Purpose: There is a growing interest in extending the axial fields-of-view (AFOV) of PET scanners. One major limitation for the widespread clinical adoption of such systems is the multifold increase in the associated material costs. In this study, we propose a cost-effective solution to extend the PET AFOV using a sparse detector rings configuration. The corresponding physical performance was validated using Monte Carlo simulations. Methods: Monte Carlo model of the Siemens Biograph TM mCT PET/CT, with a 21.8 cm AFOV and a set of compact rings of LSO crystals was developed as a gold standard. The mCT configuration was then modified by interleaving the LSO crystals in the axial direction within each detector block with 4 mm physical gaps (equivalent to the LSO crystal axial dimension) thus extending the AFOV to 43.6 cm (Ex-mCT). The physical performances of the two MC models were assessed and then compared using NEMA NU 2-2007 standards. Results: Ex-mCT showed <0.2 mm difference in transaxial spatial resolution, and, 0.8 mm and 0.3 mm deterioration in axial spatial resolution, compared to the mCT, at 1 and 10 cm off-center of the transaxial field-of-view respectively. The system sensitivities for the mCT and Ex-mCT models were 9.4 AE 0.2 and 10.75 AE 0.2 cps/kBq respectively. The higher sensitivity of Ex-mCT was due to four additional detector rings required to double the mCT AFOV. PET images of the NEMA Image Quality (IQ) phantom showed no artifacts due to detector rings sparsity, and all spheres were visible in both configurations. Ex-mCT achieved percent contrast recoveries within 5.6% of those of the mCT for all spheres and a maximum of 36% higher background variability at the center of the AFOV. The Ex-mCT, however, showed a more uniform noise distribution over an axial range of almost twice the length of the mCT AFOV. Conclusions: Using the proposed sparse detector-ring configuration, the AFOV of current generation PET systems can be doubled while maintaining the original number and volume of detector crystal elements, and without jeopardizing the system's overall physical performance. Despite an increase in the noise level, the Ex-mCT exhibited an improved noise uniformity.

We report on the NEMA-NU2-2012 performance of a hypothetical Monte Carlo (MC) model, Ex-PET, of the Siemens Biograph Vision positron emission tomography (PET)/CT (Bio-Vis) with sparse detector module rings and extended axial field of view (AFOV). MC simulations were performed with the detector module rings interleaved with 32-mm gaps, equivalent to the axial dimension of each detector module, yielding an AFOV of 48.0 cm (Bio-Vis has 25.6-cm AFOV). 3D-PET acquisition combined with a limited continuous-bed-motion (limited-CBM) was used to compensate for the loss in sensitivity within the gaps' regions. MC simulations of the Bio-Vis were performed for comparison purposes. All MC simulations were performed using GATE MC toolkit. Ex-PET exhibited 0.49, 0.16, and 0.16 mm deterioration in axial resolution at 1, 10, and 20 cm off-center of the transaxial field of view, respectively, compared to Bio-Vis. Only 1% reduction in system sensitivity and 6% reduction in peak NECR was observed with Ex-PET compared to Bio-Vis. 3D-OSEM image reconstruction, combined with CBM, allowed compensating for the lack of counts within the gaps' regions. NEMA Image Quality test showed <6% reduction in contrast recovery with Ex-PET versus Bio-Vis, yet the background variability was increased by up to 8%. The feasibility of PET imaging with an easily adoptable sparse detector configuration was demonstrated. This can lay the pathway for future development of cost-effective PET systems with long and conventional AFOV's. Index Terms-Continuous bed motion (CBM), extended axial field of view, Monte Carlo (MC), positron emission tomography (PET), sparse detectors.