Effects of out of field-of-view activity on imaging performance in a 1mm3 resolution clinical PET system

Abstract

We are building a dual panel PET system dedicated to breast imaging. The edge-on layout of the crystal arrays provides 1mm resolution in all 3 dimensions, while the dense packing of crystal elements provides high sensitivity. The dual panel breast-specific geometry, and the close proximity to the chest wall of our system, makes it susceptible to out-of-field-of-view (OFOV) activity from body uptake, particularly from highly vascular organs like heart and liver. OFOV increases dead time and results in increased random coincidence rates. Good coincidence timing resolution is crucial to mitigate the effects of randoms on the final image. The degradation due to OFOV activity is investigated by imaging a micro-Derenzo phantom in the system with and without OFOV activity from a hot source, which was designed to mimic uptake in the nearby heart. The system timing resolution is measured for both scans, and the spatial resolution and contrast-to-noise ratio (CNR) from the resulting MLEM-reconstructed images are quantified and compared. Randoms rate is estimated using the delayed coincidence window method. In the presence of OFOV activity, the 1.2mm diameter rods are resolvable, the CNR degrades by roughly 4.37%, the system timing resolution degrades by about 9.65%, and the randoms fraction increases by 75%. The results show that our system is able to preserve good spatial resolution, CNR, and timing resolution in the presence of OFOV activity from a hot source mimicking the heart.