LCPR-Net: low-count PET image reconstruction using the domain transform and cycle-consistent generative adversarial networks

Abstract

Background: Reducing the radiation tracer dose and scanning time during positron emission tomography (PET) imaging can reduce the cost of the tracer, reduce motion artifacts, and increase the efficiency of the scanner. However, the reconstructed images to be noisy. It is very important to reconstruct high-quality images with low-count (LC) data. Therefore, we propose a deep learning method called LCPR-Net, which is used for directly reconstructing full-count (FC) PET images from corresponding LC sinogram data.Methods: Based on the framework of a generative adversarial network (GAN), we enforce a cyclic consistency constraint on the least-squares loss to establish a nonlinear end-to-end mapping process from LC sinograms to FC images. In this process, we merge a convolutional neural network (CNN) and a residual network for feature extraction and image reconstruction. In addition, the domain transform (DT) operation sends a priori information to the cycle-consistent GAN (CycleGAN) network, avoiding the need for a large amount of computational resources to learn this transformation. Results:The main advantages of this method are as follows. First, the network can use LC sinogram data as input to directly reconstruct an FC PET image. The reconstruction speed is faster than that provided by model-based iterative reconstruction. Second, reconstruction based on the CycleGAN framework improves the quality of the reconstructed image.Conclusions: Compared with other state-of-the-art methods, the quantitative and qualitative evaluation results show that the proposed method is accurate and effective for FC PET image reconstruction.

Evaluation of a respiratory motion-corrected image reconstruction algorithm in 2-[18F]FDG and [68Ga]Ga-DOTA-NOC PET/CT: impacts on image quality and tumor quantification

“…Furthermore, we used OSEM reconstruction with an acquisition time of 3 min per bed position in the chest and upper abdomen according to the vendor's default settings. The influence of acquisition time and advanced reconstruction algorithms, such as BPL (28) or deep learning techniques (33,34), on MCIR should be studied in the future.…”

Personalization of 99mTc-sestamibi activity in SPECT/CT myocardial perfusion imaging with the cardiofocal SmartZoom® collimator

“…We should mention here the Artificial Intelligence (AI) algorithms that were recently proposed for de-noising nuclear medicine images [21] and specifically for optimizing the MPI dose [22]. These techniques are based on deep-learning applied to complex neural networks and may be able to significantly reduce the administrated activity without degrading image quality [22].…”

Deep learning-based image reconstruction and post-processing methods in positron emission tomography for low-dose imaging and resolution enhancement

“…Comparisons were made with DeepPET and a Unet for evaluation, with the FBP-Net being more robust to overfitting and previously unseen anatomy. Other works incorporated unfiltered back projections as a domain transform with no sinogram space filtering [ 54 , 55 ]. Zhang et al [ 54 ] presented the bpNET which used an unfiltered back projection as a pre-processing step followed by a residual encoder decoder network trained on synthetic data.…”

“…Zhang et al [ 54 ] presented the bpNET which used an unfiltered back projection as a pre-processing step followed by a residual encoder decoder network trained on synthetic data. Xue et al [ 55 ] also use an unfiltered back projection as pre-processing with a cycle consistent GAN network [ cycleGAN] trained on clinical data. Whiteley et al [ 56 ] presented a network termed Direct-PET which learned an optimal sinogram compression and performed a more efficient domain transform by masking the sinograms and mapping to a patch in image space.…”