In modern large-scale PET systems, transferring the digitized raw detector data at high count rates to a centralized processing unit is a challenge. Processing data on FPGAs close to the detectors can reduce data early on and improve scalability of the PET system. We present and evaluate an FPGA implementation of gradient tree boosting (GTB) for one-dimensional position estimation of gamma interactions in the scintillator. GTB is a supervised machine learning algorithm based on building ensembles of binary decision trees. Models were trained offline and inferred in an FPGA (XC7K410T-2FFG676 Kintex-7). Input features and GTB parameters influencing both positioning performance and model size were varied while evaluating the inferred models concerning data throughput and FPGA resource consumption as well as positioning performance. We achieved throughputs per detector between 2.94 × 10 6 and 4.55 × 10 6 gamma interactions per second. For an optimized GTB model, resource consumption could be reduced by factors of 17 and 10 to less than 1 % (2.51 × 10 3 look-up tables) of available logic and 1.26 % (20 BRAMs) of memory resources, while maintaining a positioning performance of 98.63 % when compared to the model with the best positioning performance. The presented framework can be easily adapted to other photosensors and scintillator configurations.
Background and purpose
The restricted bore diameter of current simultaneous positron emission tomography/magnetic resonance imaging (PET/MRI) systems can be an impediment to achieving similar patient positioning during PET/MRI planning and radiotherapy. Our goal was to evaluate the B
1
transmit (B
1
+
) uniformity, B
1
+
efficiency, and specific absorption rate (SAR) of a novel radiofrequency (RF) body coil design, in which RF shielded PET detectors were integrated with the specific aim of enabling a wide-bore PET/MRI system.
Materials and methods
We designed and constructed a wide-bore PET/MRI RF body coil to be integrated with a clinical MRI system. To increase its inner bore diameter, the PET detectors were positioned between the conductors and the RF shield of the RF body coil. Simulations and experiments with phantoms and human volunteers were performed to compare the B
1
+
uniformity, B
1
+
efficiency, and SAR between our design and the clinical body coil.
Results
In the simulations, our design achieved nearly the same B
1
+
field uniformity as the clinical body coil and an almost identical SAR distribution. The uniformity findings were confirmed by the physical experiments. The B
1
+
efficiency was 38% lower compared to the clinical body coil.
Conclusions
To achieve wide-bore PET/MRI, it is possible to integrate shielding for PET detectors between the body coil conductors and the RF shield without compromising MRI performance. Reduced B
1
+
efficiency may be compensated by adding a second RF amplifier. This finding may facilitate the application of simultaneous whole-body PET/MRI in radiotherapy planning.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.