The objective of this study is to develop a convolutional neural network (CNN) for computed tomography (CT) image super-resolution. The network learns an end-to-end mapping between low (thick-slice thickness) and high (thin-slice thickness) resolution images using the modified U-Net. To verify the proposed method, we train and test the CNN using axially averaged data of existing thin-slice CT images as input and their middle slice as the label. Fifty-two CT studies are used as the CNN training set, and 13 CT studies are used as the test set. We perform five-fold cross-validation to confirm the performance consistency. Because all input and output images are used in two-dimensional slice format, the total number of slices for training the CNN is 7670. We assess the performance of the proposed method with respect to the resolution and contrast, as well as the noise properties. The CNN generates output images that are virtually equivalent to the ground truth. The most remarkable image-recovery improvement by the CNN is deblurring of boundaries of bone structures and air cavities. The CNN output yields an approximately 10% higher peak signal-to-noise ratio and lower normalized root mean square error than the input (thicker slices). The CNN output noise level is lower than the ground truth and equivalent to the iterative image reconstruction result. The proposed deep learning method is useful for both super-resolution and de-noising.

The most investigated semiconductor photosensor for MRI-compatible PET detectors is the avalanche photodiode (APD). However, the silicon photomultiplier (SiPM), also called the Geiger-mode APD, is gaining attention in the development of the next generation of PET/MRI systems because the SiPM has much better performance than the APD. We have developed an MRI-compatible PET system based on multichannel SiPM arrays to allow simultaneous PET/MRI. Methods: The SiPM PET scanner consists of 12 detector modules with a ring diameter of 13.6 cm and an axial extent of 3.2 cm. In each detector module, 4 multichannel SiPM arrays (with 4 × 4 channels arranged in a 2 × 2 array to yield 8 × 8 channels) were coupled with 20 × 18 Lu_1.9Gd_0.1SiO₅:Ce crystals (each crystal is 1.5 × 1.5 × 7 mm) and mounted on a charge division network for multiplexing 64 signals into 4 position signals. Each detector module was enclosed in a shielding box to reduce interference between the PET and MRI scanners, and the temperature inside the box was monitored for correction of the temperature-dependent gain variation of the SiPM. The PET detector signal was routed to the outside of the MRI room and processed with a field programmable gate array–based data acquisition system. MRI compatibility tests and simultaneous PET/MRI acquisitions were performed inside a 3-T clinical MRI system with 4-cm loop receiver coils that were built into the SiPM PET scanner. Interference between the imaging systems was investigated, and phantom and mouse experiments were performed. Results: No radiofrequency interference on the PET signal or degradation in the energy spectrum and flood map was shown during simultaneous PET/MRI. The quality of the MRI scans acquired with and without the operating PET showed only slight degradation. The results of phantom and mouse experiments confirmed the feasibility of this system for simultaneous PET/MRI. Conclusion: Simultaneous PET/MRI was possible with a multichannel SiPM-based PET scanner, with no radiofrequency interference on PET signals or images and only slight degradation of the MRI scans.

Simultaneous reconstruction of activity and attenuation using the maximum likelihood reconstruction of activity and attenuation (MLAA) augmented by time-of-flight (TOF) information is a promising method for positron emission tomography (PET) attenuation correction. However, it still suffers from several problems, including crosstalk artifacts, slow convergence speed, and noisy attenuation maps (μ-maps). In this work, we developed deep convolutional neural networks (CNNs) to overcome these MLAA limitations, and we verified their feasibility using a clinical brain PET data set. Methods: We applied the proposed method to one of the most challenging PET cases for simultaneous image reconstruction (¹⁸F-FP-CIT PET scans with highly specific binding to striatum of the brain). Three different CNN architectures (convolutional autoencoder (CAE), U-net, hybrid of CAE and U-net) were designed and trained to learn x-ray computed tomography (CT) derived μ-map (μ-CT) from the MLAA-generated activity distribution and μ-map (μ-MLAA). PET/CT data of 40 patients with suspected Parkinson’s disease were employed for five-fold cross-validation. For the training of CNNs, 800,000 transverse PET slices and CTs augmented from 32 patient data sets were used. The similarity to μ-CT of the CNN-generated μ-maps (μ-CAE, μ-Unet, and μ-Hybrid) and μ-MLAA was compared using Dice similarity coefficients. In addition, we compared the activity concentration of specific (striatum) and non-specific binding regions (cerebellum and occipital cortex) and the binding ratios in the striatum in the PET activity images reconstructed using those μ-maps. Results: The CNNs generated less noisy and more uniform μ-maps than original μ-MLAA. Moreover, the air cavities and bones were better resolved in the proposed CNN outputs. In addition, the proposed deep learning approach was useful for mitigating the crosstalk problem in the MLAA reconstruction. The hybrid network of CAE and U-net yielded the most similar μ-maps to μ-CT (Dice similarity coefficient in the whole head = 0.79 in the bone and 0.72 in air cavities), resulting in only approximately 5% errors in activity and biding ratio quantification. Conclusion: The proposed deep learning approach is promising for accurate attenuation correction of activity distribution in TOF PET systems.

We propose a new deep learning based approach to provide more accurate whole-body PET/MRI attenuation correction in comparison to the Dixon-based four-segment method. We use activity and attenuation maps estimated using the maximum likelihood reconstruction of activity and attenuation (MLAA) algorithm as inputs to convolutional neural network (CNN) to learn a CT-derived attenuation map. Methods: The whole-body ¹⁸F-FDG PET/CT scan data of 100 cancer patients (38 males and 62 females, age = 57.3 ± 14.1 years) was retrospectively used for training and testing the CNN. A modified U-net was trained to predict a CT-derived μ-map (μ-CT) from the MLAA-generated activity distribution (λ-MLAA) and μ-map (μ-MLAA). One point three million patches derived from sixty patients’ data were used for training the CNN, data of 20 others were used as a validation set to prevent overfitting, and the data of the other 20 were used as a test set for the CNN performance analysis. The attenuation maps generated using the proposed method (μ-CNN), μ-MLAA, and four-segment method (μ-segment) were compared with the μ-CT, a ground truth. We also compared the voxel-wise correlation between the activity images reconstructed using OSEM with the μ-maps, and the standard uptake values of primary and metastatic bone lesions obtained by drawing regions of interest on the activity images. Results: The CNN generates less noisy attenuation maps and achieves better bone identification than MLAA. The average Dice similarity coefficient for bone regions between μ-CNN and μ-CT was 0.77, which was significantly higher than that between μ-MLAA and μ-CT (0.36). Also, the CNN result showed the best pixel-by-pixel correlation with the CT-based results and remarkably reduced differences of activity maps in comparison to CT-based attenuation correction. Conclusion: The proposed deep neural network produced a more reliable attenuation map for 511 keV photons than the four-segment method currently used in whole-body PET/MRI studies.

Personalized dosimetry with high accuracy is crucial owing to the growing interests in personalized medicine. The direct Monte Carlo simulation is considered as a state-of-art voxel-based dosimetry technique; however, it incurs an excessive computational cost and time. To overcome the limitations of the direct Monte Carlo approach, we propose using a deep convolutional neural network (CNN) for the voxel dose prediction. PET and CT image patches were used as inputs for the CNN with the given ground truth from direct Monte Carlo. The predicted voxel dose rate maps from the CNN were compared with the ground truth and dose rate maps generated voxel S-value (VSV) kernel convolution method, which is one of the common voxel-based dosimetry techniques. The CNN-based dose rate map agreed well with the ground truth with voxel dose rate errors of 2.54% ± 2.09%. The VSV kernel approach showed a voxel error of 9.97% ± 1.79%. In the whole-body dosimetry study, the average organ absorbed dose errors were 1.07%, 9.43%, and 34.22% for the CNN, VSV, and OLINDA/EXM dosimetry software, respectively. The proposed CNN-based dosimetry method showed improvements compared to the conventional dosimetry approaches and showed results comparable with that of the direct Monte Carlo simulation with significantly lower calculation time.