Positron emission tomography (PET) is an advanced medical imaging technique widely used in various clinical applications, such as tumor detection and neurologic disorders. Reducing the radiotracer dose is desirable in PET imaging because it decreases the patient’s radiation exposure. However, reducing the dose can also increase noise, affecting the image quality. Therefore, an advanced image reconstruction algorithm based on low-dose PET data is needed to improve the quality of the reconstructed image. In this study, we propose the use of a direct PET image reconstruction network (DPIR-Net) using an improved Wasserstein generative adversarial network (WGAN) framework to enhance image quality. This study provides two main findings. First, our network uses sinogram data as input and outputs high-quality PET images direct, resulting in shorter reconstruction times compared with traditional model-based reconstruction networks. Second, we combine perceptual loss, mean square error, and the Wasserstein distance as the loss function, which effectively solves the problem of excessive smoothness and loss of detailed information in traditional network image reconstruction. We performed a comparative study using maximum-likelihood expectation-maximization (MLEM) with a post-Gaussian filter, a total variation (TV)-norm regularization, a nonlocal means (NLM) denoising method, a neural network denoising method, a traditional deep learning PET reconstruction network, and our proposed DPIR-Net method and evaluated the proposed method using both human and mouse data. The mouse data were obtained from a small-animal PET prototype system developed by our laboratory. The quantitative and qualitative results show that our proposed method outperformed the conventional methods.

One of the potentially important applications of three-dimensional (3D) intracranial vessel wall (IVW) cardiovascular magnetic resonance (CMR) is to monitor disease progression and regression via quantitative measurement of IVW morphology during medical management or drug development. However, a prerequisite for this application is to validate that IVW morphologic measurements based on the modality are reliable. In this study we performed comprehensive reliability analysis for the recently proposed whole-brain IVW CMR technique. Thirty-four healthy subjects and 10 patients with known intracranial atherosclerotic disease underwent repeat whole-brain IVW CMR scans. In 19 of the 34 subjects, two-dimensional (2D) turbo spin-echo (TSE) scan was performed to serve as a reference for the assessment of vessel dimensions. Lumen and wall volume, normalized wall index, mean and maximum wall thickness were measured in both 3D and 2D IVW CMR images. Scan-rescan, intra-observer, and inter-observer reproducibility of 3D IVW CMR in the quantification of IVW or plaque dimensions were respectively assessed in volunteers and patients as well as for different healthy subjectsub-groups (i.e. < 50 and ≥ 50 years). The agreement in vessel wall and lumen measurements between the 3D technique and the 2D TSE method was also investigated. In addition, the sample size required for future longitudinal clinical studies was calculated. The intra-class correlation coefficient (ICC) and Bland-Altman plots indicated excellent reproducibility and inter-method agreement for all morphologic measurements (All ICCs > 0.75). In addition, all ICCs of patients were equal to or higher than that of healthy subjects except maximum wall thickness. In volunteers, all ICCs of the age group of ≥50 years were equal to or higher than that of the age group of < 50 years. Normalized wall index and mean and maximum wall thickness were significantly larger in the age group of ≥50 years. To detect 5% - 20% difference between placebo and treatment groups, normalized wall index requires the smallest sample size while lumen volume requires the highest sample size. Whole-brain 3D IVW CMR is a reliable imaging method for the quantification of intracranial vessel dimensions and could potentially be useful for monitoring plaque progression and regression.

The short frames of low-count positron emission tomography (PET) images generally cause high levels of statistical noise. Thus, improving the quality of low-count images by using image postprocessing algorithms to achieve better clinical diagnoses has attracted widespread attention in the medical imaging community. Most existing deep learning-based low-count PET image enhancement methods have achieved satisfying results, however, few of them focus on denoising low-count PET images with the magnetic resonance (MR) image modality as guidance. The prior context features contained in MR images can provide abundant and complementary information for single low-count PET image denoising, especially in ultralow-count (2.5%) cases. To this end, we propose a novel two-stream dual PET/MR cross-modal interactive fusion network with an optical flow pre-alignment module, namely, OIF-Net. Specifically, the learnable optical flow registration module enables the spatial manipulation of MR imaging inputs within the network without any extra training supervision. Registered MR images fundamentally solve the problem of feature misalignment in the multimodal fusion stage, which greatly benefits the subsequent denoising process. In addition, we design a spatial-channel feature enhancement module (SC-FEM) that considers the interactive impacts of multiple modalities and provides additional information flexibility in both the spatial and channel dimensions. Furthermore, instead of simply concatenating two extracted features from these two modalities as an intermediate fusion method, the proposed cross-modal feature fusion module (CM-FFM) adopts cross-attention at multiple feature levels and greatly improves the two modalities’ feature fusion procedure. Extensive experimental assessments conducted on real clinical datasets, as well as an independent clinical testing dataset, demonstrate that the proposed OIF-Net outperforms the state-of-the-art methods.

Purpose: To evaluate the diagnostic performance of a newly developed non–contrast‐enhanced MR angiography (NCE‐MRA) technique using flow‐sensitive dephasing (FSD) prepared steady‐state free precession (SSFP) for detecting calf arterial disease in patients with diabetes.Materials and Methods: Forty‐five patients with diabetes who underwent routine contrast‐enhanced MR angiography (CE‐MRA) of lower extremities were recruited for NCE‐MRA at the calf on a 1.5 Tesla MR system. Image quality evaluated on a 4‐point scale and diagnostic performance for detecting more than 50% arterial stenosis were statistically analyzed, using CE‐MRA as the standard of reference.Results: A total of 264 calf arterial segments were obtained in the 45 patients with 88 legs. The percentage of diagnostic arterial segments was all 98% for NCE‐ and CE‐MRA. The image quality, SNR, CNR was 3.3, 177, 138, and 3.5, 103, 99, for NCE‐MRA and CE‐MRA, respectively. The average sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of NCE‐MRA were 97%, 96%, 90%, 99%, and 96%, respectively on a per‐segment basis and 90%, 84%, 82%, 91%, and 87%, respectively, on a per‐patients basis.Conclusion: The NCE‐MRA technique demonstrates adequate image quality in the delineation of calf arteries and consistent diagnostic performance for detecting significant stenosis with CE‐MRA in patients with diabetes. J. Magn. Reson. Imaging 2014;40:1422–1429. © 2013 Wiley Periodicals, Inc.