Concurrent iterative reconstruction algorithms (CIRAs) in multi-headed SPECT systems

We are developing concurrent iterative reconstruction algorithms (CIRAs) that use partial sets of projection data to perform iterative reconstruction concurrently with data acquisition. In this study, we investigate CIRAs for multi-headed SPECT systems by simulating data for clinically relevant configurations. We study CIRAs for multi-headed SPECT under three different clinically feasible instances and identify the camera geometries and data acquisition protocols that are favorable for the convergence rates of CIRAs. We demonstrate that CIRA converges faster if projections from widely differing angular views, rather than from consecutive angular views, are available early during the data acquisition. Specifically, we show that CIRAs with single-headed and triple-headed systems show similar convergence rates, and are more efficient than CIRAs with the opposed dual-headed systems. In all cases, CIRAs take less total time to acquire and reconstruct images than conventional reconstruction methods with single-headed or multi-headed SPECT systems.