconstruction quality was assessed under severe undersampling.
• Numerical phantom studies demonstrated the benefits of individual regularizers.
• Semi-automatic segmentation-based volume measurements of the left ventricular blood pool on undersampled images correlated at 99% with less sparsely sampled data.

Conclusions:

• Accurate left ventricular blood pool volume measurements are achievable with significantly reduced projection data in 4D cardiac C-arm CT.
• This approach supports reduced scan times, lower radiation doses, and decreased contrast agent administration.
• The findings validate the potential of compressed sensing techniques for efficient cardiac imaging in clinical practice.