Acquisition Protocol Optimization of a Dual-Panel PET Scanner Dedicated to Prostate Cancer Imaging

Provision is a dual-panel, prostate-specific PET scanner featuring a compact, cost-effective design with state-of-the-art detectors. The <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\sim 10 \times 10 ~\text{cm}^{2}$</tex> detector panels are mounted on mechanical motors, enabling expanded coverage through multi-step acquisition protocols (APs). Assuming a fixed total acquisition time, AP design balances improved field-of-view (FOV) coverage (more acquisition steps) with statistical quality (shorter frame duration per step). This study aimed to identify the optimal AP to maximize image quality under practical imaging constraints. A fixed detector-to-detector distance of 294 mm was selected based on pelvic thickness analysis of 120 prostate cancer patients. Four APs comprising 3 to 17 steps were designed to progressively enhance angular coverage and inter-detector module gap filling, targeting a fixed FOV of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$10 \times 10 \times 10 ~\text{cm}^{3}$</tex>. A scaled-down NEMA IQ phantom (spheres diameters 3-13 mm) with a 10:1 activity ratio was used to mimic small, high-uptake lesions in <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">18</sup>F-PSMA PET imaging. Experiments were simulated using GATE Monte Carlo toolkit. Images were reconstructed using list-mode Maximum Likelihood Expectation Maximization (MLEM) incorporating a Gaussian shift-invariant Point Spread Function (PSF) kernel. Four images were produced with each AP: short and long scans (10 and 20 min), each with and without PSF. The 20 -minute PSF-enabled scans showed the best performance across all APs. From AP1 to AP4, the average (across all sphere sizes) contrast recovery (CR) was 57%, 55.1%, 54.1%, and 53.7 % (5.9% drop from AP1 to AP4); noise was 12.4%, 11.8%, 9.8%, and 9.2% (25.8% reduction from AP1 to AP4); and CNR was 44.5, 44.6, 53.5, and 55.5 (24.9 % gain from AP1 to AP4). In the 10 -min scan, the gain in CNR was less pronounced (14 % from AP1 to AP4). While increasing AP complexity did not improve CR, it significantly reduced noise, leading to improved CNR and image quality. This improvement was less pronounced in scans with shorter duration.