Simulation study on sensitivity and count rate characteristics of “OpenPET” geometries

An “OpenPET” geometry which consists of two axially separated detector rings, has been pro LORs posed to image the 360-degree open gap. The open gap is less stressful for patients and suitable for simultaneous PET/CT, entire whole-body PET and in-beam PET. It has been shown that OpenPET geometries can maintain high spatial resolution using a depth-of-interaction (DOI) detector. However, when the PET scanner is used for an extended axial field-of-view (FOV), several characteristics of oblique lines-of-responses (), such as decreased sensitivity due to smaller acceptance angle and increased scatter fraction (SF) and attenuation, should be carefully discussed. It is expected that these characteristics are dependent on the gap size and the target object size. In this work, we study sensitivity and count rate characteristics for three OpenPET geometries using GATE simulation. The dual scanner has 15-cm open gap and 45-cm axial FOV. The quad scanners have 30-cm open gap and 90-cm axial FOV. When the gap is extended, the axial FOV is extended at the cost of decreased sensitivity of the skirts of the main peak. Maximum sensitivity of the dual scanner with the open gap is almost the same as the scanner without the open gap. In the quad scanners, however, maximum sensitivity decreases by 25 % due to the decreased acceptance angles. Noise equivalent count rate (NECR) curves depend on not only the maximum sensitivity but also the target object size, because the extended gap results in a broadened sensitivity profile. For the 100-cm cylinder phantom with 100-cm line source, the loss of the peak NECR is slight for both scanners. For the 100-cm cylinder phantom with local activity, however, the NECR of the quad scanners is maintained at a value approximately half that of the NECR without the open gap.